GB2203617A - Embedded test system for communications systems conformance testing - Google Patents

Description

2MBEWM ITS7 SYSTEM PM M19.MCAnCM SYSTEMS C2n= WSTM 2203617 Backcmound and

&zwnarym of the Invention 1he pret Invention relates generally to the tasting of Coomnications watem software. H= Opecificallyr the tim ten to a Wethod of detanair.IM whether & ccm]n4Catlang cc

Conform to a specification of ccm=nicati pr=1% &g the

MP specification$ based an the "n Systems 1 established by the

Inte=ational 0zation for Standardization (Iso).

manufacturiM prooemses today we ccintrolled by retworked canx.er systems which t cmcate with em another.

In the Interest of standardization, the autmotive Andh, trv,, and other mwmfacd=irbg industries, haw prop a specification of cxrmmtcaticrLq Protecols which are to intercmmct carqp system and robotic systerm t o e=unicate with one another. ow gwh specification is the MAP specification.' The MAP specification is based on a =del of ccraputer 03Caticc established by the 1.90 called the Open Systems 1. Cle Open Systems 1 (and the MAP specification) divides the C=='n4r-at-iens lem into me sets of related isEue called leverge which am disc below.

7b ensure =z"tibility with a manufacturer's ociatinq =1;xltP-r and x^cmation eqt, liers and vend= of new CRUi Mut lop COWatible. cxx=m4cations software. 1he nw OWt mist = to the MP specification.

Che MAP specification and the Open Systems 1 upon which it in based are aratively c=plex n4ca prcyu=la,, Testing a rlew 03=nica Program or a r pince of cocating hardware to dmim if It em with the mAP specification is not

2_ As explained more fully below# In mter to test all 18 of a c for h-PtF-,r 110M It has haf=e b rwxmsary for the new product developer to write Include special tent progr. directly into, the cccations software. MUs aws =qplexitr and expense to the software and can leave the c=mnimtiom systern with extra software age that In riot needed cecm.- c=mance has been deter. Possibly m mor troublesome Is the fact that tent routines have heretofore been closely linked to a particular 1 within the MAP specification. This close

1 has lt very difficult to Inte certain fumw between adjacent layers without the tent routines getting in the way.

lhe present Invention overcomes this problem by a tectv of ram an of upper layer flow control nochani ws In order to manage lowier layer masage transmission and reception for omf=wx" testing purposes. The invention allow the new product to be tested without the need to include acme of the previously required special test functims in the new product software. 1he Invention permits =aroe testing to be effected without Interfering between adjacent layers of the software being tested. M=, the tim allows the emmud=tims software to lay or interrelated adjacent layers,, such as between the network and trwupcxr t layers in order to inrove software performance.

mor specifically, the Invention permits the protr=l cakfarmance testJxbg to be perfanned on tM network layer of the protocol erren under tent (IUT) 0 that in, the Antenwt protocol entity In the smnw under test,, without alte or &Mending test code to the netwmkltranap=t Interface (1xetween the network. wd transport of the product. We Invention 2 - ndirectly inanages me flow betwem the networkitrwuqxxct Interface of the product In mtkr to detc=ine whether the netw=k layer of the Iriplementation under test Is passing data to and accepting data from the trrt 1 correctly. ltion perf=m the network layer prot=l testing In a ronive lon.

W Invention ow=ls the flow of Intion through the transport ,, to thereby Indirectly control the ratwork ce, interface th the transport# by Indirectly taking control amr the =edit and acledt miqmle ratically and periodically qenerated by the trmuw;xwt 1.

Ftw a wore emplete urderstanding of the Invention,, Its cbjects and advantag",, reference way be had to the follo specification and to the acompanying drawings.

Brief Description of the Drawings

Figure 1 is a block diagram depicting the seven layered conainica±i=s system protocol; Figure 2 is a stic block dim illustrating the ical cmmmication topology iroplemented by the prwently preferred commicatjj" "tom; Figure 3 is a block dia Illustrating an Intenradiate r or router for transferring mus between.nonadjacent rc; Figure 4 Is a me sequence dia Illustrating a flow of wessage transmissions and lwl between initiator wd recipient; P 5 Is a diagram illustrating the wl prle used by the transport layer of the cx=rlm w.v to te the flow of W. 0 1 3 V,-, %-W. - Figure 6 Is a block diagram of the tent program of the invention, illustrating the mannar in which an cplary systern progrun is tested; Figure 7 is a diagram illustrating the form of protocol data units at each of the even levels; and Figure 8 (Ba and 8b) Is a flow dwwt depicting the testing methed of the invention.

Description of the Preferred Crbnt before a detailed explanation of the invention is given, a brief overview of the MAP specification will be pented. W MP ification is des= in, General Motors, Manufacturinq Autnmation Protocol Version 2.1,,' General Motots rlmr;,rcation 1985.

The MAP specification is based mpon the Open Systems Interecramiction (OST) 1 established by the ISO. A detailed explanation of the OST 1 may be found in 150 IS7498, "Information Systems P=cessing Open Syst Inte=onnection - Basic]Reference 1,1 1984 which in incorporated herein by reference.

Figure 1 diwmatically illustrates the seven Uyers into which the MAP communications p=tocol In divided. At the lowermost or most primitive level is the first level,, the Physical 1. Cw Physical Layer is responsible for encoding and physically trtting messages between two, adjacent modes an the net. MhP atw uses a bus topology Picturcid in Figure 2. The Physical layer IWts the M= 802.4 standard. 0==1cation Is mw a coaxial cable 10 which provides two cot- cat channels. One channel Is a low frequency dumel which propaqates in a first dire and the c channel is a high frequency dmwl which pates in the 4 opposite direction. A head end unit 12 terminates cme end of the --,ble for shifting the low frequency camunicatirXIS up to the high frequency to therebv change the propaqation direction. The communicating units 14 place ommmications an coaxial cable 10 at the low frequency where they propagate to the head end unit 12 for retransmission at the higher frequency. All carrmz%icatinq units vwpive data at the high frequencv.

Returning to Figure 2,, imadiately above the Physical Lwjer is lavel two, the Data Link Layer. The Data Link Laver is responsible for framing the binarv data handled by the MRical laver into octets (eight bits). A plurality of octets are qroupd into data packets hv the Data Link Layer. 41he Data Link Layer connects a link address to each packet which identifies which cmmmicating unit 14 or node is to re"eive the packet of data. 7he Data Link Layer is respmsJble only for addressing data intended for'physically connected devices on the coaxial cable. In order to send a message to a nonadiacent node (one which is not physically connected to the coaxial cable) an intermediate ovsterm unit or router is emploved.

The third level, or Network Lwrw,, handles the routing of information through intermediate sYstems or routers. 7he Network Laver establishes the protocol for transferring messages from a node on cable to a node on a physically separate cable. Figure 3 depicts an intermediate system or router 16 which in connec a first camnunication system 18, which cam&z-4cates over a coaxial cable 10, with a second canmunication system 20, which cam=icates over a fiber optic cable 22. 7he system or router 16 aqploys two Physical Layers 24 and 26 which are ccupatible with the coaxial cable and fiber optic prot ls, respectively. The Data Link 6 yers 28. and 30 res to the Physical layers 24 and 26, reively, and in turn canTmmicate with the Network laviax 32 of the router. In effect, the Network laver 32 of the zmter can be considered as a bridge between the respective Network Layers of the first and second conrunications avstems 18 and 20. In this fashion, it is possible to establish conmmications between tbK> rdjacent nodes "loying similar or dissimilar phvsical protx=1s.

In theory, the Network layer can determine the recipient node of a given comnmication. by a variety of different schemes. ".

scheme is a connection oriented protocol in which a permanent virtual pipe is created between two devices. Using such a protocol,, it is not necessary to place a signature on the ressage being =w, nor ia it necessary to place anv routing information on the weasaqe, as the pemxnanent vil pipe automatically describes the idientitv of both the initiator and the recipient. Arr protocol is the connectionless oriented protocol. The connectionless oriented protocol requires that a message be enclosed in or a=ned by an add ss and a signature identifying the desired recipient of the wessaqe and also identifying the initiator of the imsEaqe. 1he connectionless oriented protocol, unlike the connection oriented prot 1, does not guaranty that masages are conve.M in anv particular order. 7he MAP specification presently adopts the

Ln"ctimless prot 1. Ibis comectionl ss protocol is owrnonly known as the Internet or IP protocol.

use the IP protocol does not quaranty that wessages are received in the order in which they are mmt,, same picn nust he "a to ensure reliable delivery and servicing of messages. Referring back to Figure 1. the fourth level or Transport Layer provides this 6 - ssurance that mssages am reliably delivered. The W specification uses a positive acknowledgment technique In which the recipient of a nessage acknowledges receipt by returning an acJmwledge loignal (ACK) to the initiator. The details of this acknowledgment system are ocplained below.

The Transport Layer also handles ineaaaqe flow control. Flow control is the process used to ensure that the pace of cmymudcations is apiate for the two commimicating devices. For cxumle, if cocation is to be established between a high a minframe cter mid a low speed pers computer,, certain provisions t be e to establiah a working communication rate. This Is done by a pi.mr.ess of rxxjotiat-ton between the two cammunicatinq anits, and specifically between the TrarL"rt layers of the two cxxLa,Lmicating units. The two ommunicating ervatem aqree In advwice of omminication on the mx ressage (or Protocol Data Unit. PDU) size. As the exchange of wessages proceeds, each communicating unit can send control parameters to control the rate of 1M trssion.

The hanism, bv which the trssicn rate is controlled is based an the PM sequence number. A unique sequence Identification numtier In attached to ear-.h PM,, so that the sequence, numbers can be used to place the packets in the proper order ld they become Mien communication is established, each cormunicating unit cs to the other the sequence number =r spending to the ma nuimber of PWs which it is prq to receive. Cw ex of data is full duplex, either side can send, Pme; to the other up to the maxbm sequence number given to it by the o.

As a le ele, if both units am Initiating communication and unit A is prd to receive three PMs, unit A 7 -. 1 2 Ad send the number 3 to unit B. Thereaftert if unit A wishes three more PDus to be sent, it would update the sequence number to 6. Unit B could, at the same time, establish its own desired rate of communication with unit A. Unit B could initially inform unit A that it is prepared to accept PDUs up to and includinq sequence mmter 10.

After these. ten are received, unit B could request arxYwx ten, by updating the sequence number to 20,, or it could slow the rate down or speed the rate up by selecting either a lower or a hiqher sequersee number. The initial sequence numbers are established during a nnection establishment phase of cam=%ication. Thereafter, the sequence numbers are updated as part of the message received acknowledgment (ACK) signal which is relayed from recipient to initiator.

To guard against system hang up which would occur if both communicating units were waiting on the other to acknowledge, the Transport Laver entities in each unit are designed or required to routinely send redundant acknowledWent signals which convev the current allowable data messages which can be received. This allowance information is referred to as a window.

For a more complete understanding of the positive acknowledgment system inpleimented by MP. refer to Figure 4 which illustrates an exemplary cmmunication between an initiator and a recipient. It should be kept in mind that MAP Implements a fall duplex communication svstem: both cam, inicating units can be initiators and recipients simultaneously. Referring now to Figure 4', two columns are depicted, designated as Initiator and recipient. in these columns are a sequence of message transmissions with arrows indicating the direction of transmission. Commmication begins with a 8 - xction establishment sequence. The connection catablist.

portion of the communication is indicated generally at 40. lone sequence begins as at 42 with the initiator @ending a =rawbction request (CR) to the recipient. The recipient responds by transmitting a connection confirmation xwksage (M). The initiator. then acknowledges that it received the connection omf irmation by mendine.r an adahowledgment (ACK). During this COnneCtiM CEtAblint Se, both initiator and recipient specify the mi and maxi secpjence number which it is prepared to receive (i.e.#, the wi). For purposes of illustration,, only communications initiated by first unit and received by a second unit are illustrated. it will be understood that a similar diagram would result if the e=unications initiated by the second unit for reception by the first unit were to he illustrated.

When the connection is first established, communication is assumed to begin with sequence number 0 (the first message of transport user information). The hiqhest permissible sequence number established by the recipient,, together with the initial sequence rp 0 establishes the number of messages which can be sent to the recipient. This = may be conveniently repr sented as a window illustrated in Figure 5 bounded between the last seiquence request (in this case the initial sequence 0) and the hiqhest permissible sequence number rsted. If& for example, the recipient wishes to limit the reception to 100 mssages,, then the lower edgre 44 of the window would initially rest at sequence rr 0 while the upper edqe 46 would be at sequence 100.

Returning to Figure 4, ence the connection establishment 40 has cocurredo the initiator commences data trasion PWs (M).

to r purposes of illustration, it will be asoumed that the recipient has requested three messages (PDUs) to be sent. Accordingly,, the initiator transmits the first three messages rresponding to sequence numbers 0,, 1 and 2. This is indicated qenerally at 48. W= the recipient receives the third sequence, it sends an acknowledgrtent, signal (ACK) back to the initiator. This is indicated qenerallv at so. The acknowledgment siqnal contains the sequence number of the highest sFquence received, in this case message 2. 7t= the initiator knows that the recipient has received all three pwkets. If desired, the recipient could, with its acknowledgment, change its window size by requesting an additional group of packets different fxm the number originally requested. Flor example, the recipient may wish to allow the transmission of up to ten messages. In this case the recipient would revise its highest permissible sequence number to 13. if the recipient does not revise the highest permissible sequence number, then it remains at the earlier selected value,, 3, and no further sequences can be trara;mitted.

Assuming that more sequences have been requested, the initiator would continue to send messages, and would expect to receive acknowledgments thereafter. This is shown for messages 3 and 4 as at 52. It is, of course, possible that a particular message inav riot be zvoeived in the pr r sequence. A qiven packet can become temporarilv delayed or lost during transmission. At 54,, the initiator sends messages 5 and 6. The recipient..es only message nxrher 5 to illustrate the case in which message number 6 is riot sequentially received. meanwhile, the initiator continues to send messages 7 and S. as at 56. The recipient, however, cannot - :)cledge receipt of message numbers 6, 7 and 8, since it has mt __1 yet adcncwledLged message number 5.

In aco=dance with the zard operation of the Transport layer protocol# the recipient is required to periodically send an adledgment reflecting the last sequence number which it haF; received. In this case, the recipient has received message number 5.

Accordingly, at step 59 the recipient sends a periodic ad=wledcj,m t of receipt of message number 5. The Initiator determines frm thir, adlt that the recipient has not yet received message number 6 (which was sent earlier) at 54. Accordingly. the initiator resends message number 6,, at 60 and the recipient then responds with -an adaw:ywledgrnent of mssacze number 8. In this example, it is presumed that only message rr 6 was not received by the recipient and that the message numbers 7 and 8 were received and were stored pendinq receipt of message number 5.

It is important in understandinitl the invention to reize that the ow def. by the upper and lr edges Illustrated in Figure 5 can chanqe In size with each wJledt. The lawer window edqe 44 represents the last sequence nu adcrxywledged and the upper window edge represents the mn sequence number permitted. 1he present invention ia primarily corbc with cmtrollinq da" flm between the level three Network Layer entities thr xL" of level fcx= Transport Layer zwxutnisms. For o"leteness, a brief description of the level five, six and seven layers will be presented.

The level five or Session layer illustrated In Figure 1 is primarily imlved in rpwhrmizing the ownunicating units to take turns In carrying out a communicaticin. The. Session Layer is roughly VOL _' Alogous to aentence punctuation, which breaks one cxx, lete thought, allowing another message, possibly from the other unit to reply.

The sixth level or Presentation Layer is responsible for handling the data representation between cormunicating units.

C=Lmicating units may not ccrnmunicate usinq similar enicaded languages or similar file types. The Presentation Layer provides an agreed upon cmTnon language,, encoding or file type by which the two inoerrpatible Systat my commmicate. The Prepmentaticn Layer can also perform data encoding and decoding for data encryption and the like.

Although the OSI Model permits cxx,g, micatinq systems to negotiate protocol at this level, the presently Inplemented MAP specification does not employ such negotiation.

7he final and seventh level is the Application Layer. Tt Application Laver defines for oextain contexts what language is to be used during amwunication. The context can be either implicit, based upon the nature of the communication or it can explicitly stated. 'The Application Layer is needed because data being oammunicated ray be organized in fundwnentally different ways by different cm-oputer "tans. The Application Layer defines a nutually agreed upon wav in order to refer to things. One or both of the ccm=icating units my have to translate its data to the iwtually agreed upon format if such I format is not inherently provided.

As stated above, the present invention is involved primarily with testing the conformance of the level three Network layer internet prot I entity of a system under test. The cbjective is to transmit both valid and invalid data to the Network layer entity in order to determine whether it properly distinquishes between the two. Another objective is to cause the Network layer to pass data up to the 12 - Zransport Layer and also to receive data passed down from the Transport Layer without exror. The Network Layer is O=H=ticnless in that it does not iffplement cmusec-tion-oriented protocols as some of the layers do. Prior to this invention,, in order to test the Network Layer, it has be" necessary to place a special test program in the Transport layer of the system under test (or at the Network/Transport Layer interface) for the purpose of sending predefined test messages through the trariMort/network interface to the Network Laver in order to see if the Network Laver fmoperly receives the same. The test program also mist respond to predefined test messages sent through the transport/network interface from the Network Layer in order to detammine whether the Network Layer rrmperlv convevs messages. The aAdition of this test software adds to the complexity of the exmmmicatien system and makes it difficult to design high performance software which may need to integrate certain closelv related functions of the Transport and Network Lavers.

Entring thus presented a general overview of the invention and the presently preferred operating - otment, a more detailed explanation of the embedded test system will now be given.

The invention is referred to as the lenbedded" test system or technique because the network service interface in the system under test is allowed to remain naturally coupled to the Transport layer.

The cximen test method in which the network service interface is required to be decoupled, from the Transport Layer and. driven bv specialized software in the system under test is refexred to as the Oexposed" test system or technique.

The invention utilizes as its implementation base the software which was developed to perform the exposed test technique.

13 - g, the tion Is le of mmuting the same act of t&gt Mme me the expo test "tun. the Choice of techr Is left to the designer of the system under tent. In words# the Owbew tent systam uses the odating we' lra test g software as Its lementation b. in facto only one test system Is raintalned, with a user comand to @elect Ww an cnied er an servioe control l& available.

The version continues to use the MS defined Test ftment Protocol Data Units MPM) the associated Upper ter for control of the TUT service Interface. A descr of the IM (test it) Protocol can be found in IMISMAS-7,, National M of dards,, The Design of a Byatem for MWlwmta of 150 Ctnnect ess Network Protocol,,' July 1985 (NBS 86).

Figure 8 illustrates the presently preferred a of Implementing the cmed test system er=lc!ting the Immention. The sequence starts at step 100 and prom to set up the comection oriented between the test r software and the r under test at 102. We rernainder of the tent sequence illustrated in the flow chart of Figure 8 ises a plurality of br points (at at 104, 110, 120, 122, 128g 136 and 144) where specific tent& can be performed.

For eyle,, if it Is desired to test the internet protocol (IP) data trtters, flow bran at 104 to execute steps 106 and 108. In step 106 a predefined unit of data, referred to as an an AX IM Is sent with a t one qter than the previous sequence =. This cgww the credit window to allow one message unit to be ret t= the r u test. In 1Wlmmting these tiests the system under test would v=ally be 0- 0 IS' ,eorditi to send and expect to VC a large =ber of octets of data. cy c the =edit wi with a St cxw greater than the previous sequence, cine of the plurality of data cctets in permitted to tlow fram the w u tact to the test r. At step 108, the teat Watem sets a flag to C to ive a M TPM with a ene greater than the iow.

If It In desired to test the IP data Teceiver, stop 11n brw control to 112 which send an 1NSDU containing a DT TW1 with a sequence x= one greater than the =. A decision has been at step 114 whether to corrupt the nWo which carry the MMU. If corrupted, a flag is set to tall the test "tom to expect an error message ER MPDU or no response. If the data In rbot %&.%wted, a flag in set to tell the test ro tn expect an aclet nessage AX MU with a sequenoe number one greater than the previous.

As explained above# the rop m%, rainications protocol requires periodic sending of &lplicate aclt signals to guard against system hang up. C duplicate -.lmilg pet@ signals are 4 t---,ted at branch point 120. If a licate.].,9-,,9ir3edgrm!nt signal in received, no action in t by the "tem and flow control ly returns to the tep of the in ardSer to test whether the mxp)ence fie In p 18 is ved,, 122 nonitors receipt of the capected 9 a and bran to 124 If the low et in ved. Step 124 tea the wJamlodt to ace if It was In we-;roe to the test am If loo no follt 13 190p &1oe ja twilt In rq= at 126.

is - A similar test in performed at step, 128 to determine if the M 7M flnged at step, 108 has been rem If sop step 130 testa to determine whether this was in response to a p e m m. If so, an aclet signal AX TPM is went at 132. Odmr a fault oondition in signaled at 134.

In a similar fn at step, 136 a test Is perforeed to determine whether the expected error fed in 116 in received. if nog a determination in de as to whether this er= signal BR 1M la in response to a receiver test. If so, a licate DT ^ is went. Otherwise a fault =ndition in signaled, at 142.

Ply, if any other r=!P-M or nm-ER 1Ms am "ocived or if wT lnazopletely reassembled DT 1Ms are z x control bran at 144 to signal a fault at 146. At 148 a decision is whether to terminate the testing cw whether.to return to the top of the loop for further testing. If testing terminates, the connection is broken at whereupon the test program ends at 152.

The emed version uses standard Transport Protocol Data Units (TPDU) and a MAP 2.1 forming Transport Entity for TUT network service interface cxx,&=1. The IM Trariaport Entity is. in turn, controlled by the Remote Scenario Interpreter an used in Transport testing. Only one transport scenario (test cime) le required for use in wded 1P testinq. no direct moons to the IMIs network service lface is required..

Controlled uae of MAP11M Class 4 Transport P=tocol In mdx testing prcs the same functionality for cxxxtzollúM and d)eerving Network Service Interface Aftivity as the cut= protocol required In the oq wethod. ( we vc x P;, but the liost IM functionality is unneonsary to begin with.) During Internet 16 - Otocol (IP) Wat Case oxwution,, only M MUs and XX TMs are carried in the IM data portions. Transport cannection establiat and tennination are carried on outside the mocution of. any 1P Tent Cue.

Cegineca establist in initiated th Lmance of a mcxxuwict command. A CR-CC-AX TM occhanged is @2". 2he test syntom cent to trt CRa for a raaamLble before qiviM up. 2P %at Cam only execute if a cmvAction is deemed open.

port t mer and parameter values are cotabli th CR-CC negotiation, or predefined In the wbe 1P Wat Plan. Wiese am expo to he similar to values used In current transport protocol confo=anoe tests.

Valid nWs data sent by the Inwer Te are correctly aw M TMs with data format observing the P^p. Som=in Interpreter cxpecUtians. The Ratote Scenario Intarpr (RSI) In the a as used in standard transport protocol testing. 1M3 data ew!nt In invalid Mus is also mmt as correictly sequenced DT 2.

If the invalid 2PM in, then the tent r aignals a fault.

If a correc error response Is returned,, then the testing cant an passing. In c to maintain =&.cc transport me& oeing,.

the DT TEM will be reUmismitted In a valid 1PM after the correc error response in received.

An AX WM with correc ly increiased lower wi edge is In response to a DT WW carried In a valid IPDU.

In order to restrict the IM misport Zntitv frcn execraturely ending DTJMs, AX 1Ms went by the test r ly no t. an MMU "t Service Data Unit,, 1.. 0 a correctly I -m DT MU) from the TUT Is desired, an AX WM granting one 1 edit is sent by the looer %ster. If the In ter 1 a valid (i.e. a valid IDT IMU) the window is reclosed and the M 7M acle. 2As AX TYM qranting credit dom. not specify size. Such a specification me to be undesirable, since it would attaqpt to control an Urp _ementation specific parameter.

for omf=onoe testing purposest- the AX with credit Is sufficient.

Note that the DT -15MU data tram the IM in oqm to follow =&.cc Piewte scenario interpreter format. Eeth the In Teater and the IM are i to periodical transmit an AX MM to

satisfy the Transport wi thm t. This lea both parties to be ant of c' contAnued viability.

Embedded Service Control and Observation:

The exposed mode av. uses a confo=ance entry to keep track of all expected nyDU fr= the IM. This table Is a;Fzcpriately filled upon the test Watem mending an IPDU,, and the table entrv is rem upon reception of satisfying IM.

In enbedded mod, this table ia not used in connect and dlaconnect processes. For connect/disconnect processes only the transport mmuction context data structure is consulted.

During IP test case execution# both the cmfcxcmence table entries and trwt context is used. Since enly DT TEMs and AR TMs we umd daring IP tent awe execution. this Is straightfo. The test system uses the foll imlen:

1. DT 1M in received: if the DT is in sequence, it Is In the open window, the data portion Is correctly. and a on munce entry generated by an AX TM with credit has been lq went. then the entry is rmolved. the Is updated, AW IM In sent by the Inwer 7"ter.

2. AX IM with duplicate seq# is received: te the Tctxw in an app.ropriate wanner.

3. AX MM with' greater mqf is received: if a availd Data nW conformarboe entry exists and an DT TPM has 'hem, trwuniitted, Cen the entry ia ved, and the In updated.

4. DT TPW within valid data MU is amtt a good DT nnXJ ia to he sent of size n octets of data according to the Test Case. A Valid Data nWw Confoinance entry is generated, and the nw in sent.

5. UT TM within invalid data TPDU is sent: a had DT IM is to be sent Carrying n cc of data, and ex= report flag is set on, as according to the Test Case. A valid Data nnyj,-type conformance entry is generated, and the IM is sent.

6. AK MU with duplicate is sent: the wrctxln indicates that the window t has expired.

7. ^IM with duplicate W4 credit is bent: an carrying a correctly sequenced DT MU is wc;x. A tance entry is and the data 1PM with AK T= is went.

8. AX TM with greater a" is sent: a Valid Data IPDUw entry has been resolved and a data nW is sent.

9. ER nW is zpmp: if wErrm-IPDUw cxx.finunce entry ocists,, then the entry in rlved.

Inteigration with Current Internet Software:

The md) test flrct4Oniblity is with the ex software in wanner such that the es are isolated and 19 - cicarly struct:. An overview of the ax "tom softmm structure is presented, followed by the dwnes to w4p=t the Ced systern.

The exp "tom software Is structured an a Continually ing Process. The =difiealaltara to support the a system work within this structure. Cc fc)Umdng de ibes the &al. system. Statenients which are additions are marked Aus rodifications are narked M.

forever if (ca frem user terminal) m execute ca; A if (added A A send connectim raintenance, MUs A AKs to confirm valid Ms; if (test case has been queued && no tent case in-progress) read tent case; m build abstract nWsl queue abstract Mus for transformation; m fce entries for z p m P iving; wt up test ca is cenal dd698--M-&09ZCSS; if (abstract nMUs are queued) transform abstract MUs into corscrette "Me queue cancrete nWs for - - if (concrete IPWs are queued) Send Concrete IPDUsj if (inamirQ mris ars, in receive qusue) validate nW rl M Chec data and reassemble; M evaluate -- to test ca if (test case timer event) take sn=opriate action; if (test in progress && all confdrwrice entries resolved) test is finished; Modifications to mport the embedded systsm, are an follows:

1. The global data struct:ure 07L4=" provides necessary data to vjgxxrt oPeratiOn of the TransPort connection. This include the variables needed to generate and analyze data for the Pawte Scenario interpreter.

2. The MPDU data generated at abstract M?DU fo]= time is pr rided by a function rarced gen"(). This function is =odified to generate approswinte Dr TPDUs in embedded wA Minor =difications to the o=anoe entry-mking logic are 3. required.

3PDUs are generated to request the IM to send data. The data portion of such JPDU9 carries wSend-a;DU" WaWs in exposed mode.

These TK=s are generated at abstract UW fc=ation time by the function "nodu dataO." This function in =dified 9br admidded mode to generate AX TFWs with credit (a). Min= wd 4 fications toi the cm 1fmmmice entry-making logic are required.

ived nMa have their data portions Yzed by several differeint routines In exp Mode. Come routines @wit to dkmxrding upon whe ther the data was identified to be a lyym,, er p Just acquenced data. These =u are called from mither the ffireceivier 0 a cr wreasler 0 1 function.

5. In Mode. only a single routine ramd Oscan data()" is called. This routine is called from within Orasler (). 11 All IPDU deta portions are passed thr to sreasleros ther reasly is re or not (this in the existing practice for exposed also). The Owan dataOl routine Is Modified to apply transport PM type Vain. Fd, - Ived Me are m" to be within the alle wi. mcelvea me am m" to acknowleidge only outstanding Ms, or he duplicates for the purposes of connection Maintenance. The Oscan dataow function returns a count of exrors detected and raaalero" prints a masage indicating either encoding or,.1 violations.

6. Transport connection Is Managed by the routine "tr=to.,m This routine is called in the forever loop after the user comwe prmconsing has been called. User connect, disconnoct, and close camands (if valid) met variables indicating that a desired service be achieved. The "trwuqxwtO routine a my required concrete nMUs directly. Received nnXJs are handled bv scan datao within Orea ler 0. The Oscan data 0 0 its eh upon the indicators In Tctx.0 ormanoe Analysis:

Thie existing omf=mnce analysis mectv is used nearly dified for ewimffied Mode. The clwunes am (1) the @eqwnoe er variable in entry in adjusted to correlate with the - 22 U U - 1.3 cansport PDU activity; mid (2) an entry may require more than one anse to be resolved. This inplies that any existing to the mm.f.axame analysis wtnien still have not been emm@.

Transport level errors are annaumed by error mm.a to the qperat= cormhole. It in the atcw must 16 wtm such an er= message indicates 4M1 activity,, nm-teilitys, or omfo=cance W1ure.

Transport Ck:mmction Usaget The transport connection is always initiated by the test r. " LT proFmnes lao-bids on me negotiations. giving the IM m cc. Ihese bids are: checksum onj, n=nal tomatr mid "no ted data. " mx MU size Is;rrp!lmed to be 512 octets.

This size value is chosen to wj the current hare limitations of the test avatem. The TUT may negotiate this down.

aegardless of the negotiated ma size, the test system never sends an nW greater than the max message size in the test systern (currently 600 octets). " current set of IP Wet Cases does root znuire that an larger than 512 octets be sent or received by the test system. The test systam also Initiates the Aiaoorcdr-t cennanded by the console.

Every trtted AX carries the flow control-confi=ation (fec) pararnater. Paception of naked (Paremeterless) Me result in a retransmission of the latest AK.

Date Structures:

The following data structure in used to maintain the transp=t context:

23 ruct ICKWIM Tetx Int state; maJor allte 1 Int 1Cref; LT reference n 1 Int RTref 1 IM reference ru 1 Int sizebid; ma neted tpau size 1 Int men 2 swaing sequence nunber 1 int. 1 sending TA 1 int t 1 Int rwv recrbi sequence nu Int zwv psq it iving eqmme rurber 1 Int rw%e:cdt 1 1 r it 1 Int wi; 1 window t value Int Inact Ebw; 1:uwwuvity ti value Int giveLip tine; givmy tlimer value 1 Int retrai;-tirne; retranamission V value 1 Int retrad-count; 1 retransimission count 1 int er lw;-,. 1 C= pau 1 1 Int a)-len # 1 ak pau 1 1 int dt7hien; 1 dt pau headier 1 1 int dt:tlen; at pclu total 1 Int dr len; dir Pdu 1 1 int dd-len; 1 dc pau 1 1 Int ip:an; 1 ipdu header 1 1 int ip dai; 1 next dui to use 1 char erpdIu[S12); 1 cr tpau for mending 1 =signed char (512); 1 ak tpdu for sending 1 mnsigned char dtpduTS121; 1 at tpdu for mending 1 unsigned char dxpduTS121; 1 dr tpau for mending 1 wisi char dcpduTS123; 1 dc tpdu for me 1 unsi char ippda[7001; 1 data iou for sending 1 long count; count of tran data mot 1 long serxC; total wount of data to he sent recv count; count of transport data rpcpived 1 long recC; total amount of data to be ived int tat; classification of ak type Int It sequence ra of last dt mint 1 int 1 snce n=ber of last ak cent 1 Int send-pot; cot bent 1 Int recV @ú1t2 cot xempived 1 Int. tim at program Int send tive; 1 time last data transmitted 1 Int rea'"tim; 1 time last data vmrpived 1 Int sowCvector; 1 vector Inidicating rdtxt pdu to Int reason; reason 1 Int, dest; Mt-of-contmet pau dest 1 24 -nt arci 1 cut-of-context pdu arc 1 int WIsent =111 1 count of dt sent sequence number rollover Int Mack r0117 count of ak sent sequence number rollover Int IRTsent roll; count of dt recv sequence number rollover Int RTack roll; count of ak recv sequence mzdxw rgilm= char fault; 1 pointer to pdu error 1 struct Urre911011 registers to buffer, but unacked dta struct MODU 1 pointer to received pdu 1 tenance:

Transport TMO we fCd fOr bending for bic) purposes.

Mwy are either ired to form the nW data on definied by an IP Test Case,, or they are required to set upo, Zzintain cr te=dnate a transport Pm n-octim.

11M IM5 used for IP Test Case support are generated at abstract IPDU formation tine. Ctructim of them 'ZPE)Ug (only ME or AKs) reference the Tctx" structure for &Mzqxiate values.

MPDUs sent for o=mction maintenance am sent directly (the r=mal ing queues are bypassed). e MUs are sent after console cam processing. Ihe IM header is predefined with the MI (data unit Identifier),, segment 1 Indicator,, t Iength Indicator and header ctubc %;!mmrr ly wadified. All this ---t is found in the T&=. 1 DUI values start at (decimal) 1.000 so as not to conflict with IP Tent Case DUIs. IP Twt Cases 1 must refrain from, specifying IMUs with MI values greater than 1.000 to r potential conflict. Since all current IP Tmt C& use DUIs beg at (decimal) 1 (or for Inactive w tests no DUI at all).

MMUs umbd for control are easily le, t= MPDUs =xler test.

- W 11 6h All ved MMs (both mode) are acceased thr the wan data () 0 fon. In oqx 0 the MU and its data are ated dharing rmsly,, wid a data structure mmd a j w of type OR SES# 0 is created to maintain pertinent lnfcwmation about the aawment. Cw buffer holding the actual oet Is then released.

em rwmly p=ess maintains a 1 list of structiares for use in reasly.

1he data type JREASM SEGI is modified to support embedded mode by additional Infornation relevant to TMs. This iwludes partially rmslcd MUs. if the nw data is V&lid ao=rding to trort protoool rules and Soenario Interpr a, then only the confo=ance entry r be satisfied. This entry is satisfied if the sequence ru of the im=ing DT or AK =L i to that of the conformame entry.

Duplicate i=ming AKs ainply reset the inactivity tine and are diwarded. If the dhaplicate is r&," a flag is set in Tctx" indicating that an AX is to be trtted. All trtted Ncs contain the fee parwwter. All other cut-of-sequenoe " and An are Identified an errors.

Trasion of MM Is triggered by flags being set in Tvtx.w in the main loop, 41at ly after checking for owdiole c=s, the ft=tion ltransporto is called. If flags are set, then an &p p r c; T iate IM is sent. The machanism in used for mending all TMs e for tlx)se Ms and ARs which are req as MPM data during an IP Test Case.

llate, TMs are maintairmed In the OTcbc struct^ Such a TPM in ed Into a buffer an IPM, and the a;pprm;,.riate parweters are modified In the IM he and the TM.

2.7 O"his buffer in then passed to the device write rmztim, bming the queue structure for MMs ated by the IP Twt C.

The utr&n5PCrto" routine a TMs rx tpo wipp=t a Use CO (CCMCCt,, Clp dia=xct) p AX a valid 1M and cUme a windw (which previmsly was by an jp %at Cum ated AX) and to satl fy star trwurport t.

27 -

Claims (1)

1. C L A I M S

   A uke of tasting the Of an Inplenentation of an Intexnet protocol within a "tea test to its "cificationt =iprizing:

   using the flew control and nowloqgir:e,t niom of trumport, protocol to control the Internet (IP) protocol entity In. the "M U tent.

   2. The method of Claim 1 wh said flow control and acknowlet we used to waintain a controlled tent cannent.

   3. onbe me of Claim 1 Wherein maid flow control wuUams apecify the womt of musages expected to be tranemitted by the internet protocol entity in the oym under test.

   4. The rmthod of Claim 1 wherein said wJledgt ambanism is used to dete=dne whether the Internet W0t0001 Ontitv in the r under test c=rectly receives deliv data to a hi user level in the watem under tent.

   S. i%e nethed cif Claim 1 wherein mid flow control is used to define a w for all le transmissions in the internet protocol entity In the "stem u tent.

   28 - 6. A method of testing the of an AWlementation of an internet protoool within a vistam urxkx test to its specification,, omqprizing:

   using the flow control machanim of trans;xxrt protocol to control the internet (IP) Vmmtocol entity In the system w-der test.

   47. A uethod of testing the nce of an inplamentation of an ir&a=wt pratcool, within a wpftm under test to its specification, amprising:

   using the A. vadjanimm of transport protocol to control the intexnet (IP) protocol entity in the systm urAer test.

   8. A method of testing the conformance of an implementation of an internet protocol, such method being substantially as hereinbefore described with reference to the accompanying drawings.

   29 Publihhed 1988 at The Patent OffIce, State House, 86/71 High Holborn, London WC1R 4TP. Further copies may be obtained from The Patent OffIce.

   Sales Branch, St Mary Cray, Orpington, Kent BR5 3RD. Printed by Multiplex techniques ltd, St - Mary Cray, Kent. Con. 1187.