CN107807818B

CN107807818B - Multilayer model for SaaS multi-tenancy and construction method thereof

Info

Publication number: CN107807818B
Application number: CN201710964172.4A
Authority: CN
Inventors: 李响; 李彤; 谢仲文; 林英; 秦江龙; 莫启; 何云; 成蕾; 韩煦; 王晓芳; 郑交交; 杨真谛; 郑明�
Original assignee: Yunnan University YNU
Current assignee: Yunnan University YNU
Priority date: 2017-10-17
Filing date: 2017-10-17
Publication date: 2021-02-05
Anticipated expiration: 2037-10-17
Also published as: CN107807818A

Abstract

The invention belongs to the technical field of software, and discloses a SaaS multi-tenant oriented multilayer model and a construction method thereof. The characteristics of multi-tenant, scalability and configurability of the SaaS software are embodied; the color set of the colored Petri network can represent different tenant requests, the object-oriented Petri embodies the encapsulation of the service, the service is invisible to the outside, and the internal details influence the operation of the system. The colored Petri net and the idea of the object-oriented Petri net are used for reference respectively, the Petri net is expanded, and the CSOP and SOP net systems are provided. The invention provides a formalized tool for modeling of the SaaS software; the formalized model provided by the invention can run in a colored Petri net tool, so that the running of SaaS software can be simulated, and the hidden danger existing in the running of a system can be analyzed.

Description

Multilayer model for SaaS multi-tenancy and construction method thereof

Technical Field

The invention belongs to the technical field of software, and particularly relates to a SaaS multi-tenant oriented multilayer model and a construction method thereof.

Background

SaaS is a new software model that has emerged in recent years, and is an abbreviation of software as a service (serviceas software). It is very different from the traditional software model. The SaaS mode refers to a mode in which a software service provider provides network-based software services and is completely responsible for background work such as software and hardware platforms required for maintaining the software services, and a user rents the software services provided by the software service provider in a renting manner and uses the services through a network. Mention of SaaS one can think of cloud computing (cloud computing), which divides the services offered into three levels: iaas (infrastructuraasa service), paas (platformasa service), SaaS. And what we often mention as SaaS is actually the case in cloud computing. Com, the most successful SaaS software provider, its CRM (customer relationship management) solution, has a profound impact on the SaaS industry. Currently over 15000 businesses rent their services worldwide.

The SaaS not only changes the business model of the traditional software, but also changes the lifecycle of the traditional software. Compared with the traditional software, the most remarkable characteristics of the SaaS software are as follows: scalability (Saleable), Multi-Tenant architecture (Multi-Tenant), and configurability (Configurable). The methods and models for analyzing legacy software are no longer applicable to SaaS. To facilitate analysis of SaaS, the literature classifies SaaS 4 grades of SaaS according to the maturity of the SaaS software: custom development (multiple development); configurable (multiple deployments at one development time); single instance, multi-tenant (one development and one deployment); multi-instance, multi-tenant (scalable). Each layer realizes the function of the layer on the basis of meeting the requirement of the previous layer. It describes the process of stepwise evolution of SaaS software. The SaaS software maturity model is not only an industry-recognized standard, but also a specification that is followed by SaaS development.

The Petri net is a formal tool for describing software overactivity, but the basic Petri net faces the problem of state space explosion of a complex system. Colored Petri nets (ColoredPetri nets) are expanded on the basis of basic Petri nets, and different color sets are used for describing the system state. It can effectively change the folding system and compress the state space. Compared with the traditional software, the SaaS software has quite complex structure and operation condition. Especially, due to the characteristics of multiple tenants in the SaaS software, the SaaS software is difficult to describe.

The traditional method describes the SaaS software as a single-layer service composition relation by analogy with the composition relation of a component system. It is described and analyzed in a basic Petri net. The analysis actually analyzes the operation condition of a software system formed by combining components, and the software system is still traditional software, and the characteristics of configurable, scalable and multi-tenant of the SaaS software are not reflected. If the running condition of the multi-tenant is reflected by adding the number of the Petri nets. Firstly, the number of transitions is increased, so that the state space of the system explodes, and the analysis of the model has no need to be done; meanwhile, the relation among multiple tenants cannot be embodied by describing the requests of different tenants in multiple Petri networks.

In summary, the problems of the prior art are as follows: the traditional software has poor flexibility, no multi-tenant architecture and low configurability; the SaaS software maturity model is an important standard in the SaaS software industry, and meanwhile, the SaaS software maturity model gradually becomes a standard to be followed for developing SaaS, and the prior art does not support the SaaS software maturity model and does not reflect the characteristics of multiple tenants, scalability and configurability of the SaaS; especially, due to the complex architecture and operation condition of the SaaS software, the traditional method faces huge challenges; the SaaS software is described as a single-layer service combination relation, and is described and analyzed by using a basic Petri network, so that the problems of state space explosion and the like exist, and the relation among multiple tenants cannot be embodied.

Disclosure of Invention

Aiming at the problems in the prior art, the invention provides a multilayer model facing SaaS multi-tenancy and a construction method thereof.

The software-as-a-service (SaaS) software meta-model is provided by layering responsible SaaS software systems. Each layer is described top-down according to a different granularity. And the development standard of the industry standard and the SaaS software is supported by a SaaS software maturity model, and the characteristics of configurable, multi-tenant and telescopic performance of the SaaS software are fully embodied. And then, filling and modeling the actual content of each layer by taking the proposed meta model as a framework according to different SaaS software. And finally, analyzing the obtained model, and putting the CSOP into CPNtools for analysis to obtain an operation state analysis result of the SaaS software.

The multilayer model facing to the SaaS multi-tenant specifically comprises the following steps:

the application layer is used for providing the maximum granularity of the SaaS software and providing practical value for users;

software instance layer: the method is used for solving load balance by using multiple instances according to the SaaS software maturity model; different tenants are distributed into different instances by a load balancer of the software instance layer; one SaaS software has a plurality of SaaS software instances, different software instances are independent from each other, and the existence of other instances cannot be sensed, and the mutual influence cannot be further realized; a software instance is described as an extended colored Petri network of CSOP, and different color sets represent requests of different tenants and are used for visually representing the running process of the SaaS software instance. And the overall situation of the SaaS software can be analyzed, whether potential safety hazards such as deadlock and fault exist in the software instance or not, whether unfair competitive relations exist among different tenants or not and the like are judged. Different software instances have no relationship to each other and are not affected, so the different software instances are described as multiple CSOPs.

A tenant layer: the software is used for representing different tenants in the SaaS software; configuring the settings of different tenants to realize the configurability of interfaces, data and services; different tenants are allocated to different software instances, and the different software instances describe the behaviors of the different tenants;

presentation layer: for interacting with a user; and providing a function rented by the user, wherein the user can meet the requirement of the user by renting the function of the layer. And the functions of the layer are combined by atomic services.

And a service flow layer: a combined list for providing services, and configuring the combination mode of the services in the form of the list; the layer can flexibly combine atomic services, and different combination modes are added to evolve the same function. It is described using the extended Petri King SOP.

Service entity layer: a specific definition for providing a service; atomic services contained in SaaS software are described, as well as their internal structure. The complex functions in the SaaS software are all composed of these atomic services. Its interior is described as a basic Petri net, made up of different building blocks.

Infrastructure layer: the software-as-a-service (SaaS) software storage capacity, computing capacity and network connection function are provided, and are boundaries of the SaaS software.

Further, the application layer is a triplet, denoted as a ═ { N, I, S } where:

n is the unique identifier of the SaaS software;

i is a software instance layer of the SaaS software and describes a software instance contained in the SaaS;

and S is a service entity layer of the SaaS software and comprises all atomic services of the SaaS software.

Further, the software instance layer is a tuple, and the software instance layer is represented as I ═ { I ═ I_i,U,CSOP,F_i}, wherein:

I_iis a unique representation of an instance layer, an I_iAn element represents a software instance;

u is a tenant of the SaaS software and describes a tenant layer;

CSOP a colored service-oriented web structure describing the structure of this SaaS software instance, where different colors represent different tenants;

F_iis I_iThe corresponding relation of U and CSOP represents that different tenants are distributed to different SaaS software instances and represented by a COSP;

I_ithe relationship with U is a one-to-many relationship, I_iThe relationship with CSOP is a one-to-one relationship.

Further, the tenant layer U is a 3-tuple, denoted as U ═ I_u,R,F_u}, wherein:

I_uunique identification of tenant, I_uThe software is provided with a plurality of tenants;

the SaaS software is a presentation layer configured for each user; different tenants have different presentation layers;

F_uto represent the correspondence of a tier to a tenant,

one tenant corresponds to one representation layer, one representation layer corresponds to a plurality of tenants, and the relationship is a one-to-many relationship.

Further, the representation layer R is a tuple, represented as R ═ { I ═ I_r,SOP,S_r,F_r}, wherein:

I_ridentifying a presentation layer for a unique identification of the presentation layer;

the SOP is a service-oriented network;

S_ris a collection of all functions in a presentation layer;

F_ris I_r、SOP、S_rThe corresponding relationship of (a) to (b),

SOP and S_rThe relationship between is a one-to-one relationship, I_rAnd S_rThe relationship between them is a one-to-many relationship.

Further, the business process layer is a network structure SOP serving an object, where SOP is { P, PN; f, M₀,I_s}；

The service entity layer is an extended basic Petri network and is expressed as S ═ Ps, Ts; fs, M0, Is }, wherein:

wherein (Ps, Ts; Fs) is a basic net structure; ps is a library of the network, Ts is a transition, Fs is a directed arc defined from Ps to Ts;

m0 is the initial identification of the Petri net;

the Is the unique identifier of the network system and represents the whole network system to the outside.

Another object of the present invention is to provide a method for constructing a SaaS multi-tenant oriented multi-tier model,

according to the characteristics of the SaaS software, constructing a SaaS software hierarchical meta-model supporting the maturity of the SaaS software, and modeling and analyzing each layer by using a formalization method; the meta-model is a modeling framework for building different SaaS software models. And describing the software from top to bottom in a layered manner by using the meta-model according to the information such as the maturity of the actual SaaS software, the number of software instances, the number of tenants, the number of atomic services, the number of renting functions and the like. From the beginning of the application layer to the end of the service layer, the infrastructure layer describes the boundaries of the SaaS software, and the content of this layer is not contained in the specific software.

Constructing a service-oriented network structure SOP and a service-oriented network structure CSOP based on the object-oriented Petri network and the colored Petri network; the CSOP is used for describing a software instance layer, on one hand, different color sets represent different tenant requests, and the change can be effectively folded; on the other hand, the operation of the SaaS software instance is simulated through CSOP, so that the constructive safety and the evolutionary performance of the SaaS software instance are analyzed. The SOP can efficiently represent the composition relationship between atomic services.

Representing the service by using the encapsulated library element, and reflecting the invisible of the service to the outside and the influence of an internal composition structure of the atomic service on the running state; different color sets represent different tenant requests, and the characteristic of the SaaS multi-tenant is reflected.

Further, the service-oriented network structure SOP is an extended basic Petri network, SOP ═ P, T; f, M₀,I_s}, wherein:

{ P, T; f is a basic network system, P is a network depot, T is a transition, and F is a directed arc defined from P to T;

wherein, P tableService entities, the service entities are all from the service entities in the application layer

Meanwhile, the P is an expanded basic Petri network and represents the internal structure of the service;

M₀is an initial identification of the SOP;

I_sthe SOP is a unique representation of the SOP, and the whole network system is externally represented.

Further, the colored service-oriented network structure CSOP is an 8-tuple CSOP of an extended colored Petri network { P, T; f, C, W, M, I_s,II_s}, wherein:

{ P, PN; f is a basic network system, P is a network depot, T is a transition, and F is a directed arc defined from P to T;

p represents service entities, and the service entities are all from the service entities in the application layer

c is a color set of the SOP, is a limited non-empty set and represents different tenants in the service; the number of elements in C is the same as the number of tenants, and C is Ir;

W:F→L(C)+，

I_S：PN→L(C)+,

M:P→L(C)，

l (C) represents a non-negative coefficient linear function on the color set C, L (C) + represents L (C) whose coefficients are not all 0; w is a weight function from F to the color set, I_sFor mapping on transitions, M is the mapping of the library set to the color set;

II_sis a unique identification of the web structure.

The invention has the advantages and positive effects that: the invention provides a layered modeling method for SaaS software according to a SaaS software maturity model. The SaaS software is divided into seven different layers according to different particle sizes from top to bottom. The characteristics of multi-tenant, scalability and configurability of the SaaS software are embodied; the color set of the colored Petri network can represent different tenant requests, the object-oriented Petri embodies the encapsulation of the service, the service is invisible to the outside, and the internal details influence the operation of the system. The colored Petri net and the idea of the object-oriented Petri net are used for reference respectively, the Petri net is expanded, and the CSOP and SOP net systems are provided. A formalized tool is provided for modeling of the SaaS software; the formalized model provided by the method can run in the colored Petri network tool CPNtools, so that the running of SaaS software can be simulated, and hidden dangers existing in the running of a system can be analyzed.

The characteristics of multiple tenants are reflected by adding a Petri network in the traditional method, and the complex SaaS software system can be described in one Petri network, so that the transition is greatly folded, and the state space is compressed. So that the modeled SaaS software can be analyzed in CPNtools. And the description mode can analyze the concurrent relationship of different tenants requesting in the same service. The invention can describe the maturity of SaaS software of different levels, which cannot be realized by the traditional method. Grade one: for traditional software, multiple developments are required. Only one service in the service entity layer is added. All the other layers above are empty; grade two: multiple deployments are developed at one time. The software at this level is configurable, and only a service entity layer, a business flow layer and a presentation layer are added. The remaining upper levels are empty. Grade three: single instance multi-tenant. This level of software has all the levels, except that there is only one instance in the software instance layer. Grade four: multi-instance, multi-tenant. The software at the level is scalable, has all layers, can have a plurality of instances at a software instance layer, and can dynamically add and delete software instances according to system conditions.

Drawings

Fig. 1 is a flowchart of a method for constructing a SaaS multi-tenant-oriented multi-tier model according to an embodiment of the present invention.

Fig. 2 is a schematic diagram of a SaaS multi-tenant oriented multi-tier model provided in the embodiment of the present invention.

Fig. 3 is a SOP graphical representation provided by an embodiment of the present invention.

Fig. 4 is a detailed diagram of the SOP provided by the embodiment of the present invention.

FIG. 5 is a structural diagram of a CRM system provided by an embodiment of the invention.

Fig. 6 is a Petri net model diagram of a service entity layer provided in the embodiment of the present invention.

Fig. 7 is a Petri net model diagram of a business process layer according to an embodiment of the present invention.

FIG. 8 is a diagram of an example of software, a Petri Net model c1, according to an embodiment of the present invention.

FIG. 9 is a diagram of a two Petri Net model c2 as an example of software according to an embodiment of the present invention.

Fig. 10 is a diagram of example simulation software 1 provided by an embodiment of the present invention.

FIG. 11 is a diagram of example simulation software 2 provided by an embodiment of the present invention.

FIG. 12 is a diagram of an analysis report of a state space of an example software provided by an embodiment of the present invention.

FIG. 13 is a diagram of a software example two-state space analysis report provided by an embodiment of the present invention.

FIG. 14 is a bounded nature analysis report diagram of an example software provided by an embodiment of the invention.

FIG. 15 is a bounded nature analysis report diagram for example two software provided by an embodiment of the present invention.

FIG. 16 is a report of activity analysis of an example of software provided by an embodiment of the present invention.

FIG. 17 is a report of activity analysis of an example software application provided by an embodiment of the present invention.

Fig. 18 is a fairness analysis report diagram of the first and second software examples provided in the embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

The application of the principles of the present invention will now be further described with reference to the accompanying drawings.

As shown in fig. 1, the method for constructing a SaaS multi-tenant-oriented multi-tier model provided by the present invention includes the following steps:

s101, according to the characteristics of SaaS software, providing a SaaS software hierarchical meta-model supporting the maturity of the SaaS software, and modeling and describing each layer by using a formalization method;

s102, inspired by the ideas of object-oriented Petri nets (OOPN) and Colored Petri Nets (CPN), providing Service-oriented network Structures (SOPs) and Service-oriented Petri nets (CSOPs);

s103, representing the service by using the packaged library element, and reflecting that the service is invisible to the outside and the operation is influenced by an internal structure; different color sets represent different tenant requests, and the characteristic of the SaaS multi-tenant is highlighted.

As shown in fig. 2, the SaaS multi-tenant oriented multi-tier model provided by the embodiment of the present invention,

the system comprises seven layers, namely an application layer, a tenant layer, a software instance layer, a representation layer, a business process layer, a service entity layer and an infrastructure layer. The SaaS software is described from seven different layers respectively, and a meta model is proposed.

An application layer: the maximum granularity of the SaaS software emphasizes the value provided by the SaaS software for the user;

software instance layer: according to the SaaS software maturity model, the SaaS software uses multiple instances to solve the problem of load balancing. The load balancer can evaluate the load condition of the current software instance, and different tenants can be distributed into different instances by the load balancer. If a certain SaaS software is a single instance, an instance exists in a software instance layer;

a tenant layer: representing different tenants in the SaaS software. The SaaS can customize different SaaS software according to different users, and the software passes through the layer and is realized by configuring the settings of different tenants, and the interface, the data and the service can be configured;

presentation layer: the layer mainly interacts with a user, represents an interface of the user, and bears all functions of tenants of the user on the SaaS;

and a service flow layer: each function corresponds to a service combination mode, the layer has no specific service, only a service combination list, and different versions generate different service combination lists. Configuring the combination mode of the service in the form of a list;

service entity layer: the layer is mainly the specific definition of the service, the inside of each service is invisible to the outside, and the inside of the service realizes specific functions through the combination of the components;

infrastructure layer: the software-as-a-service (SaaS) layer is a hardware environment operated by SaaS software and abstracted into storage capacity, computing capacity and a network;

the characteristics of multiple tenants are embodied in a tenant layer, wherein the layer is provided with multiple tenants using SaaS software, different tenants can be distributed to different software instances, and the different software instances describe the behaviors of the different tenants on the software instances;

the extensible feature is embodied in a software instance layer, and one SaaS software can have a plurality of software instances. Multiple instances enable different tenants to be allocated to different software instances depending on the load situation. Then the software instance layer is proposed separately;

the configuration can be embodied in a presentation layer, and the tenant presentation layer is configured to enable different tenants to have personalized user interfaces, user management, data management, application programs and the like. Each presentation layer function corresponds to a specific service combination mode. The functions in the presentation layer are abstract, and users can add, delete and modify the functions at any time;

the meta-model mainly describes the SaaS maturity of level3 and level 4. And two popular modes of the SaaS are also provided at present. In level3, the SaaS software is a single-instance, multi-tenant schema. There is only one software instance in the meta-model's software instance layer, which is shared by all tenants. In level4, multi-instance multi-tenancy is its feature. The number of the software instance layers is increased to be multiple, and different tenants are distributed to different software instances through the load balancer;

the meta-model is mainly used for modeling the SaaS software, and is a model for establishing a model of the SaaS software. And aiming at different SaaS software, modeling layer by layer according to the meta-model, and dividing the complex SaaS software into different hierarchical descriptions. Each SaaS software instance is described on a colored Petri network. If the software is described as being mature at level 3. The single instance of software is the SaaS software itself. If the SaaS is described as the level4 maturity, a plurality of software instances form the SaaS software, and each software instance has a colored Petri net. The running conditions of different software instances form the running condition of the SaaS software;

the established SaaS model describes the subordination and corresponding relation of each layer element formally by using a set, so that the model is clear and accurate in description and has no ambiguity. So that the model is very convenient to analyze. For example, which tenant is assigned to which software instance, in turn rents those functions. Can be quickly obtained from the model. At the same time, the visualization portion of the model is capable of simulating the operation of the software instance. By analyzing the characteristics of the Petri network such as fairness, activity, boundedness and the like, the problems and hidden dangers existing in the system can be found and found out. For example, analyzing the activity of the Petri net can find out dead chains existing in the system, which are faults existing in the system and should be solved in time. And the fairness in the Petri network is analyzed, and unhealthy competitive relations among tenants generated due to design can be found out in time. And the bounded analysis ensures that the system cannot run without limitation.

The invention is further described below in conjunction with a formal modeling approach.

The formalized modeling method provided by the embodiment of the invention comprises the following steps:

the model is described in three places respectively by using a Petri network, and is a visualized part of the SaaS software model: the internal composition structure of the atomic service, the combination mode of the atomic service of the business process layer and the running conditions of different tenants of each software instance in the software instance layer. Inspired by the idea of a high-level Petri network-object-oriented Petri network, the invention provides a service-oriented network structure. The object-oriented Petri Net is proposed to solve the object-oriented related problem. One of the most important features of object-oriented encapsulation, object internal composition is not visible to the outside. It is just a feature of the service. The services are independent of each other, and the information exposed to the outside is very little, and almost only input and output are realized. The outside is not visible to the inside of the service. And inside the service, each service has its own composition mode. The internal composition of each service also affects the operation of the system. On the basis of meeting the integrity of system description, the characteristics of the model are highlighted, and branches are ignored. The invention introduces an expression description mode of an object-oriented Petri network. A service is described as a structure containing input and output libraries. The specific definition is as follows:

definition 1 (service oriented network structure SOP) the SOP is defined as an extended basic Petri net. SOP ═ { P, T; f, M₀,I_s}, wherein:

1) { P, T; f is a basic network system, P is a network depot, T is a transition, and F is a directed arc defined from P to T.

P here denotes the service entities, which are all from the service entities in the application layer

Meanwhile, P is an extended basic Petri network and represents the internal structure of the service.

2)M₀Is the initial identity of the SOP.

3)I_sThe SOP is a unique representation of the SOP, and the whole network system is externally represented.

4) In this net system, the graphical representation of the firing rules and transitions is identical to that of the Petri net, except that the libraries representing the service entities are represented by rectangular boxes with libraries. The libraries on the service entity represent the input and output of the service, respectively. Services are communicated and combined through transitions; as shown in fig. 3;

the service entity without refinement represents a library element, the input and output libraries of which are equivalent, wherein the token element can be arbitrarily moved.

5) The service entities can be refined by service subnets, and the interior of the service entities is a concrete implementation of the service entities and can be filled by a basic Petri network. The service is ensured to be invisible to the outside, and other services can only interact with the service through the input library and the output library. On the other hand, the integrity of the internal structure of the service is embodied, and the internal composition of a single service influences the running condition of the whole system; as shown in fig. 4.

The input and output libraries on the refined service entity are no longer equivalent, and the internal occurrence rules strictly obey the ignition rules of Petri. Token cannot move around in input and output libraries at will.

In order to enable the network structure to describe the running condition of multiple tenants of a software instance layer. And expanding the network, and describing the operating conditions of different tenants by using a color set in the colored Petri network, so that the network system can describe the state of the software instance used by multiple tenants simultaneously. The specific definition is as follows.

Definition 2 (colored service-oriented mesh architecture CSOP) CSOP defines an 8-tuple CSOP of an extended colored Petri mesh as { P, T; f, C, W, M, I_s,II_s}, wherein:

1) { P, PN; f is a basic network system, P is a network depot, T is a transition, and F is a directed arc defined from P to T.

2) C is the SOP color set, which is a limited non-empty set used to represent different tenants in the service. The number of elements in C is the same as the number of tenants, and C is Ir.

3)W：F→L(C)+

I_S：PN→L(C)+

M：P→L(C)

L (C) watchA non-negative coefficient linear function defined on the color set C is shown, l (C) + denotes l (C) where the coefficients are not all 0. W is a weight function defined from F to the color set, I_sTo define the mapping on transitions, M is the mapping of the library set to the color set.

4) The graphical representation of CSOP is the same as SOP.

5)II_sIs a unique identification of the web structure.

The software instances are independent of each other, and the existence of other software instances is not sensed. The sum of the operation conditions of each software instance constitutes the operation state of the whole complex SaaS software system.

According to the layered SaaS meta-model, the granularity described from top to bottom is smaller and smaller, and there is an inclusion relationship between layers. The application layer is the maximum granularity of SaaS and should contain all layers. Since the tenants are distributed into different software instances, the software instance layer and the tenant layer are in an inclusion relationship. Each tenant has its own presentation layer, which contains the presentation layer. Each function in the presentation layer requires a business process layer to describe the composition of the atomic service. The presentation layer and the business process layer are also in an inclusion relationship. For the SaaS, no matter how the instance of the SaaS software changes, the underlying atomic service does not change. The service entity layer is thus defined in the application layer.

The specific definition of each layer of the SaaS software meta-model is as follows:

definition 3 (application layer) the application layer is defined as a triplet, which can be expressed as a ═ { N, I, S } where:

1) and N is the unique identifier of the SaaS software.

2) I is a software instance layer of the SaaS software, describing the software instances contained in the SaaS.

3) And S is a service entity layer of the SaaS software, comprises all the atomic services of the SaaS software, and all the functions are composed of the service entities. S represents the internal structure of the atomic service by the basic petri net.

The application layer is the maximum granularity of the SaaS software, and the definition of the application layer can be regarded as the definition of the whole SaaS. SaaS includes the name of the application, the software instance, and the service entity. Since different tenants are allocated into different software instances, all service entity layers comprise the tenant layer. Due to the configurability of the SaaS software, the tenant layer includes all layers below the tenant layer. I.e. the software instance layer contains all the layers below it. Since the atomic service in SaaS does not change with the configuration of the software instance tenant, it is proposed separately to the application layer.

Definition 4 (software instance layer) defines a software instance layer as a tuple, which can be expressed as I ═ I_i,U,CSOP,F_i}, wherein:

1)I_iis a unique representation of an instance layer, an I_iAn element represents an instance of software.

2) U is the tenant of the SaaS software and describes the tenant layer.

3) CSOP a colored service-oriented web structure is used to describe the structure of this SaaS software instance, where different colors represent different tenants.

4)F_iIs I_iThe correspondence relationship between U and CSOP indicates that different tenants are allocated to different SaaS software instances and is represented by one COSP.

Define 5 (tenant layer) tenant layer U as 3-tuple, denoted as U ═ I_u,R,F_u}, wherein:

1)I_uunique identification of tenant, I_uThere are multiple tenants in the SaaS software, representing all tenants in the SaaS software.

2) The SaaS software is a presentation layer configured for each user. Different tenants have different presentation layers.

3)F_uTo represent the correspondence of a tier to a tenant,

one tenant corresponds to one presentation layerOne representation layer can correspond to a plurality of tenants, and the relationship is a one-to-many relationship.

Definition 6 (presentation layer) presentation layer R is a tuple, denoted R ═ I_r,SOP,S_r,F_r}, wherein:

1)I_ra presentation layer is identified for its unique identification.

2) The SOP is a service oriented network.

3)S_rIs a collection of all functions in a presentation layer.

4)F_rIs I_r、SOP、S_rThe corresponding relationship of (a) to (b),

Definition 7 (traffic flow layer) describes a traffic flow layer as a network structure SOP of one object service, SOP ═ P, PN; f, M₀,I_s}。

The business process layer describes the composition of the services. This combined relationship is described in terms of a service oriented network architecture SOP.

Definition 8 (service entity layer) describes the service entity layer as an extended basic Petri net, denoted S ═ Ps, Ts; fs, M0, Is }, wherein:

1) wherein (Ps, Ts; fs) is a basic mesh structure. Ps is the place of the net, Ts is the transition, Fs is the directed arc defined from Ps to Ts.

2) M0 is the initial identification of the Petri net.

3) The Is the unique identifier of the network system and represents the whole network system to the outside.

Second, the present invention is further described below in conjunction with case analysis.

Case analysis

SaaS was first successful on CRM (customer relationship management) systems, and most SaaS applications are currently CRM systems. The invention takes a CRM system as an example to verify the feasibility of the method.

This case is a simplified CRM system. Four types of tenants are provided, and each type of tenant rents different functions. Each function is formed by combining atomic services. The first tenant, the third tenant and the fourth tenant are respectively employees of the merchant. The tenant is responsible for analyzing the business opportunity of the market and finding customers for the business. And the third tenant is used for analyzing the market and making a market plan for the merchant. And the fourth tenant is responsible for making a market activity plan. The second tenant is a customer which mainly uses the system to inquire market information and find the needed merchants. The structure function diagram of the CRM is shown in figure 5. Tenants have roughly rented several functions:

and (3) recommending by the client: and recommending proper customers for the merchant according to the comments of other people. The system is formed by combining a customer information service and an evaluation service.

Customer management: customer information is managed for the merchant, and searching, modifying and deleting of the contact person are facilitated. The system consists of a customer information service and a market information service.

And (3) commercial opportunity analysis: and drawing a report form for the merchant according to the market information and the customer information to analyze the existing business opportunity. The system comprises a customer information service, a market information service and a report service.

Market information sharing: and sharing the searched market information to a third-party social network. The system consists of a market information service and a third-party social network service.

Recommending by the merchant: according to the comments of other people, the required merchants are recommended for the user, and the system comprises merchant information service and evaluation service.

Analysis of potential customers: and drawing reports to analyze potential users for the merchants according to the market conditions and the behaviors of various customers. The system consists of a customer information service, a market information service, a comment service and a report service.

Market management: the merchant inquires the market information and manages the market activity. The system consists of a customer information service, a market information service and a merchant information service.

According to the SaaS software meta-model, the model is built from bottom to top.

Service entity layer:

for convenience of writing, names of services are abbreviated respectively, wherein the customer information service is customer information service, the market information service is MIservice, the merchant information service is MIservice, the third-party social software service is SNservice, the report service is RFservice, and the comment service is Cservice. The service entity layer is described as a Petri network, and the specific structure is shown in FIG. 6.

And a service flow layer:

SOP＝{s₁,s₂,s₃,s₄,s₅,s₆,s₇}；

the graphical representation is shown in fig. 7.

Presentation layer:

R＝{Ir,SOP,Sr,Fr}；

Ir＝{r1,r2,r3,r4}；

Sr＝{ClientRecommend,ClientManage,BusinessAnalyse,MarketInformationShar e,OrganizationRecommend,PotentialClientAnalyse,MarketManage}；

SOP＝{s1,s2,s3,s4,s5,s6,s7}；

Fr＝{(r1,ClientRecommend,s1),(r1,ClientManage,s2),(r1,BusinessAnalyse,s3),(r2,MarketInformationShare,s4),(r2,OrganizationRecommend,s5),(r3,ClientManage,s2),(r3,PotentialClientAnalyse,s6),(r4,MarketManage,s7)}；

a tenant layer:

U＝{Iu,R,Fu}；

Iu,＝{rt1,rt2,rt3,rt4}；

R＝{r1,r2,r3,r4}；

Fu＝{(rt1,r1),(rt2,r2),(rt3,r3),(rt4,r4)}；

example layer:

I＝{Ii,U,CSOP,Fi}；

Ii＝{Instance1,Instance2}；

U＝{rt1,rt2,rt3,rt4}；

CSOP＝{c1,c2}；

Fi＝{(Instance1,r1,c1),(Instance1,r2,c1),(Instance2,r3,c2)}。

c1, wherein the color sets R1 and R2 represent the requests of tenant one and tenant two, respectively. R1 and R2 on the arc respectively limit the flow direction of the color set, which are the limiting functions on the arc, the color set R1 flows through the arc labeled R1, and R2, like R1, only flows through the arc labeled R2. Both color sets can flow through an arc labeled r. Wherein, there is arc suppression on transition t6, t8, t10, and only t9 transition has right to happen when there is token in service CIservice and minservice r 1. Enabling the services to be combined as normal according to the functions rented by the user. The graphical representation is shown in fig. 8.

c2 similar to c1, c2 shares two color sets, R3 and R4, representing tenant three and tenant four, respectively. R3 can only flow through the arc limited by the R3 function, and R4 can only flow through the arc limited by the R4 function. Both arcs R3 and R4 labeled R can pass through. Wherein, the transition t6 and the t9 are provided with the arc suppression, which is similar to the software instance I, and only the transition t7 has the right to happen when both services, namely, the R3 and the Miservice, have token. As shown in fig. 9.

An application layer:

A＝{N,I,S}

N＝{CRMSystem}

I＝{Instance1,Instance2}

S＝{CIservice,MIservice,OIservice,SNservice,RFservice,Cservice}

two software instance models were separately placed in CPNtool for simulation. All library elements can contain a composite color set RN, and the RN contains color sets of different tenants. The colors are numbers on the color set, and respectively represent tenant numbers. The arc is provided with a limiting function flow1(), a flow2, a flow3(), and a flow4(), which limit the flow direction of different color sets and only specific tenant requests can pass through. The normal operation of multiple tenants of the SaaS software in the same system is ensured. All tenant requests marked with r can pass through, indicating that the requests of a plurality of tenants share the arc. For reasons of space, the services in the Petri network are not refined. The simulation cases are shown in fig. 10 and 11, respectively.

As shown in fig. 12 and 13. The first software instance has 56 nodes and 119 arcs, the second software instance has 48 nodes and 88 arcs, and the models of the two software instances are strong communication graphs, which shows that the software models have better communication performance.

The data in FIGS. 14 and 15 show the number of tokens with the most and least ownership per library. It can be seen that the number of tokens in all libraries of the two software instances does not exceed 2. The system is bounded. And will not be in unrestricted operation.

FIG. 16 software example an activity assay report;

FIG. 17 software example two Activity assay report;

fairness analysis reporting for software instances one and two of fig. 18;

as can be seen from the activity analysis report, both software instances have only one dead label, and both are respective endpoint nodes, which is consistent with the design of the system. Furthermore, there are no dead chains. The system is reasonable in design and free of faults.

Neither software instance has an unfair sequence, indicating that there is no unhealthy competitive relationship between multiple tenants.

By modeling the SaaS software, the complex CRM system is exposed from different levels. The dependencies between modules are quite clear. The colored Petri network of the visualization part can simulate the running conditions of multiple tenants in respective software instances. The operation report of each software instance is analyzed, so that the SaaS has no hidden danger and faults such as deadlock, dead cycle and the like.

In the analysis work of SaaS software, many works are developed by modeling SaaS software using a Petri net. The literature provides a Petri network model, which models the cost condition of SaaS software and provides a basis for an enterprise to select traditional software or SaaS. But does not describe the behavior of the SaaS system. The literature uses a colored Petri network to model SaaS, and provides a method for predicting SaaS behaviors. But does not embody the characteristics of multi-tenant, scalability and configurability of the SaaS. Still other efforts have used other methods to model SaaS software. The literature provides a method for supporting the dynamic evolution of a multi-tenant architecture SaaS, which provides a strategy for the upgrade and update of the SaaS, but does not formally describe an analyzed system. The document proposes a SaaS software model of a multi-tenant architecture, which strictly defines service quality by using a mathematical method and reflects the service condition of the SaaS under the multi-tenant condition, but the model has no graphical representation and cannot intuitively express the behavior of the SaaS.

In response to the problems presented in the above-mentioned documents, the present invention proposes a SaaS software meta-model and defines it using formalized methods. The meta-model supports the maturity of the SaaS software. Fully embodies the characteristics of multiple tenants, scalability and configurability of the SaaS. The relationship among the modules is modeled by using a Petri network in a software instance layer, a business flow layer and a service entity layer of the meta model. The graphical representation intuitively reflects the internal behavior of the system. The software instance layer adopts a colored Petri network, different color sets represent requests of different tenants, the problem of explosion of a Petri network state space is solved to a certain extent, and system behaviors are simulated better.

The SaaS software meta-model provided by the invention provides a framework and a method for modeling SaaS software. The configurable and telescopic characteristics of the SaaS software are reflected, and the complex SaaS software is clear in venation due to the gradually refined granularity. Meanwhile, the colored SaaS software example described by the service-oriented network structure solves the problem of multi-tenant and embodies the description of the mature model of the SaaS software. The multi-layer SaaS software model describes the structure of the SaaS software, and the Petri network model simulates the running condition of the SaaS software. The hidden danger of the SaaS software system can be analyzed. And the interactive relation among the services can be analyzed according to the system operation result, and a foundation is laid for analyzing the dynamic modeling performance of the SaaS software.

Based on the existing work, a multilayer model is established for SaaS software, and reachable graphs of all color sets are obtained. On one hand, the behavior of the multi-tenant request service and the interaction relation among the services are analyzed, the correlation and consistency of the service behavior are analyzed, and the correct implementation of the dynamic evolution of the SaaS software is ensured. On the other hand, the description of the SaaS software multi-tenancy provides a method for further analyzing how the multi-tenancy influences the dynamic evolution of the SaaS software.

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.

Claims

1. A SaaS multi-tenant oriented multi-tier system, comprising:

the software instance layer is used for solving load balance by using multiple instances according to the SaaS software maturity model; different tenants are distributed into different instances by a load balancer of the software instance layer; one SaaS software has a plurality of SaaS software instances, different software instances are independent from each other, and the existence of other instances cannot be sensed, and the mutual influence cannot be further realized; the software instance is described as an extended colored Petri network of CSOP, different color sets represent the requests of different tenants and are used for visually representing the running process of the SaaS software instance; the overall situation of the SaaS software can be analyzed, whether potential safety hazards of deadlock and fault exist in a software instance or not and whether unfair competitive relations exist among different tenants or not are judged;

the tenant layer is used for representing different tenants in the SaaS software; configuring the settings of different tenants to realize the configurability of interfaces, data and services; different tenants are allocated to different software instances, and the different software instances describe the behaviors of the different tenants;

a presentation layer for interacting with a user; providing a function rented by a user;

the business process layer is used for providing a service combination list and configuring a service combination mode in the form of the list; combining the atomic services, and adding different combination modes to evolve the same function;

a service entity layer for providing a specific definition of a service; describing atomic services contained in SaaS software, and their internal structure;

and the infrastructure layer is used for providing SaaS software storage capacity, computing capacity and network connection function.

2. The SaaS multi-tenant oriented multi-tier system of claim 1, wherein the application layer is a triple, denoted as a ═ N, I, S } where:

n is the unique identifier of the SaaS software;

3. The SaaS multi-tenant oriented multi-tier system of claim 1, wherein the software instance layer is a tuple, the software instance layer being represented as I ═ I_i,U,CSOP,F_i}, wherein:

u is a tenant of the SaaS software and describes a tenant layer;

4. The SaaS multi-tenant oriented multi-tier system of claim 1, wherein the tenant tier U is a 3-tuple represented as U ═ I_u,R,F_u}, wherein:

F_uto represent the correspondence of a tier to a tenant,

5. The SaaS multi-tenant oriented multi-tier system of claim 1, wherein the presentation layer R is a tuple, denoted R ═ I_r,SOP,S_r,F_r}, wherein:

the SOP is a service-oriented network;

S_ris a collection of all functions in a presentation layer;

F_ris I_r、SOP、S_rThe corresponding relationship of (a) to (b),

6. The SaaS multi-tenant oriented multi-tier system of claim 1, wherein the business process tier is a network structure SOP of one object service, SOP ═ { P, PN; f, M₀,I_s}；

The service entity layer is an extended basic Petri network and is expressed as S ═ Ps, Ts; fs, M0, Is }, wherein: PN is { P, T; f, M₀Is a Petri net; t is transition; p is the place of the net, F is the directed arc defined from P to T, M₀Is an initial identification of the SOP;

m0 is the initial identification of the Petri net;

the Is the unique identifier of the Petri net and externally represents the whole Petri net.

7. The method for constructing the SaaS multi-tenant oriented multi-tier system of claim 1, wherein the method for constructing the SaaS multi-tenant oriented multi-tier system constructs a SaaS software hierarchical meta-model supporting the maturity of SaaS software, and a formalization method is used to perform modeling analysis on each tier; describing the software from top to bottom in a layered manner by using the meta-model according to the maturity of the actual SaaS software, the number of software instances, the number of tenants, the atomic service and the number of renting functions; starting from the application layer to the end of the service layer;

constructing a service-oriented network structure SOP and a service-oriented network structure CSOP based on the object-oriented Petri network and the colored Petri network; the CSOP is used for describing a software instance layer, on one hand, different color sets represent different tenant requests, and folding and transition can be carried out; on the other hand, the operation of the SaaS software instance is simulated through CSOP, and the constructive security and the evolutionary property of the SaaS software instance are analyzed; SOP is used to represent the composition relationship between atomic services;

8. The SaaS multi-tenant-oriented multilayer system construction method according to claim 7, wherein the service-oriented network structure SOP is an extended basic Petri network, SOP ═ P, T; f, M₀,I_s}, wherein:

wherein P represents service entities, and the service entities are all from the clothes in the application layerBusiness entity

M₀is an initial identification of the SOP;

9. The SaaS multi-tenant-oriented multilayer system construction method according to claim 7, wherein, in the service-oriented network structures SOP and CSOP, the colored service-oriented network structure CSOP is an 8-tuple CSOP ═ P, T; f, C, W, M, I_s,II_s}, wherein:

W：F→L(C)+，

I_s：PN→L(C)+，

M：P→L(C)，

II_sis a unique identification of the colored service-oriented network architecture CSOP.