CA2997478A1 - Common platform architecture - an alternative solution for streaming data integration in a system of systems, multi-layered business transformation (specific context: byte, map, graph, stream) [english] - Google Patents

Description

Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration in A System Of Systems, Multi-layered Business Transformation (Specific Context: Byte, Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' Aerovition Digital Inc. www.a3ic.org lAerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, T 604.440.4224 1 showy al onail.4(ter ot, itiondigi t _cm Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (OS). Domains in pie-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (ROD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OSI), for seamless integration across industry partners and or body . bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap---bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space im-customers, who execute product lines in a Multi layered. Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-!natty to meet business goals. It can be an industry such as sioning, for building a capable , Data As A Service (DAAS) financial institutions. Insurance, Government departments, platform, or to build a capable platform , to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard sok,-8 , JSR 311. compliant JAX-RS interface), thru Software As lion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark. Zookeeper. Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser- Classifier Processing.
vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OS!) is an ongoing challenge in In-The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (R.DD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with H.Base or HTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to intemet or to efit numerous engineering domains and financial industries.
open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other ,,..., than variable data entered by user as input, or bulk catalog data, or applications variables posted via http post(), get () "*" -- town el.
, ...õ.
MOP
methods submitted to utilize a service controller gateway, to parse and to access any application, using the calling con-OtTP :
troller or sub (nested) controller, that recognizes the browser ster dook tance) in a Cl ,-,t variables data types, mime types. The incoming data, at - ¨
,., E-,_,,,s, .
, .......
. a insu 1 ' = 1 ter passing demilitarized zone. (DMZ, B in picture) is then . .
I-1 .. , transformed thru a data type transformation processing layer ::-, =I samin "--- t .:::::.-7.: .F.*:;.7. = ¨
, ( C). categorized by source systems types, to convert data , , -- !
1.stµ
......... ,:i- , ; = ; ;i:7:1 :
¨ ' .....
in a single common format, byte data, for non Stream data ........ =T t /4.= V.
, ..x..*.ft==
types. .¨
----. l i Sqoop uses a relational database driver, JDBC driver to Looprown. 4.. ..
...An .......
obtain these data set from source systems residing in ( C) to l `,7.,6;:f.. : ,; ,1 7-1 ,:=, ,;,-...,-, õ-:-.,:, move them into HBase HTables. via utilizing mapreduce or != ;,p-s.4-1,,,- , - Ell Niliirilft '''.' ,::, :4, ,,,C1 = : . Input Uri, ItemfulAPI Ciatit) ..,,,,..õ..
Hive. Tuple data are processes same way. After processing ......
the result set is either sent to an ETL source or to each source system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data, Figure 1: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F). and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances FIDES or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bin medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from Fifa-uct or services, The solution is proprietary to Aerovition bles' Classifiers" Columns , graph_node. graph_relationship.
Digital Inc. Our subsequent work include LLD, and pro-graph_search.weight_algorithm (tables ) to build. runtime in memory data (D). to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One -- Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in-Performance Test 0 0 yoke JSON, X ML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table 1: liLD Analytic Pipeline for An Ecosystem, In A
System of Systems.
Search engines can invoke data residing in Hadoop tables of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services lbr Big numerous verticals, using this platform as a service, for data Data Management.

fUe543 [filtration . Traral µ.1.1 faaiini Vita Aerospace Automotive 0.11,40 Pantile& Peewee Saiiretitiaaawn Reference Guam to, Dom Psne=Pf peenkents = DW
- (o)trt ^"""'".'" S , - ; A OLTP
_ a R t 4 ma õ
ORRIN ; a (E)tiive To an iiiimainEl . , IG) /*obi* Mel urnoos Hadoop(instance) in a Cluster , RietweceII , I
Pe e we. p el 1I. , c.c.-PPP". cro., Ulu ler...
Novo s tOp iireallini,ra It. loupe Rowe!. GPU Preceture Ernl:WA Clecceres t "-"rewt= 144 'gill 411 it SADX,SEO, Mince, See " ' 200Getfer Ensemtle 1 r r'¨

aehmenment Ave.*
Ecosystems, Processin. er Ma = Pipeline Ilea' five 4.5f, R @stk. API - Source Source SCurrit Consuorer a., ant p..
ETL
(A)sennce = = Input (aver, Restful 01 ,CU(C) EMP0U511 = Ineepposon.
fooncng MAL ar)f Oaks. NNW./
VineSSY Protest , Wan..
Arnuallan PrOtatien Otero! inr Goer .104 Following products at Acrovition Digital apply the above design. Specific Product Line. Context:
Line I Guided Systems For Autonomous Vehicles, Space-craft. Terrain and Underwater Systems Navigation, AisX17 Line 2 Bio Medical Scanning:MRI .CAD, Physiological Or-gan Design. AiXBIO
Line 3 Forensic Medicine: Advanced Security - Body Parts Fingerprints, Face Recognition Identity Management Systems, AisXFI7 Line 4 Model Driven System of Systems For Neuro Compu-tational Medicine. Drug Pathways Clinical Trials. AisX1)S17 Line 5 Accelerator Physics and Big Data Mining for Complex Analytics. AisXA17 Line 6 Autonomous Commercial Airplane. Space or Space Vehicle Systems Integration. AisXSA17 Line 7 Bio Medical Product Design. Clinical Research. Analytics.
AisXCRO7 Line 8Cyber Product Development Framework. AisXCF07 Line9Cardiovascular. Brain, Neural Disease Research, AisXBR09 Line 10 Solution For Auto Industry Product Line Optimization:
Autonomous Vehicle Guidance, Model Driven Configuration In System Of Systems Integration. XilencerAisCX07 Line 11 Expert Solution For Integrated Healthcare. Phar-maceutical Industry. Insurance Company. Billing Systems, DigiHealthAiXria1201 Line 12 Applied Artificial Intelligence for Integrated Compliance Management Systems.ITAR, SOX. ICA, liSRixK200T

Legend Used in Patents Pictures:
[Connected components > Bi directional Data transmission _____ 7 __________________________ Uni directional transmission Uni directional transmission EU/ISO Cnline teaming Travel Stock Market Entertainment Aerospace Automotive -. grakritaa *Me i Pertnersl Research Salestonce.corn Reference - "?."."7 '71;1 . .. ' A ' r,;11,A4',7,'"' . .1r-I - ¶ri4 - -7 Data Global SAP, Partners/
..
-4 payments .*
DW
-1- Job -ii _._ syste ms *
, ,..
:.: ::--õ __________________ ---'''" 2i ID) In frlemory Computing, RDDIGenirat '' V-1 -,' 1 1111 t; s i ' 4, ., if OLTP
Insurance it ,-'-,....i*,. A = , p_ .
, , A
R __________ i r-4 , K (E)Hive _ i :-=
Banks -.1, , 11? <-.....31- Job ... -- 1 o ..
-,, CZ3 a rgr'=1111916. , _____ .,-2 2 j < l't 4, ';
, HCatalog I, ) ; 1 I
1G) Htable (Column Ob) Hadoop(instance) in a Cluster _i _________ =, Social == - ______________________________ 4.
I i , --- -----, Network ..-fHl Re =at onsh p W el ht. Ai. ML I I i'"="4 "'a D'i'm ,, Target Systerns õ.7...._ Target System's Graph Nodes', .... ....
'N.¨ __ ----`
' = VO e., = Nodeis) Log:.- Spool Source Sink ;logger;
Parquet, GPU Processing Embedded Classifiers µ-----------IveL1...
== 1111..t al _ . ___ - - ¨ ---' Healthcare, .r..
- -=. MDX,SEO, Hospital 1' A .=
r)tence,Soir ZooKeeper Ensemble , =
E , sec --art-.
Cluster Instances, cuffv-I
_______________________________________________________________________________ _____ Government -' =
ko ' o ¨I 1 m Ecosystems, Processin : La er Anal ic Pipeline.i -- - , 1 , .'==;r-,,,..=:.:-,..`,MMV:, i t 1 1 co Real Estate I ,- ----%='7,-,Mkt:=44- '-: .,: , ., 1 V " ,'"0:k.:V.K==;',`" 4:4 1. o . 40 .: ..1,.', ,..; ...:,:.,/,':',..,:s4,4, - . =,-.,..: csi 1 Ke.=;ii*i...I:z./=;',N:z%I..,',..k:'=;:*;k:,*::::::,',.;,:
Pesti u) API *E-1 ...A'14:11.4:1$,Z.:,'SgIx co 4%4:5;:kiµ$:V.4::::MY,::5'.., Source Source Source , Consumer 1 1 . .,..;.,z,...,w-*- -:24.:,.',$:SV
Systems oa.hoo :::',:',$..,.,--,:::-4r.A. ,-,~';''`,''?:,,,...4:, ' Systems systems I 6.-..'=
V
''''''- ¨ '';'" ETL I, õ..), iiwarami ::::::;.4*. ".., ,. :"-- 11' i ., e4s,'q . 350 :;'''Sf:"..'='::=';',4'',/,,,,,,:i:: 11,-,',:', 1:."4.4V t .4., 01 ' -----)P. ' ,,, (A) c,Nururiegs Ilk ....SK:,:,;:t== ....
,,':1,=-af,P=:=4 =`= = 4 Input Layer, Restful 4PI -.1Call(C) OPeri SYsthniF tsi ...;'''''';A;;;;....y,''t,s4eõWs x'.:.:,',p; ' - '' 1 *
== õ_--..."-`77"--õ
.A.-,µ,,,-._,õ :, ArAY.,...f.v.,:;-,-,,A,,44.W. Integration, Strewing -----L4 ' EllOpOkItS. ' i.V...4.".i...';',i1:**;.,::I=?e, -' 2 CVS, t4 ,.õ....%::.*..;',4',.;:=..;:!,:.,%:2,41. XIAL JSON
ROF Coen, Fedor/rod (-) deI imIter ''-µr=-::: ' .,;:.V.,021'..."--;%-lf-: Processor Processor Processor Solution Aorterition Processor Digital Inc, Copyright 2038 Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' lAerovition Digital Inc. www.a3ic.org 1 Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, T 604.4404224 sh zifr ohm .. r ovition digital .coin Abstract ware As A Service (SAAS). from the same platform, that is built on Opens Systems Standards (OS!). Domains in pie-The following article demonstrates a solution proposed to turn (I), are the sources of data, in varied farms, such as, design highly resilient distributed streaming data (FWD, Byte Data. Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OSI), for seamless integration across industry partners and or body . bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analyties instruments used in ownership of Aerovition Digital inc. as a standard approach drug pathway discovery: electron microscope. satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space int--customers, who execute product lines in a Multi layered, Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-many to meet business goals. It can be an industry such as sioning, for building a capable Data As A Service (DAAS) financial institutions. Insurance. Government departments, platform. or to build a capable platform . to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of conunon business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEF
this solution to integrate with open system's, standard solu-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop. and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive. Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems.
types and stream transformation, we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser-vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OS!) is an ongoing challenge in In-The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru resot1ree SPARK and its execu-Patent Sought: Why this is an invention t.or services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with 11Base or ['Tables. YARN enables SPARK to , run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain ) source data . from Ser-that we can resolve by applying our solution, which can ben-efit numerous engineering domains and financial industries.
vice End Point Entry, where API is exposed to intemet or to open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http post . get 0 ..
, amp methods submitted to utilize a service controller gateway, .. ¨
OW
n:W -..,<.,..:..;-"=:.: , to parse and to access any application, using the calling con-,Illl , ._ t : ..
OLTP
. ¨
troller or sub (nested) controller, that recognizes the browser . = =
, .= ' : ' ir/ViiVe variables data types, mime types. The incoming data, af-4.. ¨ :
, õ ,..a..1prtnst ter passing demilitarized zone, (DMZ, B in picture) is then ...
¨ ====== . Htdoo ance) in a Cluster transformed thru a data type transformation processing layer ...
( C), categorized by source systems types, to convert data .
.....
in a single common format, byte data, for non Stream data ...... ....
types. .
-....
Sqoop uses a relational database driver. JDBC driver to .,..õ.õ.õ,,, : , e .,.. w=PpAne obtain these data set from source systems residing in ( C) to i ,..1Z.:;..*.. j -., .... =
move them into HBase IlTables, via utilizing mapreduce or !' -- r r.4.6-,,,, -, --- -,--¨ "---7: , ,,, = , MVO tiryer,11estfulArl :C.Ntl ..........
Hive. Tuple data are processes same way. After processing I 1..--' . 4-' .,.:'.. ..:.".;:;.., :,.,;:z .,..:::,.. .=...7..
the result set is either sent to an ETL source or to each source ........
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data, Figure I: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stow-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa-uct or services. The solution is proprietary to Aerovition bles' Classifiers' Columns , graph_node, graph_relationship, Digital inc. Our subsequent work include LLD, and pro-graph_search_weight_algorithm (tables ) to build. runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. ____ I Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes house simultaneously using JDBC drivers to interact with LLD

Hadoop platform either via ETL systems that can send data Prototype to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 yoke ISDN, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convection to inte-grate with XML or Streaming via platform component. then send to Hadoop platform.
Table I: HLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service. for data Data Management.

'P
. ... 0..= lawn Trite. St t.10.tot I...qty..... Ateosoece Autism:0o twsnott/ IIIN1161/1 t ialadar=mn iefee0t0 WO
G,thol SAP, PV1.4.1) 111111117-7 -'f''.----YrEinr..' ...!
..,=====.... = DW
. t MM..
,:7711;1864*. 1 :K 1 , v....... .
..
-,. _________________________________ / = , OLTP 1 i ..,1,4, i, i. * .- 1 (E)Hive 1 I I 11 .

.. - I .I. - - -. .... .......Ø
MI /*Shit ICOIYINI Obi Hadoop(instance) in a Cluster i%.1 1 :
So"
hittwaYtt ' IttY to . yr., r, Ne=g4 1 04 ? ,I 0.,....=`^* 1."-- Tv,.
Storm . , st.,...1 ,..,:t -.i, l' . (i=evh.
el trietsi '.....¨...¨...- I =
ft tr... t Iv; : 1 I Wt. taunt I,.. Heir, tor rq I that r- , V I ...foetid& Cl.,,,11.,, 1 , 1 I 'T....1i 1, Meat:nu.", - ......, , . , tiospnal Zooiteeper Insem 0,e . It tar C, mitt. ,rtstences ...õ= , tvytty-. Ct......
Gotemmon 1 = ' ..;-'q '-, r ',, Ecosystems, r . essin: La er Ana k Pipeline, I

::..,x^':::".311 1 ett4i.i444.......St i i e, i A..e,,AP, ' =67,V,'õ,-*:;;;',;=:;...,;i s'""`' Sow(' Sauna i Consomer i '''. ''. '';'0,1.17.._ :3'. '.' .. ;:;''','",:
3.01...,. Systems I 44-4-1Y3tel" ETL
. IA) serwc= ! - ¨ 4,.pti-:,,-c.1,14.4'1- . , '.. , .. Input I.,.ayer,Restful4P1 $0,N(C) I
, friftintmt, vow, , Endpoints oc44,4=2W.44 61,..11,.....
I ;,,,,,,,,l'IM:V="' ,Alt . SON C./S
ROI Oar horkorttyd del .Y.rit I . =,=;',. . 't " A Urrt t la ' ' ' " ''' f=Ott 557 0.0e SW, P=ticessor ken*.
Aogrimw Petal iv Cop*. umi Following products at Aerovition Digital apply the above design. Specific Product Line. Context:
Line I Guided Systems For Autonomous Vehicles, Space-craft, Terrain and Underwater Systems Navigation. AisX I7 (specific context. Graph. Stream, MAP) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' I Aerovition Digital Inc. www.a3ic.org lAerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2. Canada, T 604.440.4224 I sir alum hmanOaerovitiondigit. al . cot n Abstract ware As A Service (SAAS), from the same platform. that is built on Opens Systems Standards (0S1). Domains in pic-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as.
design highly resilient distributed streaming data (RDD, Byte Data. Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OS1), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints.
independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope. satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space im-customers, who execute product lines in a Multi layered. Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-tnally to meet business goals. It can be an industry such as sioning, for building a capable, Data As A Service (DAAS) financial institutions, Insurance, Government departments, platform, or to build a capable platform to interact between . Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solu-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platfomis , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper, Sqoop, Pig.
faster execution. We propose a solution based on Hadoop Hive. Kafka for stream management. The solution converts ecosystems. useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS
Present Research Challenges Description: Common Platform Solution's Building a multi-channel multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS).
Data As A Service (DAAS) platform, Software As A Ser-vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In-The platform provides parallel in memory data computation, formation Technology or Systems Engineering.
for faster processing classifier objects such as Map. Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform. Soft-interact with HBase or liTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain) source data , from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to internet or to efit numerous engineering domains and financial industries.
open public domain, (A). passes thru a common authentica-tion and authorization security layer that restricts data other : - . õ , _. . ...
than variable data entered by user as input, or bulk catalog --- ¨ ¨ -........
data, or applications' variables posted via http post(), get () .¨

...
.....
......., Nem methods submitted to utilize a service controller gateway, --, DW
0,,,, .....4.16,"=^=11;
µ , ii ;
to parse and to access any application, using the calling con- Ei OUP, , troller or sub (nested) controller, that recognizes the browser .
I ' , .
variables data types, mime types. The incoming data, af- P; E,......a.
,...
, ,,.., nce i ter passing demilitarized zone. (DMZ, B in picture) is then ...., .... Hactooptinsta ) n= Cluster , . .
transformed thru a data type transformation processing layer -. . ,-. "..7' - ÷:, 7 ' ' '" " ''' ' %::=:.:7: - ':::;.:- = -( C), categorized by source systems types, to convert data ,:,-.= , ......- f..-: ,7,..i :
....., in a single common format, byte data, for non Stream data -7.
¨ ¨.
types. Sawa,. !' :
.....
Sqoop uses a relational database driver, JDBC driver to .
tiainw :
obtain these data set from source systems residing in ( C) to :,..,. .1. ';',.' .r...- ¨
ETt move them into HBase HTables, via utilizing mapreduce or ... !,. .
.,¨
*. .: , 7 : input eswitlitofuIAPI cakc) ,,,.,õ
Hive. Tuple data are processes same way. After processing the result set is either sent to an ETL source or to each source .......
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data, Figure 1: LSSI System Of Systems. Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa-uct or services. The solution is proprietary to Aerovition bles' Classifiers' Columns , graph_node, graph_relationship.
Digital Inc. Our subsequent work include LLD, and pro-graph_search_weight_algorithm (tables I to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in-Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table 1: HLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

Idecetreg 0,41.11, Uwe, Two. S.0:1. Mort et 0,r-i.e.-Pm Av. coyote Automotive Down.
1.W.Www... liteterince -- Ow embat ter, Pertversi .t. . INICIII-- = .
leerelleel DW
.....
(0)kiflient* 40¶11,44"ft, 4,40,4wintts. , 4.--- = S , IIII 1 1111011.111 OLTP
w : r.:=-:, L=--- E.
, I m , ; = : . 1 A . 1 !
*i.11iii, ' * I
u(iIIIIIII 1111 1401; K 4,1 vac ., D . I ' . i (E)Hive - Ii111111111110141 H,1 j;14x-' 1 a44.4. , op alovninosEll i i . ' I Klelle. 1 ' i IQ IRMA 'Colwell:4e) Hadoop(instance) in a Cluster . r v Sows P N 1 _ * -.1 etvrolt . is.
' ' pew it-ceve o AO ph A . Mr. ir=vc..g.,== Ter therm 4., veriest Some*:
r ----F -G. so, Pieces t -Note',. 106 c t- - ---- - Means. me sviceenee.
. 0.1 ¨ 1 i I
I
I
. . ..... la Ella IN a a a 1 .,- 'i - " : ' . : - -7.;
'.' -.=-,r- ' moicski, ....... 4. u ..:,--434= - dooms*
I niognim Insynb,e , VI' :=-cwiltt.kvIll õ, .
cr.. l.,.t..nc... 11104, ,, .1,' . t =
Government I
Ecosystems, Processin_ An .' Pipeline.
Iteolfrnate - -= i .
. .
Reetternep e-- Source Source . : ,...., Consumer *seems Eli .
immovimi, , , = tr v==54-4,,e r , Input etnyer,RnstfuliP1 4,¨_i.aiip41 , ii......w.
CVS RCP
ettaireeee.tosetee .1011 Vocessor ' g Pro(????, felmiseesesseteme Processor Oleereirs.
Mon*. Me .
Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 2 Rio Medical Scanning:MRI ,CAD, Physiological Or-gan Design, AiXB 10 (specific context. Graph, Stream. MAP) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahmanl lAerovition Digital Inc. www.a3ic.org lAerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, 1604440,4224 1shawr'ahr,uwaevovIioodigta1.com Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (0S1). Domains in pic-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (ROD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (0S1), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory. for customers, who execute product lines in a Multi layered, Very aerospace product design, space research data for space im-Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous necting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sioning, for building a capable . Data As A Service (DAAS) financial institutions. Insurance, Government departments, platform, or to build a capable platform , to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solu-8 , JSR 311, compliant JAX-RS interface), thru Software As lion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products. to that supports both abilities, control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs. training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper. Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems. useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation.
we propose Linux Cent OS
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables. Platform As A Service (PAAS).
Data As A Service (DAAS) platform, Software As A Ser-vice (SA AS ), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OS!) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YARN enables SPARK to , run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using IITable data of Masc. which stores classifier data types or We have completed System Architecture High Level De-graph data elements (0) .
sign ( HLDs) . Alongside we have filed proposition to de-velop products to solve multiple industry complex issues, The data from (H) domains, are in general referenced as Domain Data, The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-ctn. numerous engineering domains and financial industries.
vice End Point Entry, where API is exposed to intemet or to open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other , than variable data entered by user as input. or bulk catalog data, or applications' variables posted via http post(), get 0 Noevre.
1 W '' ISCOM
Dm =
methods submitted to utilize a service controller gateway.
-- ffi DW
= ....
¨
to parse and to access any application, using the calling con-10. kte.,. (4,tomit, -,=....... , : .. M
, OLTP
:
: ---troller or sub (nested) controller, that recognizes the browser i ,....
variables data types, mime types. The incoming data, af- = - .' ,,-, ........5 ....
..,..,,,...... clop,,,sta ilusr i . h ter passing demilitarized zone, (DMZ, B in picture) is then ¨
Heol nce)n a Cte ki.... .
..
transformed thru a data type transformation processing layer ( O. categorized by source systems types, to convert data in a single common format byte data for non Stream data 4..10 =
. . ......
types.
..:.:.. .. 1 Sqoop uses a relational database driver, JDBC driver to obtain these data set from source systems residing in ( C) to -,. ......
¨
move them into HBase HTables via utilizing mapreduce or 14-Ti 1 ,.......
........ ,,õõ, En.
, PONY' = ..,M4., 1:.(..,,..004., 4.14q lifie*
Hive. Tuple data are processes same way. After processing ¨ ---, r ,, ........... --,....- the result set is either sent to an ETL source or to each source -...,,,...
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data.
Figure I: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories. as HFile, and then to HBase for star-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring . Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from UM-uct or services. The solution is proprietary to Aerovition bles' Classifiers' Columns , graph_node. graph_relationship, Digital Inc. Our subsequent work include LLD, and pro-graph_search_weight_algotithm (tables ) to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One F
Sti-iie--fi. Stage Three A Customer can maintain OLTP or Relational data ware- IILD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 Hadoop platform either via ETL systems that can send data Prototype to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data. thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table I: FILD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables .. System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development. performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

- - 0.4. = les-egg T====1 :!a:11.1sItt 1,==1.r. -et At)ososet A.AorAnqm= Iducsboni ''. tPOPPIP PIO
Portron: Roseman takmacAsA., Pefeteme illegal UP, Den rangers/ - , - ...
P,.......= DW
. I
i ID) tri tAtrnory Ce;iiiiiiiine, 4.,',......;,¨ --7- s . EN
. , e......................../ ' 1, P ., I .
- = , A i , .,...... . OLTP
. 1.
I a 1 ,, . ,_,,,, ,i .., ..........

, ,.. i IC I (E)Hive E?....iirga , `.-. - = - . , = Katelas ,. . i IG) litsta le 'COWRY, ON __ Hadoop(instance) in a Cluster [, - ' `j = 1 i = =
Soc.a, _- , , .
, e V a-.0- G ..,, to et a ki. < I .nam-.. I- T.t},....,.. .
Tv.,?;,.,,,,, , 1 t ',tr.,. La; . - --- - Steel koce yr wig*, ..., L
I , Hialkell. ^ I = ' Igo Ilan al = iii oillis4...k, ---, .. . , 1 ZooKeeper linse,b,e =
Cutter ,ns=enceS :Oa 4.=
i=
GIOMIGIAlit 1 _r ...
.....
. .
- Ecosystems, Pr. . ....L. .._/i a ic Pipeline= , 11 Neal twice .
Source SOur(le Sow., ., .
Consume) ...... .' -AA.4)01.A,AAAerog systems Yriterns s', xte rrs ETL
. ' (a)seivice "s . = - " ' ' f , ' Input layer, Restful API 'Call (C) 00====, Syq=10'd ' L-6. EntiPekill Vs "
tat=rec.., term, C
ANL ..,01,1 -U.), Om. kmeerlrog1 Voreis...7 I mess,' P OE f 510f %caution A..irlon . . ..... . - Prace ;SC, 0.04, ,..c=
Cory... SDI&
Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 8 Cyber Product Development Framework. AisX('.F117 (specific context. Graph, Stream, MAP) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' Aeowition Digital Inc. www.a3ic.org Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2. Canada, T 604.440.4224 s awrahm am:4(1er ovi t iondigit a Leorn Abstract ware As A Service (SAAS), from the same platform. that is built on Opens Systems Standards (OS!). Domains in pie-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as.
design highly resilient distributed streaming data (RDD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (0S1), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory. for customers, who execute product lines in a Multi layered, Very aerospace product design, space research data for space lin-Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data . telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-many to meet business goals. It can be an industry such as sioning, for building a capable , Data As A Service (DAAS) financial institutions, Insurance, Government departments, platform, or to build a capable platform , to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over WE
this solution to integrate with open system's, standard solo-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large.
A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms . and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase. Parquet, Flume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems. useful for more than half a dozen complex data complex data types into byte data. For Operational systems.
types and stream transformation.
we propose Linux Cent OS
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser-vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards OS1) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS). Data As A Service (DAAS) platform. Soft-interact with HBase or HTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues.
Domain Data. The ( multi domain ) source data , from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to interact or to efit numerous engineering domains and financial industries.
open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http post(), get 0 ¨
-¨ um methods submitted to utilize a service controller gateway, _ . ow to parse and to access any application using the calling con- gine tte,,,,,, tisliaise, ...ewe troller or sub (nested) controller, that recognizes the browser , io.k. ,,,.. .
variables data types, mime types. The incoming data, af-ter passing demilitarized zone, (DMZ, B in picture) is then ,. .
............ tledooplinstance) in a Cluster -transformed thru a data type transformation processing layer .L 1L
--( C), categorized by source systems types, to convert data .?..t. -in a single common format, byte data, for non Stream data fAXIWI I
an... !
types. a' I
Sqoop uses a relational database driver. JDBC driver to I...., r - ,11-j,:?!:. "PO"
obtain these data set from source systems residing in ( C) to ......,, i ' l'''' : 4 ' ' ,..=== 'O.,. ,,, .
move them into HBase HTables, via utilizing mapreduce or ,,...õ
- i',--ETL
= Mi... t4 4 .,,,, ., = .
, . = hymn 1,wet,11.tbi. , i r APO) Hive. Tuple data are processes same way. After processing , ,.. ., ,N=
,,,, ...;,.. .' lon.sainwe the result set is either sent to an ETL source or to each source teen. .1 system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data.
Figure I: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HOPS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance . first to HDFS directories, as HFile, and then to HBase for stor- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa- uct or services. The solution is proprietary to Aerovition tiles' Classifiers' Columns , graph_node. graph_relationship.
Digital Inc. Our subsequent work include LLD, and pro-graph_search_weightAgorithm (tables ) to build. runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in-Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convection to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table I; HLD Analytic Pipeline for An Ecosystem. In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying A1-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

I-tUISO tcluciteln , 0.),) e terel T.)1.1 ',lire t., F, )). ,-. Arrospece 0.1.'1,7MCINr mange:10n t For=tos=rs/ *reefer&
1.4.4emk te., *event e ) Opts eivemesi 1111121111 , .
......t. DW
WiteMB ti lei .
=
(Mtn fAomory Computing. ottos....b... = - --= ]
te41/1.01. ' I i =,... -- _ r - i p OLTP
' .',I can 4 l It ewes .ke, o g( (E)Ilive ..8 ................................................. . , ,õõõ6--....-: ...
. -:
= .Cseeleg i (G) Pliable Ito' wr,,, obi Hadoop(instance) in a Cluster ,1----7----j sc.(.., . - -. = -.....
=e e" N..) o , e ire .. ),,, vpr true* .
A, i..! ),..., 1 .------- - 1-, ,..õ,õõ..., , ---..- , ,-, 5,2315eve = 1..11.1%.,) ' De.
rrir.7 Inemort. CPU Peotressati ImIrrefriktf tlatellites i -, . .
Heeldcare. l' f'= i ['::: '''', !..j:. F'r. -i li.'=-**i .;. i.. 1.
ft*" ' ' ' INelloal a/0M %DV
, Zookeeper Ensemble . 3 le NI, cluster 'fITUIT.Cei 1 C+.s",..
pli..11EV. i Gownsman KS
- ..... j Illerhe 1 Ecosystems, Processin : La er Anal tic Pipeline, i I
' 1 t=ell tette l 6.1,"==4" 4 , :14.".-7.1:,."0,t:.:=,',5%. .V = = .., i Restfl AP, ' , V.1,...',=t<=4,..., so,c w,õ,,. somc.
....
c,,,,õ.,. : =,.. ,)5 ci,...?:.,,X.:e S y St = rn 1. 5) Ste ,1 5 sl,sienn ETL
-,.. :.r.11..,,re.4.>%*,i.4,4 ' . r ¨ (A) SelvIcr 1 4e.iii44";o0Naotial ' Input layer, Restful API ;Call(C) ' OPPPSVPolie ' I . Endpoint% 1 t."t',W'''"'",!'X'3V'''''!',<
ImeepsUenientemieg i ra''`Wtfrg/t"'14 V.1,. .50)1 :XS
t311.=Motef sac OM, lePeneed ' Vocessor %ix puce P=ocelso, Lembo %modelle.) fOrtanot.
Covr#1 MI
Following products at Aeros ition Digital apply the above design. Specific Product Line, Context:
Line 9 Cardiovascular. Brain. Neural Disease Research. AlsX BR(19 (specific context. Graph, Stream. MAP) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' Aerovition Digital Inc. www.a3ic.org lAerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, T 604.440.4224 ishawrahmanktacrovitiondigital.com Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (0S1). Domains in pic-The following article demonstrates a solution proposed to ture (r), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (RDD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (0S1), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery : electron microscope, satellite data.
to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space ii-customers, who execute product lines in a Multi layered, Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dusty data , telecommunication or wireless mobile device data. or governance data for compliance between regulatory Introduction Target Audience agency or government partners These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous necting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sioning, for building a capable , Data As A Service (DAAS) financial institutions. Insurance, Government departments, platform, or to build a capable platform . to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry, We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solu-8 , JSR 311, compliant JA.X-RS interface). thru Software As tion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS.
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser- Classifier Processing.
vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru . resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YAR.N enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (C;).
sign ( HLDs) . Alongside we have filed proposition to de..
The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues.
Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to internet or to cfit numerous engineering domains and financial industries.
open public domain. (A), passes thru a common authentica-tion and authorization security layer that restricts data other ¨
than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http postO. get () ........
--AIM ..
methods submitted to utilize a service controller gateway. =
ti. - OW
iii..4.,..;.,... t as ;
to parse and to access any application. using the calling con- ¨ OLTP
troller or sub (nested) controller, that recognizes the browser 4, variables data types, mime types. The incoming data, af-ter passing demilitarized zone, (DMZ, B in picture) is then .-HadOtT(StanCe)ine Ouster ¨ .
transformed thru a data type transformation processing layer ...._ ( C), categorized by source systems types, to convert data in a single common format, byte data, for non Stream data -....õ
types.
¨,-õ
---¨
:
Sqoop uses a relational database driver, IDBC driver to t,,,,stsmt, l'r ,,... 1J ir, A a 0.14 obtain these data set from source systems residing in ( C) to */* 1 1 ' :=,-..4,,J ,--,-,.
move them into HBase HTables, via utilizing rnapreduce or _ ,, I '..;. : ei,-i. , 1 .7.-- ,r-, Ett.
Hive. Tuple data are processes same way. After processing 1)' 1 ; __ ,'_:, 1 ..Z... .X., ::=
the result set is either sent to an ETL source or to each source ....., system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data, Figure 1: LSSI
System Of Systems, Domain Integration. So-HCatalog is sent to Pig (F), and afterwards thru mapreduce Intim using Open systems Ecosystems operation from Pig to Hadoop instances IIDES or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for star- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga- fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa- uct or services.
The solution is proprietary to Aerovition tiles' Classifiers' Columns , graph..node, graph.selationship, Digital Inc.
Our subsequent work include LLD, and pro-graphsearch_weight_algorithin (tables ) to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HID Yes 0 house simultaneously using IDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 yoke BON, XMI., or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table 1: fiLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of fiBase invoking APIs utilizing these business data pro-cessing flows, Thus a common platform interface will re-duce operational cost and consolidate processes in one plat- totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

, ' WWI = Oallae leireou T...01 St dee S'e,e et l".e.ta.aaat die0SPINN AIROftlilelbe Iducnan/
p.r...../ RetasnA
Seltelefeeeme , addleeledt, ,õ,:"
... ,,,,=,,"1.2 õ " ,4"deidnit4"B'''''l ..
blade' MP.
Panders/ . Malli . :"
MOM = Jed ow - , , ' :=':
0.P.IYF911"1/1.1( e'aliti"4;100- ded:eftbe: '472:: 1 S i '. 1.11 i 1 ,,,,,,,:::"A" 'le = , " ' OLTP
,......- -, A we.... .t....1 .
J. 1 , = ! R
1!
* 1 1 gm . - I t lee 0 4 i 1 (E)Hive L.:
;
.
. = ,. , Hc.,,a. ,.µ
1 .
IQ Mob.< 1Co'unno 'obj Hadoop(instance) in a Cluster L , , se,*
1",; 1 s., anew". /.4 =,= T....t $...,,,, , Try. Sane.. I
GropWNW.5l, :I . --- - ifealia6r. Seq dew, lhdeeetisanwe 04e PonorM MU Prom...woo. Embedalkal rtrollts , I I.Wowii-1, r r---,- , : ' ,_ i - -Me/001cm* , .
Mospisi =
, ZooKetper Enowroble .
1 1-= = ,..
Cluny. $notonees :woo tnrasstr:
alleg...
Ecosystems, Processinklarr Ana it Pipeline.. i . 1 i Reel Estee i''' ;
..e.õ,,-..,,,a,,,,, Restful eet ^ _ ,,Istke_ i"ii,&.4 sc,,,, e Se..
IPa. t SC( aft ..... .-..,,". .41 sy... (0,,,,er. sts., sl Stems . .' Ell.
Input Layer, Restful/41)i 1,Cali(C) 0.....s,....,.
- L_,. Endrornts :-. : --k=q.444µ-6,..A.05,44 Mewed., Marne.
redS"d"V*4=Seg*.?+e= ( VS
ehl. . SON 401 Drs.
heeeeted .
L ..,õ0õ,tw=-...m ___ gi4,:p,:x.k.r, = = . --- P,eSSO t eAe Vorcsos de "'Ie.
Pr3:1[311, Or' arkodnmpltee DifM114 /et Cowed,. 10,1 Following products at Aerovition Digital apply the alx)ve design. Specific Product Line, Context:
Line 10 Solution For Auto Industry Product Line Optimiza-tion: Autonomous Vehicle Guidance. Model Driven Configuration In System Of Systems Integration, XilencerAisCX07 (specific context. Graph. Stream. MAP) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: (;raph, Stream) [English]
Petitioner: Shaw Rahman' I Aerovition Digital Inc. www.a3ic.org 'Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, T 604.440.4224 I sliawraltramAkacrovitioncligital. Corn Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (OS!). Domains in pie-The following article demonstrates a solution proposed to lure (I), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (RDD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OSI), for seamless integration across industry partners and or body , bodily parts scanners images( bnnn, finger prints.
independent subscribers. The solution is maintained under bin medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (ROM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory. for aerospace product design, space research data for space im-customers, who execute product lines in a Multi layered, Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sioning, for building a capable, Data As A Service (DAAS) financial institutions, Insurance, Government departments, platform, or to build a capable platform , to interact between = Aerospace Digital product developers. Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over SEE
this solution to integrate with open system's, standard solu-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SA AS), from the same Platform Service (PAAS) vendors numerous platforms, and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadocm, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parqu.et, Hume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems.
types and stream transformation, we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser- Classifier Processing.
vice (SAAB), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
integration Standards (OSI) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering. for faster processing classifier objects such as Map. Topic or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YARN enables SPARK to , ..
run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues.
Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to intemet or to efit numerous engineering domains and financial industries.
open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http post , get () --,... ¨
-- mai methods submitted to utilize a service controller gateway. : ¨
ei - OW
to parse and to access any application, using the calling con- ..11 0.........õ4õ....,......, .
.... ,111-troller or sub (nested) controller, that recognizes the browser ¨
' tflHirt ' .
variables data types, mime types. The incoming data, af- ;
, ...i in e utte r ter passing demilitarized zone, (DMZ ..... c. .....
Hadooplinstance) CI
, B in picture) is then ¨ :
transformed thru a data type transformation processing layer __ ......., ==,:z . .;;;= - , : - r7" ',=,::,-:: ;',::õ; .
¨ ¨
( C), categorized by source systems types, to convert data !!
in a single common format, byte data, for non Stream data ...... --...";
types.
.---...- :
, Sqoop uses a relational database driver, JDBC driver to ftdtil.
I
obtain these data set from source systems residing in ( C) to move them into HBase HTables. via utilizing mapreduce or ¨ , :,, i ,.... cz,:tv7,:. t... I'll.
. 04.**1+ - , Inc,a 41., itottiw ..r. ,(414(,) --Hive. Tuple data are processes same way. After processing ,_ --..i.,,,..) s ,,.. .......
the result set is either sent to an ETL source or to each source um_ system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data.
Figure 1: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F). and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to IIDES directories, as HFile, and then to HI3ase Mr stor-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring . Map data is nesses willing m transform to Al, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet front HTa-net or services. The solution is proprietary to Aerovition bles' Classifiers' Columns, graph_node, graph_relationship, Digital Inc. Our subsequent work include LLD. and pro-graph_search_weight_algorithin (tables ) to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using IDI3C drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype to Sqoop or via API interfaces to a search engine that can in-Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Al Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table 1: liLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of H Base invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

. .
ILPRIC Ed11011,0111 iiitaisliaik ' Ar, w.soror ,s=l=so 1.106111seh la l= rt,utor ,,.....C.,,,,i,,,,,,,,OPPIt ::. Dia Grewsl3w1 =
wen's./ 4 Oev=wwsl. DW '"
eases, CI j'''' A
I D) In 11.1tmory GynputIng, sap.,,,..n., ; S 111 i !
i OLTP ' i ,. r, = A wane.. = = =
tants , 0 ; (E)Hive 161 Setable ICol wren Dbl Fladoop(instance) in a Cluster 1 , , 1 Soca) I l.---------..- 1 ' , Network . õr . I, I I:
I
iH: t.,* r >a, c We ins w 1.1. , I trrsorserwww ifs-s,,,,, T.q..,,,..õ. . .,.,e,,..... 1 eA nish tbere. ' ' No.e.s lerg : I'M low re= S..

p..... Grst= Pe.ormsrwfr f.wresschirwet sists.,eors ( 4..iww:1-1 I ..' : isostaili "*""e ''¨^ = i'ii MON rp til et WI W4, rot;4-, , Zoogeeper E,senlvie i tr,usttr tnronces '1 ?'',','fµlitt t114," 1 iV't Ctoodtte t tr....
Govenmern :71116'' t.
I.. - : -*
Ecosystems, Processtn: Later Analytic Pipeline - . , Real Mgt ................. 'vs .A,..:,,sw .....
' '4' /. l'..=
44.,,ir=ri!..f.= - 464(4' ''''k'''',. , ', ' liestfa API = i Source Source .,...-Welee>A.4". !'===4-A.%. " "
.11..Plt:' ; ', ',. S1,1,,,,, 'Meng, sYstems ' : ETL
,,.. ' --- ....., .... .=.-. '.. ..=.:11.,, ' -- (A) Senke "'"' ,;; ;
input Layer, Restful API IrallIC) i endpoints 1/,'Ar=======: >wr ,,rwowi. "
Wwwvenew. Women' k. ... 'OX,,101.1 Fr I-mem, P=M[SSOI fahrtMotAarditam C.'S(' 4r3:E{SCI Crops&
kw.
Craseerths "
Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line II Expert Solution For Integrated Healthcare, Pharma-ceutical Industry. Insurance Company. Billing Systems. Digi-ilealthAiXrial20t (specific context. Graph, Stream) =
Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Byte, Graph, Stream) [English]
Petitioner: Shaw .Rahmani I Aerovition Digital Inc. www.a3ic.org lAerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6F1 2B2, Canada, T 604.440.4224 sitawrahm ara4arrovition yi tat .com.
Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (OS!). Domains in pie-The following article demonstrates a solution proposed to tare (I), are the sources of data, in varied forms, such as.
design highly resilient distributed streaming data (RDD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OS!), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aemvition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space irn-customers, who execute product lines in a Multi layered, Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous meting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sioning, for building a capable, Data As A Service (DAAS) financial institutions. Insurance, Government departments, platform, or to build a capable platform , to interact between , Aerospace Digital product developers. Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solo-8 , JSR 311, compliant JAX-RS interface), thru Software As lion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zook.eeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation.
we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel . multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS).
Data As A Service (DAAS) platform, Software As A Ser- Classifier Processing.
vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OS!) is an ongoing challenge in In-The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map. Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( IILDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain ) source data from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry. where API is exposed to intemet or to etit numerous engineering domains and financial industries.
open public domain, (A). passes thru a common authentica-tion and authorization security layer that restricts data other . ., .
than variable data entered by user as input or bulk catalog data, or applications' variables posted via http post(). get 0 .-.. sem methods submitted to utilize a service controller gateway, ¨ .. -ow =
¨ - ai to parse and to access any application, using the calling con-' j troller or sub (nested) controller, that recognizes the browser J.
:
variables data types, mime types. The incoming data, af- ".-= - 4 -..
, - (In in ter passing demilitarized zone, (DMZ, B in picture) is then - ---H.doop stencel * Cluster transformed thru a data type transformation processing layer ?...L
CNN SESIESI " - -=:' ' ':+i:.--r: - .7--.;=72 ( C), categorized by source systems types, to convert data .
....... :- = - - . - [
in a single common format, byte data, for non Stream data .-......,,, .... .
types. --.
3.4^,-..<
...,...., :
Sqoop uses a relational database driver, JDBC driver to ,.... ..-0. r, . = l obtain these data set from source systems residing in ( C) to i i ... .,-*,-- ¨
...... ¨
move them into HBase HTables, via utilizing mapreduce or 1 "*"", '7"*" Mo.". ETL
., ,,,_--,%,. s .., = .
Input ("optliestlulAFt 'CAW/
Hive. Tuple data are processes same way. After processing =_, , -..õ...:.
the result set is either sent to an ETL source or to each source .
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data.
Figure I: LSSI System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lotion using Open systems Ecosystems operation from Pig to Hadoop instances HD/7S or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai. for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa- Oct or services. The solution is proprietary to Aerovition bles' Classifiers' Columns , graph_node, graph_relationship.
Digital Inc. Our subsequent work include LLD, and pro-graph_search_weight_algorithin (tables ) to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDFIC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in-Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convenion to inte-grate with XML or Streaming via platform component. then send to Hadoop platform.
Table I: HLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

, ' 1114" ..-, *Aloe Lorene Treeel Sioct frieoo;
/.4.10101..401 AereelOIK e Autorooleee Slerch al11061101.
1 Perteeett Ali I Splademeree Ora 1 I
. eennea./ ' ... .
Asseesois DW
, , , " ":4( t.....0,40õ.6, "1õ::111,00, 1 i=
i_ 1 ' :
f71,--,......=
11,õ,110,,,r1 4 , rõ," ,t1111, ,,noollp, ,oh,o A
.....õ,,,, . . OLTP i !
,.1=;..--,--,'. RI
11111,1011r,.1,;õõ,.1.1.1.1.1,1111.11 R I
= IA :: AIL .1 ¨.111110, .õ, ...
III ,õ1õ,111, ,, = . 1.._ , µ = , Keteke 1 ; ' %IMMO! iCeiumn 010 HadoopOnstance) in a Cluster ,.

wow =
Netwk t.- 1 = ' , ' ' = i .;-- =
g ! ;
. .raph fINPSS, i I
Fr=o?1,11-o17".4 = 1.1111gilieft . 1 IIII Ma a a a a MI , V-14".4:1111W 7r '77 I
A.-i;t:tv.;.- z : '1?-^Mil:a dithot,iter ' - = Zeolfeew &rarebit 1 1 , PC .......
(WPM IPSUPPMS. Ptarh, i ! ' tvo--sommr Nevernment i : f' ,A.A
1-1. ---, L ......
' Ecosystems, Pr =__..., _,, ,,.., .i._., ."
Pipeline*
! =
' . . .
, P epth1 API ." - - Source Source Consumer i . . =,,t0 Systems ,),Ite,,, = .
.., ETL
(Mum*. iSpraNipt¶. . Input layer, Restful4P1 Call(C) OPP.Symnal ' I ... endpoints -CVS, htegrePipt. Pummrip Mi. ;$0.4 PDF Om.
PoPenrevd Clti.PPI AP
. P.PPt OPP P,XtSSS, PrOtteser &Artie.. ASIIMIDA
. - = .. ProceSSCP 04pial PK.
CaveqPrt IPA
Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 12 Applied Artificial Intelligence for Integrated Com-pliance Management Systems. ITAR. SOX. ICA, liSRixK2OGT
(Specific Context: Byte, Graph, Stream) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Graph, Stream) [English]
Petitioner: Shaw Rahman' lAerovition Digital Inc, www.a3ic.org Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada. T 604.440.4224 shawrahmanktfiaerovitiondigital ,com Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (0S1). Domains in pie-The following article demonstrates a solution proposed to ture (r), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (RDD, Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OSI), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints.
independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for customers, who execute product lines in a Multi layered, Very aerospace product design, space research data for space itn-Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data . telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous necting multi -disciplinary business domains, to provision.
engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-many to meet business goals. It can be an industry such as sioning. for building a capable Data As A Service (DAAS) financial institutions, Insurance, Government departments.
platform, or to build a capable platform . to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solu-8 . ISR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SAAS). from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper, Sqoop, Pig.
faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems. useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation.
we propose Linux Cent OS
Present Research Challenges Description: Common Platform Solution's Building a multi-channel , multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser-vice(SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In- The platform provides parallel in memory data computation.
formation Technology or Systems Engineering. for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YARN enables SPARK to _ _ run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using !-liable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry. where API is exposed to internet or to efit numerous engineering domains and financial industries.
open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other = _____________________________________________________________________ , _ than variable data entered by user as input or bulk catalog .....
data, or applications' variables posted via http post(), get () --¨
¨
¨
sicas methods submitted to utilize a service controller gateway. 7...-:::' -to parse and to access any application, using the calling con- õ.......
OLTP
troller or sub (nested) controller, that recognizes the browser : .
variables data types, mime types. The incoming data, af- , : , .õ
,.4. +.= ....E. HarlooOnst=nce) in = Miter :--::' ' ''' ter passing demilitarized zone, (DMZ. B in picture) is then ,..,. 4 =
.
transformed thru a data type transformation processing layer ' ism ssmiss .- F.-.. .,:. .
.,_ ( C), categorized by source systems types, to convert data !,:=.t `, =.,I. .1214 ' =
in a single common format, byte data, for non Stream data ...ow ...."!.
i ' ¨ Z:=..
types. .....
=
Sqoop uses a relational database driver, JDBC driver to ta.......k r,...1,6*.,",,,,. = ; , = :
, obtain these data set from source systems residing in ( C) to .."
move them into HBase HTables, via utilizing mapreduce or ¨
",.......] - - , ,::!,:õ.., Inoue 14yKlboultiAll CACI --Hive, Tuple data are processes same way. After processing 1 L,.:.'" = :'-:.,:::.`
the result set is either sent to an ETL source or to each source _ .......
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data. Figure I: LSSI
System Of Systems, Domain Integration. So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai. for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga- fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa- uct or services.
The solution is proprietary to Aerovition bles' Classifiers' Columns , graph_node, graph_relationship, Digital Inc.
Our subsequent work include LLD, and pro-graph_search_weighLalgorithm (tables ) to build. runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects.
I Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- fILD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 Yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data. thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table I; HLD Analytic Pipeline for An Ecosystem. In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat- totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

, Onto lung Tenn 0!000 Ma A et 1010,010=0110:
MOM/CI AutailleteVe 14 Va 14" "rag= 10 1.04000000 101110100011 ' L
SOW0.010.. ; Inftlent= =
' Pertneni 11111011111-7:-: -.0],---------,-,.. i P=00,0010. 1 -. los DIN i 0101m, riellaci ....
. .:
, . ! .
: ism* , --..=i4 444 K , II
(E)HiV1 Ie ' i =
4 _ , = = . ---- _ ......Angalvd,....1 = 4 1 I
!, pteuin ' 1 .
' 1G) Nesbit (Col wan DPI Hadoop(instance) in a Cluster r ,4=i1- =
1.,µ .
I W itelitanth0 'delft. IV.GIL
*_.
, . 0.'40 N0Oeft.
hoe*** VW:. i " " 10001)00000 1, 110U010 .., 1 = i 000, nu 1.1 p=-0e , = -a IN II MI Mg. ''' .
. MOK.3E0.
Zotteeper Ensertt'e =
(tulle, I nstences , , -- .

....., .1 = : _,,,. Mat "
GOV00000111100 !
0 . -Ecosystems, Processin : la er Ana lc Pipellne, = I "1-- I -Asalt.tOn . =,"---.. ^t""`/=^P.59 ,ittAP-04.1 r:-. - ,`/=t4,60*,=)00=MO
- , el-rtrrtiteAr 1 "
. . , , . P rX.,.."4..,====1,10/5 ' .
*nth/ API 0.,..= '...., , =af; Source Source C001P/Alte ' , ", õ õ,,,_ " .'lltrtrre' ,4'. '''''''""
SyS10000 01,1001, = L ...
- .. ETL
---= : ktalitilfinerg=.,,,.. -- - - = Input Layer, Restful API ;Call(C) 00=011=00=00.
0 . - 00000 0 In0nrernielo, Strmer=Ire '1-04-111A111!!!'t= ' . 400.1.00.00000000 , 00.4000). 000000000000000 !".
..0 ON C0'S 000 Oen. fs=tent=01 dir m te = At-4,1!...f.,$4.......100 VOcessa vocesco P=rxesso= Sokol*. Ateetker.
- - .....k.vAriv,Foraweee 000:0"0 trout rrte tonne. t0111 Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 3 Forensic Medicine: Advanced Security - Body Parts Fingerprints, Face Recognition Identity Management Systems AisXFI7 (specific context Graph, Stream) CA 299'74'78 2018-03-06 Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Byte, Map, Graph, Stream) [English]
Petitioner: Shaw Rahman' 'Aerovition Digital Inc. www.a3ic.org Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 282, Canada, 1604440.4224 I sh era lint an(tlizer ovi iondig if al .corrr Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (0S1). Domains in pic-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (RDD, byte) processing, in compliance with open systems standards Byte Data, Maps (from graph baset1 systems such as face (05I), for seamless integration across industry partners and or body . bodily parts scanners images( brain, finger prints.
independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine teaming (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space im-customers, who execute product lines in a Multi layered, Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-malty to meet business goals. It can be an industry such as stoning, for building a capable . Data As A Service (DAAS) financial institutions, Insurance, Government departments, platiiirm, or to build a capable platform , to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication. Media or Auto industry. We propose (we choose restful APIs integrating platform built over MI
this solution to integrate with open system's, standard solu-8 õISR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products. to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase. Parquet, Flume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive. Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables. Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser-vice(SAAS). integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OS!) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru = resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables. Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously, in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) sign ( HLDs) . Alongside we have filed proposition to de-The. data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues, Domain Data. The ( multi domain ) source data , from Ser-that we can resolve by applying our solution, which can ben-viceefit numerous engineering domains and financial industries.
End Point Entry, where API is exposed to intemet or to open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other ,,,. ..
than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http post(). get 0 ........
...
....., mom methods submitted to utilize a service controller gateway, t--:., fti ¨
. ow to parse and to access any application, using the calling con- õ
pittOenwi,orowom ...,.."÷,* ' .' ,....
OLTP
troller or sub (nested) controller, that recognizes the browser , "... =
variables data types. mime types. The incoming data. af- ,.,j a,........,52 ......
ter passing demilitarized zone, (DMZ, B in picture) is then ---P,torre , .vol. "s, HIld OCIp tirtSt 8 Me) (II = Ouster ...
,,, = ' ,"
transformed thru a data type transformation processing layer ...
............
( C), categorized by source systems types. to convert data 1 ..
r.r.,., : ig , = ; " :"...:
in a single common format, byte data, for non Stream data ....... ^
,......1.1 types. ¨.¨.
........ i Sqoop uses a relational database driver, JDBC driver to fgrAftarlu, =-,.
:I:. Arii 1, P.N.. %
I
4...,...
, obtain these data set from source systems residing in ( C) to -- , ¨ ¨.
¨
move them into HBase HTables, via utilizing m c.¨
apreduce or ii,...õ , ......, ,v ,...¨ ..... .....:
,, IN ! :::, = .,, , wow wpm IrentfutArt sag; __Hive. Tuple data are processes same way. After processing ,.. -_,, . õ. i ..¨ 4-7,`,. z =
¨75:¨ -.. --the result set is either sent to an ETL source or to each source _ system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data, Figure I: LSSI
System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lotion using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile. and then to HBase for stor- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai. for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems naviga- fast responsive to our Partners and consumer of our prod-non. The Graph data is processes with Parquet from HTa- uct or services.
The solution is proprietary to Aerovition bles' Classifiers' Columns . graph_node, graph_relationship, Digital Inc.
Our subsequent work include LLD. and pro-graph_search_weight_algoritlim (tables ) to build, runtime in memory data (D). to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. F..¨ Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 yoke JSON, XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component. then send to Hadoop platform.
Table I: HLD Analytic Pipeline for An Ecosystem. In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat- totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platfOrrn as a service, for data Data Management.

.!
i'''' IMMO r-er Cora taws tducat<o/
',MI Rol marlin Imerow.irs AtfORNICI Airtonl on.< Cheard Ono .,. ftnneof =E .;
Recant, C' SRIBRORmain i Reference Deca frst,n=
El .1111 ' ID) in ivSernoly Computing, ottO innfn no MR
n , S = /
reale-, Le . OLTP
-- '1"""'""""---! = ! ,,tn, = w Donis , õ i ii? 0 K =
i (E)Ilive , (..., 1.,...7.-/' r In, eon! ,I-L.--:-.1 IG) ritlitk [Column Dbl Hadoop(mstance) in a Cluster 1 -LI , S.M. IV . .
. 4 Re et cran 0 .' . Z ÷' " ' trim c,,c,.... .;;.;
TwreScst,cv ;;;'; &Koh tledefsl tiod=ii: to; =
f rev kW t= 3./.111egge0 I
".s DO
' = = =
Hamm, . ail all III 0 ip ink. .
MON,510. 1 SescrEll = Metro, Id/ ;

2.01ettpe= Emend), e , ......
C, utter nrances o ....
' Govemmant =
Ecosystems, Processm laytr, Analytic Pipeline = ' 1 I
Real Eif ate ;
i - . =114..*,,e14.27%tõ, 1 0estko nal - 1 Source =
Source I
(OoSsonel = !,..v.,,., irsItInS Sy f tvn$ S,Stt ,,S En.
. , . .
¨ (A)Serrite ...i .....¨
= ' .ar, Input Layer, Restful API "Sall(C) Oreer.freeleern ' 1 * , pxIpotnts 1 . ,...m.,,,, 'µ.1eS
Megreben.S.1.Ø1.9 =
_ ot ..nre r oreiso= 1,0,330, ferhaiert Aare/411m .rOceStte NOV in r, ---_-__ _ eareleeed 102 Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 4 Model Driven System of Systems For Neuro Computa-tional Medicine. Drug Pathways Clinical Trials. AisXDS17 (Spe-cific Context Byte, Map. Graph, Stream) =
Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Graph, Stream) [English]
Petitioner: Shaw Rahman' lAerovition Digital Inc. www.a3ic.org Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 282. Canada, T 604.440.4224 s hatcrah oto,n4aer iondi9ital .com Abstract ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (0S1). Domains in pie-The following article demonstrates a solution proposed to ture (I), are the sources of data, in varied forms, such as, design highly resilient distributed streaming data (RDD, byte) processing, in compliance with open systems standards Byte Data, Maps (from graph based systems such as face (OSI), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for aerospace product design, space research data for space im-customers, who execute product lines in a Multi layered. Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transforination and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data . telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous necting multi -disciplinary business domains, to provision.
engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sinning, for building a capable , Data As A Service (DAAS) financial institutions, Insurance, Government departments, platform, or to build a capable platform, to interact between , Aerospace Digital product developers. Semi-conductors, domains via a set of common business logics or processes Telecommunication. Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's. standard solu-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief front large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms . and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose 1-ladoop, and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark, Zookeeper, Sqoop, Pig, faster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables, Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser- Classifier Processing.
vice (SAAS), integrated in, that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In-The platform provides parallel in memory data computation.
formation Technology or Systems Engineering, for faster processing classifier objects such as Map, Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables. Platform As A Ser-cache, to build runtime resilient distributed data (RDD) and vice (PAAS), Data As A Service (DAAS) platform, Soft-interact with HBase or HIables. YARN enables SPARK to =

run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase. which stores classifier data types or We have completed System Architecture High Level De-graph data elements (C).
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues.
Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to intemet or to elit numerous engineering domains and financial industries, open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input, or bulk catalog -data, or applications variables posted via http post(), get () ¨.
¨ mos methods submitted to utilize a service controller gateway. C
'' : OW
I, to parse and to access any application, using the calling con- ¨
¨"" ,, ...,... , OM
troller or sub (nested) controller, that recognizes the browser variables data types, mime types. The incoming data, af- , pp,....a.:-.... <,....x. Hadooplinste nce) in a Clutter :
ter passing demilitarized zone. (DMZ, B in picture) is then ' transformed thru a data type transformation processing layer L....A.-,7......atj , !.' r -7 ' =,...:,:.-7: =;!...!;,;.7-( C), categorized by source systems types, to convert data , .... .. .
in a single common format, byte data, for non Stream data ....... -, --.*--_...
....
types.
...¨
Sqoop uses a relational database driver, JDBC driver to M..
obtain these data set from source systems residing in ( C) to , move them into HBase HTables, via utilizing mapreduce or . := r,., ;,-.-- 1 ~,^~ "."-Hive. Tuple data are processes same way. After processing .., , worn,. IMØ1 .., ry,......
i ' , , _.,''' J
======== =-*.- :Z.:r the result set is either sent to an ETI. source or to each source OM ¨
system(s) (C). Relational database data is manipulated via Hive (13) query processing services, then the processed data, Figure 1: LSS1 System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lotion using Open systems Ecosystems operation front Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor- Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring . Map data is nesses willing to transform to Ai. for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and be bio medical devices or from Graph based systems navigit-fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from HTa- uct or services. The solution is proprietary to Aerovition tiles' Classifiers' Columns, graph_node, graph_relationship, Digital Inc. Our subsequent work include LLD, and pro-graph_search_weight_algorithm (tables ) to build. runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Threel A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 0 yoke JSON. XML or. convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
Table I: HLD Analytic Pipeline for An Ecosystem. In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

, UM) tehoCatoOn/
- C,4.to ttiareel treed StO,CIMS.10 V....I.... A,VOSIHICe .I.W.IIIDVVe Pomo./ RitS082/1 Soiesiemaeare Deference Dots 914,01 SAP, lietrumi .
ear,eme ''' I DW
' (0)1DINANDOSD COPODUittiliitm.r ,, f.t_,. S I =

1 =
lamonl.al -'f- OLTP ;
1 , A l reir ,:e = e = . - ..,-i "4/ ,, D .= ' K i (E)Hive = .
, feCoeSag 1 .
IG) lituble (Column Cibp Hadoop(instance) in a Cluster L- 'Atj Socisi j ---- ' :-.i., Dees-ors1.¨ - -^, ,,,,..,s,,,... ,k TIllet i,tfll",11 s.---1 ' GT aot, num,:
Hsu* t Los : S'm Pelt.:

HeIlltaill. MI r4N lei IN IN NI 1111 ra4;1.. U..' , HOWE, , Zoosetper ilsembls , Cluster Meances G01.1.rnms.
Ecosystems, Processin : la er Ana tic Pipeline, . ! , 4.,...,=W " =
*eel Efate , I i=vo:4'.
nest6.0 API " , Sou, ce Source COII.V.Ifte , su. our ",, ,., '''': ' svStP,. Systerm SySNMS
' .1 , '' ETL
Input Layer, Restful API I Call(C) 00=.911.11011 .,1,)141Polnis Int new.. Drewry tM
!SON RL, wt. toiervies Poseurs s,ocesso/ Salat.10.014004IIII
it,ctS3.17 0411511114 t4.0101. UN
Following products at Aerovition Digital apply the above design. Specific Product Line, Context:
Line 5 Accelerator Pliyics and Big Data Mining for Com-plex Analytics. AisXA17 (specific context. Graph. Stream) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Map, Graph, Stream) [English]
Petitioner: Shaw Rahinanl lAerovition Digital Inc. www.a3ic.org 1Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V6B 2B2, Canada, T 604.440.4224 shawroh man4aerovi1 iond .com Abstract.
ware As A Service (SAAS), from the same platform, that is built on Opens Systems Standards (OS!). Domains in pie-The following article demonstrates a solution proposed to tare (I), are the sources of data, in varied forms, such as, design highly resilieni distributed streaming data (RIX), Byte Data. Maps (from graph based systems such as face byte processing. in compliance with open systems standards tOSD. for seamless integration across industry partners and or body . bodily parts scanners images( brain, finger prints.
independent subscribers. The solution is maintained under bio medical devices, neural analyties instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery : electron microscope, satellite data, to resolved complexity, in machine learning (MU and ap-bill of material (BOW used on automotive, air vehicle, or plied artificial intelligence (Ai) alg.orithms and theory. Mr aerospace product design, space research data for space im-customers. who execute product lines in a Multi layered. Very Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data , telecommunication or wireless mobile device data. or governance data for compliance between regulatory Introduction Target Audience agency or government partners . These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream (taut provi-malty to meet business goals. It can be an industry such as stoning. !Or building a capable , Data As A Service MAAS) financial institutions. Insurance, Government departments, platform, or to build a capable platform , to interact between , Aerospace Digital product developers, Semi-conductors, domains via a set of common business logics or processes Telecommunication. Media or Auto industry, We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard solu-8 , ISR 311, compliant JAX-RS interface). thin Software As lion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities, control business, for simpler management and to minimize We propose Hadoop. and Apache software components:
operational costs, training cost and to increase revenue via FIBase, Parquet, Flume, Spark, Zookeeper. Sqoop. Pig.
taster execution. We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems, useful for more than half a dozen complex data complex data types into byte data. For Operational systems, types and stream transformation.
we propose Linux Cent OS .
Present Research Challenges Description: Common Platform Solution's Building a multi-channel . multi layered . architecture for Internal Cross Systems Data Flow For a platiOrm that enables. Platform As A Service (NAAS).
Data As A Service (DAAS) platform. Software As A Ser- Classifier Processing.
vice (SAAS), integrated in. that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In- The platform provides parallel in memory data computation.
formation Technology or Systems Engineering. for faster processing classifier objects such as Map, Tuple or Hata besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the plattOrm enables, Platform As A Set--cache, to build runtime resilient distributed data (KM) and vice (PA AS ). Data As A Service (DAAS ) platform. Soft-interact with HBase or HTables. YARN enables SPARK to run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of liBase, which stores classifier data types or We have completed System Architecture High Level De graph data elements (G ) . sign ( IILDs) , Alongside we have filed proposition to de-velop products to solve multiple industry complex issues.
The data from (H) domains, are in general referenced as Domain Data. The ( multi domain I source data , from Set-that we can resolve by applying our solution, which can hen-vice End Point Entry. where API is exposed to inferno or to open public domain, (A). passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input or hulk catalog data, or applications variables posted via http post . get 0 Mee.
a.
ft.., Mai methods submitted to utilize a service controller gateway, = =
to parse and to access any application, using the calling con- 7.; , :
JO . i -..¨
OLTP !
troller or sub (nested) controller, that recognizes the browser ,... , t =*. , .
. .... .
. .
v.triables data types, mime types. The incoming data, af- .
, . *4 4,.... ni in e Cluster ,-y---- , .. . ===
ter passing demiliotrized zone, (DMZ, B in picture) is then ..
. HedooMinstence).. 1.. . . , .
. .
= ..
transformed thm a data type transformation processing layer _.41.-1191111191.11811. ...-7..-- *. Z."-: '7:-,=:¨ -.- = : 1 :
I C 1, categorized by source systems types to convert data . .. i ...,...=== ifit : , t q y c .= .:: = -7 .=
. ..
...., = = = in a single common format, byte (Int, for non Stretun data IT R :
= . ..
..:..:
types. ........ -Sqoop uses a relational database driver, JDBC driver to eop obtain these data set from source systems residing in ( C N=fl ) to f i' = ' ,:::., .........õ ¨
move them into HBase HTables, via utilizing nmpreduce or . = , .--.-,:=-- t P.t4r..- . = ETt RnIttd 0 ,,, It ' 0 , ,,, ,,, P1 , ON 1 o.........
Hive. Tuple data are processes same way. After processing - ! _,, -,21 ..=:',".;., -:=''.. rx. ---",..-the result result set is either sent to an EIL source or to each source ..........
system(s) (C). Relational database data is manipulated via Hive (F.) query processing services, then the processed data, Figure I: LSSI System Of Systems. Domain Integration. So-HCatalog is sent to Pig if'. and afterwards thru mapreduce lotion using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories. as HFile. and then to HBase for stor-Our Target Audience are Fortune 100 companies and busi-age in HTables via zookeeper data monitoring . Map data is tresses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition.
believe the solution will stream line product lines and be hio medical devices or front Graph based systems naviga-fast responsive to our Partners and consumer of our prod-non. The Graph data is processes with Parquet front liTa-uct or services. The solution is proprietary to Aerovition hies' Classifiers' Columns . graph .node. graphfelationship, Digital Inc. Our subsequent work include LLD, and pm-graph_search_weight_algorithm (tables ) to build. runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects.
Stage One Stage Two Stage. Threel A Customer can maintain OLTP or Relational data ware- HLD 'cis - 0 house simultaneously using .IDBC drivers to interact with LLD 0 0 Hadoop platform either via ET', systems that can send data Prototype 0 to Sqoop or via API interfaces to a search engine that can in-Performance 'rest 0 0 yoke ISDN, X ML or. convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data. thru convertion to inte-grate with XML or Streaming via platform component. then send to Hadoop platform.
Table 1: HLD Analytic Pipeline for An Ecosystem, In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows, Thus a common platform interface will re-duce operational cost and consolidate processes in one plat-totype development. performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optitniie our services for Big numerous verticals, using this platform as a service, for data Data Management, fu AV) 0,1.= L=anva 7,==== `. -... `1=Pt =- f===-=,. a .
AerCKPaCt ALaranD? tr. flucattni Chialita DNS
411=100 Saluiratc=== litference e a. Data , tv Pont.aaj .
DW
(D) In Memory COPlip IA rig, POO torrnarlat. t s ' IN =
¨ , OLTP
1%. A ...p.c., = .
1 a II

o - ( E)H we .
. , ..(...... , IG) motile (Colu=nr. Db) Hadoop(Instance) in a Cluster 1 rocio , LNeSooe . __ re w .nr, o rtk r r4 P I t I , , =4 7 ,..."-. k IrrtaaS, IlDtittli. t - j lieol lawns Saw PeUer C.
p I =-=-.= I
""ilaft Sr a. a 1111 11/ a 11114( 4- -. 1011,S10.
01040111 difirdal.**
Zootlemer Insembq , Cluster Instances *NNW
111.....10, ...,.._ GCMITAIIIM _ WI
Ape.
, Ecosystems, Processl er Ana tic PipeGne.- --- I
I
' ! , , IlestioAPI 6" , source Source ,,,e,,.
Consurnw Systers St, punt ""¨'' Eli.
- -, ':Oisitivice '''' .=:.=;',".==="===;:='' = = = .===¨=,-.- ,.
input I,ey<' er,nestful4P1 ,ICall(C)XMLi -,.: .,..:-. --:=....- .., ::vs . eftran..toomal 4S0T1 KW Duiftteletesol clelmter _ +.,- =-= x-..,.. ) Practise, trecesia Pncesur iskosionAorsolik=
.....,, ...,,,,,, , , s, 4,, ,y, PtittS30, Ilyinflas.
carol& me Folios.. ing products at Aermition Digital apply the above design. Specific Product Line. Context:
Line 6 Aimmoinous Commercial Airplane. Space or Space Vehicle System Integration. AisXSAI7 (specific context. Graph.
Stream. Map) Patent Context: Common Platform Architecture - An Alternative Solution For Streaming Data Integration In A System Of Systems, Multi-layered Business Transformation (Specific Context: Graph, Stream) [English]
Petitioner: Shaw Rahman".
1 Aerovition Digital Inc. www.a3ic.org 1Aerovition Institute(s) Applied Research And Development Department 2915 360 Robson Street, Vancouver BC V68 2B2, Canada, T 604.440,4224 ',ch(rwrah?r1onct(wrotyitioi1dqita1.eo,n Abstract ware As A Service (SAAS). from the same platform, that is built on Opens Systems Standards (051). Domains in pie-The following article demonstrates a solution proposed to ture (1), are the sources of data, in varied forms, such as.
design highly resilient distributed streaming data (ROD.
Byte Data, Maps (from graph based systems such as face byte) processing, in compliance with open systems standards (OS1), for seamless integration across industry partners and or body , bodily parts scanners images( brain, finger prints, independent subscribers. The solution is maintained under bio medical devices, neural analytics instruments used in ownership of Aerovition Digital Inc. as a standard approach drug pathway discovery: electron microscope, satellite data, to resolved complexity, in machine learning (ML) and ap-bill of material (BOM) used on automotive, air vehicle, or plied artificial intelligence (Ai) algorithms and theory, for customers, who execute product lines in a Multi layered. Very aerospace product design, space research data for space im-Large Scale System of Systems structured business domain ages sent from space objects or instruments, service data for Transformation and Integration with advanced technol-from patient health care, clinical trials, entertainment in-ogy.
dustry data telecommunication or wireless mobile device data, or governance data for compliance between regulatory Introduction Target Audience agency or government partners. These data are categorized in multi classifiers, and can form a neural network, intercon-A Multi layered Complex business domain with numerous fleeting multi -disciplinary business domains, to provision, engineering or business verticals needs to integrate opti-thru a common system for digital and stream data provi-mally to meet business goals. It can be an industry such as sioning, for building a capable, Data As A Service (DAAS) financial institutions, Insurance. Government departments, platform, or to build a capable platform, to interact between , Aerospace Digital product developers. Semi-conductors, domains via a set of common business logics or processes Telecommunication, Media or Auto industry. We propose (we choose restful APIs integrating platform built over JEE
this solution to integrate with open system's, standard saki-8 , JSR 311, compliant JAX-RS interface), thru Software As tion, using it as a single method and seek relief from large A Service (SAAS), from the same Platform Service (PAAS) vendors numerous platforms , and supply chain products, to that supports both abilities.
control business, for simpler management and to minimize We propose Hadoop. and Apache software components:
operational costs, training cost and to increase revenue via HBase, Parquet, Flume, Spark. Zookeeper, Sqoop. Pig, faster execution, We propose a solution based on Hadoop Hive, Kafka for stream management. The solution converts ecosystems. useful for more than half a dozen complex data we data types into byte data. For Operational systems, types and stream transformation, we propose Linux Cent OS , Present Research Challenges Description: Common Platform Solution's Building a multi-channel . multi layered , architecture for Internal Cross Systems Data Flow For a platform that enables. Platform As A Service (PAAS), Data As A Service (DAAS) platform, Software As A Ser-vice (SAAS), integrated in. that is built on Open systems Cross Systems Data Flow Across Platform's Components:
Integration Standards (OSI) is an ongoing challenge in In- The platform provides parallel in memory data computation, formation Technology or Systems Engineering. for faster processing classifier objects such as Map. Tuple or Bag besides byte data thru , resource SPARK and its execu-Patent Sought: Why this is an invention tor services to process in each executor's processs memory The architecture of the platform enables, Platform As A Set-cache, to build runtime resilient distributed data (ROD) and vice (PAAS). Data As A Service (DAAS) platform, Soft-interact with ['Base or iiTables. YARN enables SPARK to , run on Hadoop platform as a resource to perform parallel, and software simultaneously.
in memory execution of data processing tasks (job) using HTable data of HBase, which stores classifier data types or We have completed System Architecture High Level De-graph data elements (G) .
sign ( HLDs) . Alongside we have filed proposition to de-The data from (H) domains, are in general referenced as velop products to solve multiple industry complex issues.
Domain Data. The ( multi domain ) source data, from Ser-that we can resolve by applying our solution, which can ben-vice End Point Entry, where API is exposed to intemet or to efit numerous engineering domains and financial industries.
open public domain, (A), passes thru a common authentica-tion and authorization security layer that restricts data other than variable data entered by user as input, or bulk catalog data, or applications' variables posted via http post(), get () .-1......, 01:1111 methods submitted to utilize a service controller gateway, .7,-..:1.-to parse and to access any application, using the calling con- _._ 4,,,i.--, i.,...4.õ... . . s=1;
troller or sub (nested) controller, that recognizes the browser , ; . a nlive variables data types, mime types. The incoming data. af- - :, doo tee) C -, -_...õ ep(insertine lutter r,-- = '"' ter passing demilitarized zone, (DMZ, B in picture) is then -ti ,-.., :, transformed thru a data type transformation processing layer I= :::. , ,, ....... õ-õ
( C), categorized by source systems types, to convert data ..
,...,,, in a single common format byte data for non Stream data -.. ...... I
types. ...,-, , .

Sqoop uses a relational database driver, JDBC driver to firtetuuere P =
,P= r PtOPtirw- r Iwo*.
l obtain these data set from source systems residing in ( C) to ....-move them into HBase HTables. via utilizing mapreduce or -- rn.
:, ...xi -77-:*<:..;,:...4,,, . ,,,o, key' er.ltesttoi.ePl Cart(0 ..,,,,...
Hive. Tuple data are processes same way. After processing -the result set is either sent to an ETL source or to each source ...MO
system(s) (C). Relational database data is manipulated via Hive (E) query processing services, then the processed data. Figure I: LSSI
System Of Systems, Domain Integration, So-HCatalog is sent to Pig (F), and afterwards thru mapreduce lution using Open systems Ecosystems operation from Pig to Hadoop instances HDFS or Htables.
Steam data is categorized and ingested for processing via kafa Stream controller, and sent to Hadoop instance , first to HDFS directories, as HFile, and then to HBase for stor- Our Target Audience are Fortune 1(X) companies and busi-age in HTables via zookeeper data monitoring. Map data is nesses willing to transform to Ai, for a simpler solution. We passed similar way, from sources such as face recognition, believe the solution will stream line product lines and he bio medical devices or from Graph based systems naviga- fast responsive to our Partners and consumer of our prod-tion. The Graph data is processes with Parquet from IITa- uct or services.
The solution is proprietary to Aerovition bles' Classifiers' Columns, graph_node, graph_relationship, Digital Inc.
Our subsequent work include LLD, and pro-graph_search_weight_algorithm (tables ) to build, runtime in memory data (D), to constructed aggregated in memory ob-jects, and or to manipulate data classifying objects. Stage One Stage Two Stage Three A Customer can maintain OLTP or Relational data ware- HLD Yes 0 house simultaneously using JDBC drivers to interact with LLD 0 0 Hadoop platform either via ETL systems that can send data Prototype (>

to Sqoop or via API interfaces to a search engine that can in- Performance Test 0 t) yoke JSON. XML or, convert incoming multi layered nested Big Data Ai Application 0 0 aggregated data such as catalog data, thru convertion to inte-grate with XML or Streaming via platform component, then send to Hadoop platform.
'fable 1: HID Analytic Pipeline for An Ecosystem. In A
Search engines can invoke data residing in Hadoop tables System of Systems.
of HBase invoking APIs utilizing these business data pro-cessing flows. Thus a common platform interface will re-duce operational cost and consolidate processes in one plat- totype development, performance testing and applying Ap-form for Large enterprise that interchanges data globally in plied Artificial Algorithms to optimize our services for Big numerous verticals, using this platform as a service, for data Data Management.

LOW !Outlaw,/
.. 0.....u...1 r.... !,;;;; = =ri .= in, 1 an -An, Aeroso.ce Automa,,.
Fem.& Antarth S=larloecattre Ifettrance Met.1 SAP, hateen/ MEV Ono ..
DW I
mot. m - - PrA
;;; ";
(I)) In Mornay CompuUng, $,CIDgerwralsei- r4*, , , , i S
A , .......õ., OLTP
I R =
A .
= =4. al) 4.= I , ;
tanks ' m= 0 ..
' n (E)Hive , ,..-õ,j IG) titatli iCe/amin OW ___ Hadoop(instance) in a Cluster, Sec.', õ wa......-......:. : ; -=,,,,'"--------' I
...1 j ...... .....- 1 _ I F, 1 , , w . , , , . on 4 Mt i In cr.:Pnr in ¨ TA=get Scrum's I
., A, Ii 1 4 10 ............w....,' 1 .
' . . : ¨.."- - .1501,0 4 ITI.I anu,a, APtertmemm Cre Paweparl.. OM *mem* halwatlerdClaam#Yers 1141=Anactl, I.. - ''' I,'" ' :. ' . : 1 tre ,.
Pu.... . PWILIEO, I ' liostaKti . affect fir InnKeepe, Fmen-b C uste , ntt or, CI L-ntta , GOVIMMleat .
= =
Ecosystems, Processin La er Analytic Pipeline = ,,,. , i T

Aenquede Restt.t AP Source 5,0UrC1 Sew ce r Cont.,. , ''''';Af., = , =
... a ,,,,A, = ,,,kri,,, Tvatrars 5.tems 5,,sitms .
ETL
Input Layer, Restful API ',Call(C) o.... II..., - L.. Endpoints : s,s InVorRolosi.
W.F.' '.:... i e AYIN:',0- ,` .1 , .50 rl ROI, Do. htdented CIV "II!
L_......., .. ., ,,..,.....4,4,0,..õ*, i o: es, C=o:esaa Par. e sto, P'oinso, SeillitftelliNeWit.
143111 Ina.
¨ Coonmit. nda Following products at Acrovition Digital apply the above design. Specific Product Line, Context:
Line 7 Bio Medical Product Design. Clinical Research, An-alytics, AisXCR07 (specific context. Graph. Stream)