WO2018134911A1

WO2018134911A1 - Resource allocation system, method, and program

Info

Publication number: WO2018134911A1
Application number: PCT/JP2017/001536
Authority: WO
Inventors: Shashank Bijwe
Original assignee: Nec Corporation
Priority date: 2017-01-18
Filing date: 2017-01-18
Publication date: 2018-07-26
Also published as: JP6881586B2; JP2020504552A

Abstract

A resource allocation system for communication service requests comprising: a service modeling unit which derives a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and a virtual network mapping unit which finds an optimal mapping for the virtual network components on a physical infrastructure based on the virtual network request; wherein the virtual network mapping unit sets reliability goals to each of the components by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, and finds the optimal mapping based on the reliability goals, the resource requirements and the reliability expectation for the service.

Description

RESOURCE ALLOCATION SYSTEM, METHOD, AND PROGRAM

The present invention relates to a system, method, and program of resource allocation for virtual networks. In particular, the present invention relates to a system, method and program of resiliency aware resource allocation for satisfying reliability guarantees for virtual networks.

Traditionally, in order to facilitate communication between different entities, physical networks comprising of various physical networking devices and communication links is deployed. With more and more devices communicating to each other, complexity of these physical networks have increased tremendously. Continuous increase in volume of traffic flowing on these networks have put additional challenges in the design of these communication networks. Increase in security threats, frequent demands for setting up networks, changing business needs and advances in technology among other issues make it further difficult to adapt these physical networks and results in slowing the overall growth and huge costs.

However, recently network virtualization by leveraging virtualization technology, have enabled network operators to move these communication networks from physical networks to virtual networks. Virtual networks makes it easy to deploy communication services faster, provide agility, isolation, and facilitate resource sharing and reduced costs among other benefits. These virtual networks consists of virtual nodes interconnected by virtual links. Virtual nodes may represent a component of an application, Virtual Network Function (VNF), etc., depending on the type of service.

Specifically, Network Functions virtualization (NFV) enables decoupling of physical network equipment from the functions that run on them. These functions can be implemented as software, also referred to as Virtual Network Functions (VNFs). For example, a network function such as firewall, can be realized with a plain software and commodity server. Each virtual network allows for consolidation of different network functions onto high volume servers, switches and storage devices. This consolidation reads to increased resource utilization and reduced costs. Commodity servers located in one or more Data Centers (DCs) can be used to host these VNFs, furthermore distributed DCs can be used to deploy these VNFs near customer premises to realize better service quality and satisfy various service level requirements.

Even though network virtualization have helped resolve various challenges associated with traditional networking but in order to realize true potential, different challenges needs to be addressed. One of the challenges is to optimally embed these virtual networks onto distributed physical infrastructure with hypervisor support. The physical infrastructure should satisfy various constraints associated with the virtualized networks such as capacity, bandwidth, location, etc. These category of problem is referred to as Virtual Network Embedding (VNE) problem and many algorithms have been proposed to solve the challenge which targeted different objectives such as efficient resource utilization, reduced energy consumption, failure tolerance, etc.

As virtual networks tries to become mainstream for provisioning services in telecommunication industry, there are additional requirements such as satisfying reliability and availability requirements to deliver the telecom grade services. Some arts have tried to improve the reliability for virtual networks by incorporating redundancies or only using the reliable physical resources for hosting.

For example,PTL 1 discloses a technology that employs redundancies to improve the reliability for the virtual infrastructure, furthermore redundant resources are shared among multiple virtual infrastructures.

PTL 2 discloses a technology that related design of a virtual network to a deployment on a physical infrastructure. Various entities involved in the end-to-end operation and sequence of operation between these entities in satisfying reliability and availability is disclosed in PTL2.

NPL 1 discusses approaches for end to end reliability estimation for service functions chains. It further discusses different protection schemes for VNFs, reliability issues faced during NFV software upgrades, different entities playing role in overall reliability and recommendations for deploying reliable services.

U.S. Unexamined Patent Application Publication No. US20110029675 A1 U.S. Unexamined Patent Application Publication No. US20130212285 A1

European Telecommunications Standards Institute, "Network Functions Virtualisation (NFV); Reliability; Report on Models and Features for End-to-End Reliability", ETSI GS NFV-REL 003 V1.1.1, Apr. 2016

Use of geographically distributed physical infrastructure for hosting virtual networks does lead to numerous benefits, however it breeds new challenge of optimally allocating these physical resources among different virtual networks while satisfying different requirements from the services. Furthermore complexity increases with varying constraints of physical infrastructure and different requirements for virtual networks. For example, physical infrastructure has constraints such as different locations, varying capacities, varying interconnectivity, varying reliability and availability levels, etc., and virtual networks have requirements such as capacity, bandwidth, varying reliability and availability expectations based on services, etc. Some related arts have tried to address the challenges but only could address the challenges to some extent.

Thus, the problem is that, in situations such as multiple service flows exists on single virtual network, virtual network components constituting the virtual network are shared among multiple virtual networks and have varying importance operational behavior, etc., merely satisfying the expected reliability for overall virtual network does not prove to be efficient during overall lifecycle of the service.

Another problem is that, physical resources with varying reliability values should be allocated optimally so as to maximize the acceptance of future virtual networks. Further, in some situations incorporating redundancies to match the expected level of reliability with least effort and using minimum resources by identifying appropriate location for redundancy placement is also crucial and remains to be one of the target problem of the invention.

The present invention is made in view of the above mentioned situation, and the object of the present invention is to provide a resource allocation system, a resource allocation method and a resource allocation program which can optimize allocation of reliable network resources.

A resource allocation system for communication service requests in accordance with the present invention comprises: a service modeling unit which derives a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and a virtual network mapping unit which finds an optimal mapping for the virtual network components on a physical infrastructure based on the virtual network request; wherein the virtual network mapping unit sets reliability goals to each of the components by distributing the reliability expectation for the service among them based on their relative importance in the virtual network , and finds the optimal mapping based on the reliability goals, the resource requirements and the reliability expectation for the service.

A resource allocation method for communication service requests in accordance with the present invention comprises: deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.

A resource allocation program for communication service requests in accordance with the present invention causes a computer to execute: a service modeling processing for deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and a virtual network mapping processing for finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.

The effect of this invention is that allocation of reliable network resources can be optimized.

Fig. 1 is a block diagram illustrating a configuration example of a resource allocation system according to the first embodiment. Fig. 2 is a block diagram illustrating a configuration example of the virtual network mapping unit 23. Fig. 3 is a flowchart illustrating an overview of operation of the resource allocation system 200 according to the first embodiment. Fig. 4 is an explanatory diagram illustrating a concept of virtual network mapping onto physical network. Fig. 5 is a flowchart illustrating an example of operation of a pre-processing unit 24 according to the second embodiment. Fig. 6 is an explanatory diagram illustrating a sample virtual network. Fig. 7 is an explanatory diagram illustrating list of factors useful for determining the relative importance of virtual network components. Fig. 8 is an explanatory illustrating a reliability block diagram of a network function. Fig. 9A is a flowchart illustrating an example of operation of a mapping unit 25 according to the third embodiment(1/4). Fig. 9B is a flowchart illustrating an example of operation of a mapping unit 25 according to the third embodiment(2/4). Fig. 9C is a flowchart illustrating an example of operation of a mapping unit 25 according to the third embodiment(3/4). Fig. 9D is a flowchart illustrating an example of operation of a mapping unit 25 according to the third embodiment(4/4). Fig. 10 is an explanatory diagram illustrating a function for minimizing reliability gap for node mapping. Fig. 11 is an explanatory diagram illustrating an example of fault tolerance configuration. Fig. 12 is an explanatory diagram illustrating a function for minimizing reliability gap for link mapping. Fig. 13 is an explanatory diagram illustrating a mechanism for improving reliability for network link. Fig. 14 is a block diagram illustrating a configuration example of an information processing apparatus to realize a resource allocation system according to the third embodiment. Fig. 15 is a block diagram illustrating an overview of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described below with reference to the drawing. The following detailed descriptions are merely exemplary in nature and are not intended to limit the invention or the application and uses of the invention. Furthermore, there is no intention to be bound by any theory presented in the preceding background of the invention or the following detailed description.

Skilled artisans will appreciate that elements in the figures are illustrated for simplicity and clarity and have not necessarily been drawn to scale. For example, the dimensions of some of the elements in the figures illustrating resource allocation system architecture may be exaggerated relative to other elements to help to improve understanding of the present and exemplary embodiments.

First Embodiment
Fig. 1 is a block diagram illustrating a configuration example of a resource allocation system 200 according to the first embodiment of the present invention. the resource allocation system 200 is responsible for allocating optimal amount of resources from one or more physical infrastructure(s) 27 to communication service requests. The resource allocation system 200 also referred to as a reliability aware resource allocation system 200.

The communication service requests are such that, it satisfies all the service level agreements and performance expectations specified by the user along with the service request. For example, a user can request for a Voice over IP service with throughput of 100Mbps, latency of 150ms and reliability of 99.99%. Hereinafter, the communication service requests are simply represented service request. Since the service request arrive to the resource allocation system 200 in online manner, the resource allocation system 200 does not possess the knowledge of the future incoming requests. The resource allocation system 200 accepts various types of the service request from multiple users or entities. It is also possible that operation support system (OSS) first accepts the service from user and then forwards it to the resource allocation system 200 for further processing and provisioning. In output, the resource allocation system 200 instantiates the requested service on the physical infrastructure 27 and ensures that requested performance is maintained throughout the lifetime of the service.

According Fig. 1, the resource allocation system 200 comprises a service modelling unit 21, a database 22, a virtual network mapping unit 23 and a deployment unit 26.

the service modelling unit 21 is responsible for the design of a virtual network request as per different requirements of the service. In order to design a suitable virtual network request and decide the constituent of a virtual network, the service modelling unit 21 takes functional requirements as an input. the functional requirements can be either mentioned with the service request or the service modelling unit 21 gathers this information from the database 22 by learning the type of the service request. For example, the functional requirements may involve deciding functions such as firewall, DHCP server, load balancer, encryption, decryption, etc.

The service request may include performance requirements such as latency and throughput, which helps the service modelling unit 21, decide the size of the virtual network requested. Size of the virtual network requested implies here the amount of virtual networking resources necessary to satisfy the expected performance for the service, such as CPU capacity, link bandwidth , etc. The service modelling unit 21 generates a virtual network request, which include capacity requirements for virtual nodes and bandwidth requirements for virtual links. the service modelling unit 21 may use techniques to precisely estimate the amount of the resources, such as performance modelling , estimation. In output, the service modelling unit 21 forwards the virtual network request to the virtual network mapping unit 23. The service modelling unit 21 also forwards it along with reliability requirements for the service and service time. These information can be contained in the service request or generated by the service modelling unit 21 based on the service request.

The database 22 acts as the knowledge store of the resource allocation system 200. The database 22 may hold the detailed information about various virtual network templates for various type of services. The virtual network template may consist of different network functions, their interconnection and order suitable for different types of services. The database 22 may also hold information about software instance of network functions. This information may include their operational behaviour, failure characteristics, recovery mechanisms, reliability and availability values among other information. The database 22 is synched with operation support system or external system (be)responsible for updating the information and adding new information in the database 22.

The database 22 also holds the knowledge about physical infrastructure 27, detailed network topology, their geographic location, available capacities, non-functional parameters such as reliability values and availability values for hardware resources and hypervisors, etc. This knowledge is usually synched with the resource monitoring unit and resource discovery unit (not shown in figure), and this knowledge continuously updates the information as the changes happens in the physical infrastructure 27.

The virtual network mapping unit 23 performs mapping to the physical infrastructure 27 of the virtual network, based on the virtual network request. The virtual network mapping unit 23 achieves optimal utilization of reliable resources by allocating just enough reliability to satisfy expected reliability for the service, it also improved overall reliability for the virtual network by minimum additional resources and costs. Fig. 2 is a block diagram illustrating a configuration example of the virtual network mapping unit 23. The virtual network mapping unit 23 further comprises a pre-processing unit 24 and a mapping unit 25 according to the first embodiment.

It is the task of the pre-processing unit 24 to distribute the expected reliability for the service among virtual network components. More specifically the expected reliability for the service may be an end-to-end reliability expectation in a virtual network running the service. This distribution is based on the relative importance of the virtual network components. To accomplish the task, the pre-processing unit 24 may comprises different functional modules such as virtual network analysis module, data collection module, reliability estimation module, priority assignment module, reliability goal setting module, etc. Details of these modules and their interconnection is covered in other relevant part of the description. In output, the pre-processing unit 24 forwards the virtual network request to the mapping unit 25 along with the distribution results. The distribution results are used as reliability goals and location constraints for virtual components.

The mapping unit 25 takes the virtual network request as an input from the pre-processing unit 24 and performs mapping to the physical infrastructure 27 of the virtual network components. It is the responsibility of the mapping unit 25 to satisfy the requirements for virtual network components specified with the virtual network request. For example, the requirements are that such as CPU capacity for virtual nodes, bandwidth for virtual links, reliability goals for virtual nodes & links respectively, location constraints, etc. The mapping unit 25 maximizes the revenue by accepting more and more request by optimally using the resources from the physical infrastructure 27. To achieve the goal, the mapping unit 25 employs different functional modules such as optimization module, criticality analysis module, node mapping module, link mapping module, reliability validation & negotiation module, etc. Details of these modules and their interconnection is covered in other relevant part of the description.

The mapping unit 25 outputs a deployment configuration template, which specifies the hosts from the physical infrastructure 27 for hosting the virtual network components, based on the mapping results. For example, the template may include server ID or DC ID for hosting virtual network node, physical link connecting two servers for hosting virtual link, CPU capacity requirement for the virtual node, bandwidth requirement for the virtual link, type of virtual node, etc.

The deployment unit 26 accepts the deployment configuration template as an input and instantiates the service on the physical infrastructure 27 by hosting the virtual networks on the physical infrastructure 27 according to the deployment configuration template. For hosting the virtual networks, the physical infrastructure 27 comprises hardware resources distributed geographically and virtualization layer. In addition, the hardware resources include computing, networking and storage resources. The deployment unit 26 may use virtual machines (VM) or containers for hosting the virtual nodes.

The service modelling unit 21, the virtual network mapping unit 23(more specifically, the pre-processing unit 24 and the mapping unit 25), and the deployment unit 26 are realized, for example, by a CPU of a computer operating according to a resource allocation program from a computer readable recording medium, for example, a program storage device, and operates as the service modelling unit 21, the virtual network mapping unit 23, and the deployment unit 26 according to the program. Alternatively, each of above-mentioned

components

21, 23 and 26 may be realized by separate hardware. Similarly, each of above-mentioned

components

24 and 25 may be realized by separate hardware.

Operation of resource allocation system
Subsequently, operation will be explained. Fig. 3 is a flowchart illustrating an overview of operation of the resource allocation system 200 according to the first embodiment. Operation starts with the step S31 with user or entity registering for the service request on the service portal or with operation support system (OSS). Service portal or OSS then forwards the service request to the service modelling unit 21. The service request may include functional requirements, performance requirements, service level agreements (SLAs), service time, etc.

In next step S32, after receiving the service request, the service modelling unit 21 converts the service request into a virtual network request. The service modelling unit 21 decides the constituent of a virtual network consists of and logical topology, using virtual network components. The service modelling unit 21 decides them as per functional requirements mentioned in the service request or by selecting a virtual network template for the type of service from the database 22.

In step S32, further the service modelling unit 21 decide the amount of resources such as CPU capacity , link bandwidth, etc. for each of the virtual network components as per the performance requirements mentioned in the service request, using performance estimation techniques, modelling techniques and so on. The service modelling unit 21 then forwards this virtual network request along with other SLAs such as reliability requirement and service time to the virtual network mapping unit 23.

In next step S33, after receiving the virtual network request, the virtual network mapping unit 23 performs performs mapping to the physical infrastructure 27 of the virtual network components, based on the relative importance of them and the absolute reliability of physical resources contained in the physical infrastructure 27.

More specifically, At step S33a, the pre-processing unit 24 analyses the virtual network request for topology and functional characteristics of the virtual network components, using the virtual network analysis module, etc. At this time, the pre-processing unit 24 refers database 22 for the necessary information, using data collection module, etc. Then, the pre-processing unit 24 decides the relative importance of virtual network components and assigns priorities in terms of weights, using the priority assignment module, etc. At this time, the pre-processing unit 24 plays a role in deciding the priorities by estimating the effect on overall reliability by each of the virtual network components, using the reliability estimation module, etc. Once the priorities are decided, the pre-processing unit 24 distributes the expected reliability value is mentioned with the service request among the virtual network components as per their corresponding weights, using the reliability goal setting module. After that, the pre-processing unit 24 is to add the virtual network request the distribution results.

At step S33b, once the pre-processing is complete, the mapping unit 25 gathers knowledge about the physical network, available capacities at nodes and available bandwidths on links and reliability values for each of them. Then the mapping unit 25 finds the optimal mapping for virtual nodes on physical nodes of physical network such that all the requirements are met, using the node mapping module with the support of the optimization module, etc.

Fig. 4 is an explanatory diagram illustrating a concept of virtual network mapping onto physical network. For example, fig. 4 shows mapping of virtual nodes of virtual network 11 on physical nodes of physical network 17. More specifically, a mapping virtual node 13 on a physical node 18 at DC 3, etc. In the event of reliability values are not satisfied, redundant nodes are employed to satisfy the reliability goal for that particular virtual node. The mapping unit 25 maps for all the virtual nodes in this way.

After all the virtual nodes from the virtual network are mapped, the mapping unit 25 finds the optimal mapping for virtual links on the physical network such that all the requirements are met, using the link mapping module with the support of the optimization module, etc. According to Fig. 4, for example a mapping virtual link 12 on physical link connecting DC4 to DC5. The mapping unit 25 maps for all virtual links in this way.

After that, the mapping unit 25 calculates the overall realized reliability of the virtual network and validates against the expected reliability mentioned in the service request, using the reliability validation & negotiation module, etc. In the event reliability is not satisfied, multi-commodity flows technique can be used to improve the overall reliability. In addition, the criticality analysis module helps find the optimal location for applying multi-commodity flow, such that reliability is improved greatly with less effort and cost. The mapping unit 25 maps for all virtual nodes & links onto one or more physical networks optimally again with the support of those modules.

The mapping unit 25 then writes the mapping information to the deployment configuration template and passes to deployment unit 26 for instantiating the service (step S34). In the present case, the mapping information includes all virtual nodes deployment location & size (amount of resources used by each virtual nodes) and all virtual links physical path & bandwidth requirements.

At step S35, after receiving the deployment configuration template, the deployment unit 26 reads deployment location and size of virtual nodes from the template and launches virtual machines or containers on physical nodes to host the virtual nodes. The deployment unit 26 then connects the virtual nodes by deploying virtual links on the physical path described in the template and reserving the link bandwidth as per the bandwidth requirements mentioned in the template. For hosting the virtual links, the deployment unit 26 may use techniques such as VLAN, VXLAN or VPN, etc. Once the virtual nodes & virtual links are hosted, it is ready for user to run their applications on the virtual network. The deployment unit 26 ensures that hosted virtual network satisfies the requested performance throughout the lifetime of the service.

According to the present embodiment, the resource allocation system 200 achieves optimal utilization of reliable resources by allocating just enough reliability to satisfy the reliability expectation for the service. Because, in the present embodiment, the resource allocation system 200 distributes the expected reliability for the service based on the relative importance of the virtual components at least, and maps the virtual components onto physical components using the distribution as reliability goal. It also improves overall reliability for the virtual network by minimum additional resources and costs.

Furthermore, the resource allocation system 200 by ensuring higher reliabilities for the relatively important virtual network components of the virtual network minimizes the adverse effect on the running services in the event of failures. It also ensures minimal mean time to recovery with minimal associated resource costs, and improves the acceptance ratio of the virtual network requests.

Second Embodiment
Fig. 5 is a flowchart illustrating an example of operation of the pre-processing unit 24 according to the second embodiment of the present invention. The pre-processing unit 24, of the resource allocation system 200 can be referred to understand the details of the system required to implement the process described in the second embodiment.

Network Functions virtualization (NFV), one of the child disciplines of network virtualization is getting popular recently and telecom carriers are rapidly adapting their networking approaches to benefit from reduced CAPEX & OPEX. To deliver telecom grade services it is important to satisfy high reliability demands of the services. In this embodiment, we will explain the process from the perspective of NFV, so as to make it easier for the person skilled in the art to understand the invention. However, this does not restrict the scope of invention for NFV and can be applied to approaches in other applicable disciplines with similar advantages. We will use the terms virtual network function (VNF) for virtual nodes, VNF-Forwarding Graph (VNF-FG) for virtual network and virtual network components for VNFs & virtual links interchangeably in rest of the embodiment.

According to Fig. 5, operation of the preprocessing unit 24 starts with the step S41 by obtaining the topology information of the virtual network. In the present case, the topology information is represented by

　
　
, where

　
　
is the set of virtual nodes and

　
　
is the set of virtual links. Understanding the interconnection and placement of VNF & virtual links in VNF-FG, whether parallel placement or series placement is important for the overall reliability estimation for VNF-FG.

In step S41, the pre-processing unit 24 also obtains information such as CPU demand

　
　
, bandwidth demand

　
　
, reliability expectation for the service

　
　
, service time T , etc. This reliability expectation is indicated by the service quality requirements and has similar meaning as the reliability expectation for the overall end-to-end of the VNF-FG. This reliability expectation can be referred as the reliability expectation for the virtual network, overall expected reliability for the virtual network, the end-to-end reliability expectation, etc.

Fig. 6 is a VNF-FG consisting of network functions (NFs) connected by virtual links, as a sample virtual network. Fig. 6 shows an example of the virtual network to which different NFs such as firewall (FW), network address translation (NAT) and intrusion detection system (IDS) are connected by virtual links. In virtual network, Each NFs contains a VNF which operates as the specific virtual node respectively. In General VNF is equivalent to the virtual part of the NF. Hereinafter, When referring to the virtual part of the NF, NF and VNF are not distinguished.

Then, at step S42, the pre-processing unit 24 gathers informations from the database 22 for differentiating different virtual network components. The database 22 may obtain those informations from common VNF catalogs, centralized orchestrator, OSS/BSS or other databases in the system, etc.

Note that in a virtual network some virtual network components hold more importance compared to other components. Such as, failure of important components can affect large portion of the services or lead to longer overall downtime of the service, also recovery time or recovery cost for some virtual network components can be higher than the other, etc. To deal with this situation effectively, the pre-processing unit 24 ensures higher reliabilities for these virtual network components, and ensures that they have lowest failure probability compared to others. However, to assign the reliabilities as per the importance, it is first important to identify the factors that are helpful in determining the relative importance for these virtual network components.

Fig. 7 is an explanatory diagram illustrating list of factors useful for determining the relative importance of virtual network components. These factors be explained along with their importance. However one should bear in mind that the list is exemplary in nature and not limited to these factors and can be expanded or altered to the situation or the service in consideration.

First category of the factors which is helpful for determining the relative importance of virtual network components is position in VNF-FG as a factor 72 shown in Fig. 7. It plays important role in overall reliability of the service. For example, the failure of NF which is shared among multiple service flows will affect the services much more compared to the NF which is not shared (See code 72a of Fig. 7). In Addition, Failure probability of NF placed in the series have direct impact on the overall end-to-end reliability of the VNF-FG, whereas failure probability of VNF placed in parallel combination with similar or different type of VNF affect much less on the overall end-to-end reliability of VNF-FG (See code 72b). Hereinafter, the factor 72 may be referred to as NF position 72.

Second category is type of VNF as a factor 73 shown in Fig. 7. Requirements on reliabilities for a VNF greatly depends on the type of workload. For example, Particular VNF handles such as VNFs handling signal processing, switching/forwarding, H-QoS, access control/authentication, etc., have strict requirements for higher reliabilities compared to the ones handling other type of workload (See code 73a). Assigning higher reliabilities to the stateful VNFs compared to the stateless VNFs is also important (See code 73b). It's reason that in the event of failure of stateful VNFs restoration of lost state is crucial for the service continuity , and it leads to increased downtime and increased recovery costs for the service. Another type which helps in differentiating between VNFs is, whether they execute the core function for the service or execute supplementary function (See code 73c). Structured VNF can be more prone to failure as compared to simple VNF instances, thus ensuring higher reliability for them is also important (See code 73d). Hereinafter, the factor 73 may be referred to as VNF type 73.

Third category is operational features as a factor 74 shown in Fig. 7. Different VNFs have different mean time to recover (MTTR) based on their implementation, size or complexity. Thus VNFs with higher MTTR acts as bottleneck for the overall service availability and ensuring higher reliabilities for these VNFs can improve the overall availability for the service (See code 74a). Also from the cost perspective, restoration of some VNFs involves more resources such as computing, networking or storage compared to other, and ensuring less probability of failure for these VNFs helps reduce the overall costs (See code 74b). Many a times depending on the type of service and failure of hosting server, changing environment or requirements or optimization purposes, some VNFs needs to be migrated to another infrastructure and many a times leads to downtime of services. Some VNFs are not fully supported for migration & leads to failure of the service during migration, thus ensuring higher reliability for them is also important (See code 74c). It reads to host these VNFs on reliable resources so that need for migration is reduced, and to increase the overall availability of the service.

Another feature is time slot length i.e. for a one particular service flow on VNF-FG involving multiple VNFs. Each VNF processes the information for particular portion of the time from the total processing time. Failure of the VNF having greater share of the overall processing time affects the service availability more compared to others. Thus ensuring higher reliability for them is also important (See code 74d). Also upgrades of the VNF software is essential for including new features, fixing bugs, etc., and many a times leads to failure of the VNF during upgrades and downtime for the service. Thus ensuring higher reliability for the VNFs requiring frequent upgrades can reduce the adverse effect on the services (See code 74e). Hereinafter, the factor 74 may be referred to as operational features 73.

At next step S43, the pre-processing unit 24 uses all the information collected in previous step S42 to determine the weights for the VNFs, which represents the relative importance of the VNFs in the VNF-FG. Higher the weight value, higher the importance and thus need for ensuring reliability proportional to the weight value of VNF. According to Fig. 7, the pre-processing unit 24 have categorized the factors in three categories for the ease of determining the weights by adopting the techniques such as analytic hierarchy process (AHP), multi criteria decision making (MCDM) or fuzzy set theory, etc. For example, in Fig. 7 with sample virtual network, different weights can be assigned to different VNFs such as

　
　
.

At next step S44, the pre-processing unit 24 derives the reliability equation for the virtual network. Reliability R for any component can be calculated as

　
　
, where MTTF is a mean time between failures of a component,

　
　
is a failure rate and T is service time. Unreliability F can be calculated as F = 1 - R. Reliability R of the virtual network depends on the arrangement of the virtual network components, for example, overall reliability of the two series components is the product of reliability of each one of the components, whereas in parallel arrangement it is the product of unreliability of each of the components that gives overall unreliability.

In step S44, the pre-processing unit 24 takes into consideration these factors and estimates the overall reliability equation for the virtual network. For example the overall reliability equation for the sample virtual network in Fig. 4 is

　
　
, where

　
　
represents the reliability for the network function (FW, NAT, IDS) which will be realized after instantiating on hardware and

　
　
represents the reliability for the network link which will be realized after instantiating on physical link.

At next step S45, after estimating the reliability equation for the virtual network, the pre-processing unit 24 distributes the reliability expectation for the service mentioned above among constituent components. Mathematically, left hand side of the reliability equation is set to the expected reliability and in right hand side reliabilities realized by all the network functions & links is set to common variable r, and then calculates the value of r. For example, in Fig 4,

　
　
after which r is used to balance the reliability goals proportionally among components based on the weights in the next step.

At next step S46, the pre-processing unit 24 calculates reliability expectations for virtual nodes and virtual links by considering the VNF software reliability values. Usually, the service modelling unit 21 in Fig. 1, while estimating resource requirements for virtual nodes and links also considers the type of VNF software to be used for realizing the networking functionality. Each VNF software has its own software reliability and plays important role in the overall realized reliability for that VNF.

Fig. 8 is an explanatory illustrating a reliability block diagram of a NF. According to Fig. 8, VNF software is instantiated on top of hypervisor which in turn is instantiated on top of hardware, thus a NF can translate to a series reliability block diagram for these three entities. In other words, the realized reliability for the NF is the product of reliability of VNF software, hypervisor and hardware. Since hypervisor and hardware combinations have less variations, thus can be referred together as NFV Infrastructure (NFVI) or simply hardware reliability.

In step S46, the pre-processing unit 24, first obtain the reliability values for these VNF softwares. For example, in Fig. 8,

　
　
is the software reliability for

　
　
respectively. Then, by using the weights determined in step S43, VNF software reliability and the value r determined in step S45, the pre-processing unit 24 calculates the reliability expectations for virtual nodes and links, which also can be referred as the reliability expectation for the hardware resources, the expected reliability for network components such as network link (NL) and network function (NF) , etc..
For example, in Fig. 6,

　
　
　represents the reliability goals for the hardware to instantiate the VNF and virtual links.

According to the present embodiment, the resource allocation system 200 , sets the weight based on the virtual network configuration and the functions of the network components, and then set the reliability goals accurately based on the weight and the reliability of VNF software. Therefore, resource allocation is optimized.

Third Embodiment
Fig. 9A to Fig. 9D illustrates a flowchart of an example of operation of the mapping unit 25 according to the third embodiment. Fig. 9A to Fig. 9B shows the node mapping phase of virtual network mapping operation by the mapping unit 25. In addition, Fig. 9C to Fig. 9D shows the link mapping phase of virtual network mapping operation by the mapping unit 25. the mapping unit 25, of the resource allocation system 200 can be referred to understand the details of the system required to implement the virtual network mapping operation.

First, the node mapping phase will be described below with the Fig. 9A and Fig. 9B. Process starts with step S801. In step S801, the mapping unit 25 obtains the information from virtual network configuration template. The information is such as a topology information of the virtual network, represented as

　
　
. Each virtual node contained in the above topology information is associated with CPU demand and reliability expectation, represented as

　
　
respectively. Each virtual link contained in the above topology information is associated with bandwidth demand and reliability expectation, represented as

　
　
respectively.

　
　
has similar meaning as the reliability goals for the hardware , represented as

　
　
respectively in the second embodiment.

In step S801, the mapping unit 25 further obtains information such as associated weighs

　
　
and service time T ,etc.

At next step S802, the mapping unit 25 is to obtain the substrate network information such as topology information, represented by

　
　
, where

　
　
is the set of substrate nodes and

　
　
is the set of substrate links. Each substrate node

　
　
contained in the above topology information is associated with CPU capacity

　
　
and nodes failure rates

　
　
, which is substituted in reliability formula with service time T to obtain the reliability

　
　
. Each substrate link

　
　
contained in the above topology information is associated with bandwidth capacity

　
　
and link failure rates

　
　
.

The following are use case of these parameters.

Available/residual CPU capacity of substrate node:

　
　
Available/residual bandwidth of substrate link:

　
　
As the path r, joining two substrate nodes may involve multiple substrate links, the available bandwidth capacity of the path can be expressed as follows.

At next step S803, the mapping unit 25 sorts the VNFs in the descending order of the reliability expectations

　
　
.

At next step S804, the mapping unit 25 finds the mapping for VNF with highest reliability expectations as per the objective function while satisfying the associated constraints. The objective function may consider multiple objectives as per user requirements, for example, satisfying the reliability expectations, cost minimization, lowering energy costs, etc.

Node mapping can be represented as

　
　
which denote a mapping function between virtual nodes and substrate nodes.

　
　
represent the substrate node selected for mapping the virtual node

　
　
. We propose to formulate the reliability satisfaction in objective function rather than in constraints which gives the advantage of increase in acceptance ratio for the virtual network requests. We present the equations here in order to help the person skilled in art to fully understand the invention.

Minimizing the reliability gap for node mapping:

　
　
- the function which ensures the minimal reliability gap while rewarding favorable region and penalizing unfavorable region.

　
　
- the reliability of substrate node where virtual node is mapped.

Fig. 10 illustrates the function graphically, where,

　
　
- favorable region and

　
　
- unfavorable region.

We also present the constraints as:
Node Capacity constraint (CPU demands of virtual nodes to be embedded on substrate node must be less than the available CPU capacity of that substrate node.)

　
　
Availability/ Node mapping constraint (Maximum number of virtual nodes from same virtual network request allowed to host on one substrate node)

　
　
. Value k restricts the number of virtual nodes from same virtual request to be mapped on same substrate node. Flexibility of choosing k can benefit the applications which are ok with hosting multiple virtual nodes on same substrate node.
　
Domain constraint

At next step S805, the mapping unit 25 verifies if the realizable reliability for the network function (NF) for the selected mapping from step S804 is greater than the expected reliability for that particular VNF. If yes, jumps to step S809 of Fig. 9B, else takes step S806. The realizable reliability for the network function (NF) is the product of reliability of selected substrate node (including hypervisor) and the reliability of VNF in consideration.

The mapping unit 25 takes steps to improve the reliability for the NF for which the realizable reliability is less than the expected reliability. It is possible to improve that by two approaches. First approach is fault avoidance, which involves relocating VNF to the hardware which can offer more reliability. Second approach is fault tolerance, which involves adding redundant hardware resources for VNF in the event of failure.

At step S806, the mapping unit 25 verifies the suitability of these approaches according to the requirements of the service. More specifically, the mapping unit 25 confirms that what is permitted by the service requirement. If it is fault avoidance only, takes step S807, else if it includes fault tolerance, takes step S808.

At step S807, the mapping unit 25 takes the first approach of fault avoidance for improving the reliability. The mapping unit 25 searches the next candidate satisfying constraints by passing the control to step S804. In the event no more candidate exists which satisfy the constraints, control is passed on to the next step S808.

At step S808, the mapping unit 25 takes the second approach of fault tolerance for improving the reliability. The mapping unit 25 adds redundant hardware resources for the VNF. Fig. 11 illustrates one of the aspect of fault tolerance configuration. Where, redundant hardware is allocated for the

　
　
. Redundant configuration can be set as active-passive or active-active depending on the choice of fault tolerance policies. Redundant resources can also be shared among multiple virtual nodes to improve resource usage. Fault tolerance configuration shown here is for exemplary purpose and does not restrict the invention to the particular configuration.

At next step S809, the mapping unit 25 records the mapping in virtual network configuration template and updates (recalculates) the available capacity of substrate resources.

At next step S810, the mapping unit 25 takes into consideration the anti-affinity requirement for the node placement, which restricts the two virtual nodes from being mapped on the same substrate node. Placing two virtual nodes on same substrate node, leads to failure of both the virtual nodes in the event that particular substrate node fails. Thus the mapping unit 25 confirms the anti-affinity is permitted. If yes, takes step S811, else jump to step S812.

An step S811, the mapping unit 25 temporarily set the available capacity for the selected substrate node to zero so that same substrate node is not selected for mapping the other virtual node from the same virtual network.

At next step S812, the mapping unit 25 inserts the realizable reliability value in the reliability equation and recalculates the relative reliability expectations for the remaining components using the weight values. Motive of this step is to satisfy overall reliability expectation and ensure priority for different virtual nodes by minimal use of reliable resources, which leads to increase in revenue.

At next step S813, the mapping unit 25 checks if mapping of all VNFs is completed, if yes, passes control to the link mapping phase, else passes control to step S803.

Next, the link mapping phase will be described below with the Fig. 9C and Fig. 9D. Process starts with step S901. In step S901, the mapping unit 25 inserts the realizable reliability of all NFs in the reliability equation and then calculates the reliability expectations for the virtual links.

At next step S902, the mapping unit 25 assigns weights to virtual links by sharing the VNFs weights among virtual links attached to that VNF. In other words, the virtual links connecting higher priority VNFs deserve high priority than other links. and the mapping unit 25 assigns weights to virtual links based on the weights of VNFs connected using those links. Then the mapping unit 25 sorts the virtual links in descending order of weights.

At next step S903, the mapping unit 25 selects the virtual link from the unmapped links with highest weight and finds the shortest paths between network functions (NFs) joining the selected link. There can be multiple shortest paths on substrate network joining two nodes.

At next step S904, the mapping unit 25 calculates the realizable reliabilities for the selected paths and selects the one as per the objective function and satisfies associated constraints. The objective function may consider multiple objectives as per user requirements, for example, satisfying the reliability expectations, cost minimization, lowering energy costs, etc.

Link mapping can be represented as

　
　
which denote a mapping function between virtual links and substrate links.

　
　
represent the substrate link selected for mapping the virtual link

　
　
. Selected substrate link also can be referred as network link (NL). We propose to formulate the reliability satisfaction in objective function rather than in constraints which gives the advantage of increase in acceptance ratio for the virtual network requests. We present the equations here in order to help the person skilled in art to fully understand the invention.

Minimizing the reliability gap for link mapping:

　
　
- the reliability of substrate path r, where virtual link is mapped.
The substrate path may consist of multiple substrate links, thus the reliability can be calculated as

Fig. 12 illustrates the function graphically, where,

　
　
- favorable region and

　
　
- unfavorable region.

We also present the constraints as:
Link capacity constraint (Bandwidth demands of virtual links to be embedded on substrate path must be less than the available bandwidth capacity of that substrate path.)

　
　
Domain constraint

At next step S905, the mapping unit 25 verifies if the realizable reliability for the network link (NL) for the selected mapping from step S904 is greater than the expected reliability for that particular virtual link. If yes, jumps to step S907, else takes step S906.

The realizable reliability for the NL is the product of the reliability of selected substrate link and reliability of virtual networking protocol in consideration for deploying the virtual link. In short, it depends on all the entities taking part in carrying out communication on that link.

At step S906, the mapping unit 25 takes steps to improve the reliability for the NL for which the realizable reliability is less than the expected reliability. For improving the reliability of the network link, the mapping unit 25 can take one of the two approaches as per suitability for the service and availability of resources. In first approach of path splitting, traffic of the virtual link is split among two virtual links and then maps onto two distinct shortest paths on the substrate network. In second approach, it reserves the dedicated link capacity equal to virtual link for distinct path.

Exemplary Fig. 13 illustrates the mechanism for improving reliability for network link (NL). In first approach, traffic between

　
　
and

　
　
is distributed among

　
　
and
　
　
with suitable splitting ratio. Whereas for second approach, these pass have same capacity, which is equal to the traffic between

　
　
and

　
　
. In second approach it can adopt shared backup, active-active or active-passive configuration as per the situation. Step S906 is repeated until the condition of step S905 is satisfied or until the predetermined number of time is completed.

At next step S907, the mapping unit 25 records the mapping in virtual network configuration template and updates (recalculates) the available capacity of substrate resources.

At next step S908, the mapping unit 25 checks if mapping of all the virtual links is completed, if yes, passes control to the step S903, else passes control to step S909.

At step S909, the mapping unit 25 calculates the overall reliability for the virtual network by inserting the realizable reliability values of network functions (NFs) and network links (NLs) in the reliability equation.

At next step S910, the mapping unit 25 checks to see if the overall realized reliability for the virtual network is more than the expected reliability for the service. If not it passes control to step S911 for improving reliability, else passes control to step S914.

At step S911, the mapping unit 25 derives the criticality importance measure for all the network links (NLs). The criticality importance measure helps identify the component, whose improvement in the reliability improves the overall reliability greatly, thus requiring minimum resources for satisfying the overall reliability goal for the virtual network.

At next step S912, the mapping unit 25 selects the network link with the highest criticality importance measure. Then at step S913, it improves the reliability of the selected network link (NL) by path spitting or dedicated backup similar to step S906. It then calculates the newly realized reliability for the network link (NL) and passes the control to step S909.

At step S914, the mapping unit 25 finally forwards the virtual network configuration template to the deployment unit 26.

According to the present embodiment, the mapping unit 24 maps the virtual components using several approaches based on the expected reliability for them as well as the expected reliability for the service, so that optimal mapping is realized.

Fourth Embodiment (Information processing apparatus)
Fig. 14 illustrates, by way of example, a configuration of an information processing apparatus 900 (computer) which can implement a resource allocation system relevant 200 to an exemplary embodiment of the present invention. In other words, Fig.14 illustrates a configuration of a computer (information processing apparatus) capable of implementing the system or apparatus, such as in Fig.1 or 2, and representing a hardware environment where the individual functions in the above-described exemplary embodiments can be implemented.

The information processing apparatus 900 illustrated in Fig.14 includes the following as components:
- CPU 901 (Central Processing Unit);
- ROM 902 (Read Only Memory);
- RAM 903 (Random Access Memory);
- Hard disk 904 (storage device);
- Communication interface 905 to an external device (Interface: hereinafter called "I/F");
- Reader/writer 908 capable of reading and writing data stored in a storage medium 907 such as CD-ROM (Compact Disc Read Only Memory); and
- Input/output interface 909.

The information processing apparatus 900 is a general computer where these components are connected via a bus 906 (communication line).

The present invention explained with the above-described exemplary embodiments as examples is accomplished by providing the information processing apparatus 900 illustrated in Fig.14 with a computer program which is capable of implementing the functions illustrated in the block diagrams (Figs.1, 2) or the flowcharts (Figs. 3,5, 9A-9D) referenced in the explanation of these embodiments, and then by reading the computer program into the CPU 901 in such hardware, interpreting it, and executing it. The computer program provided to the apparatus can be stored in a volatile readable and writable storage memory (RAM 903) or in a non-volatile storage device such as the hard disk 904.

In addition, in the case described above, general procedures can now be used to provide the computer program to such hardware. These procedures include, for example, installing the computer program into the apparatus via any of various storage medium 907 such as CD-ROM, or downloading it from an external source via communication lines such as the Internet. In these cases, the present invention can be seen as being composed of codes forming such computer program or being composed of the storage medium 907 storing the codes.

The previous description of the embodiments is provided to enable a person skilled in the art to make and use the present invention. Moreover, various modifications to these exemplary embodiments will be readily apparent to those skilled in the art, and the generic principles and specific examples defined herein may be applied to other embodiments without the use of inventive faculty. Therefore, the present invention is not intended to be limited to the exemplary embodiments described herein but is to be accorded the widest scope as defined by the limitations of the claims and equivalents. Further, it is noted that the inventor's intent is to retain all equivalents of the claimed invention even if the claims are amended during prosecution.

Fig. 15 is a block diagram illustrating an overview of the present invention. A resource allocation system 500 for communication service requests comprising a service modeling unit 501 and a virtual network mapping unit 502.

The service modeling unit 501 (for example, the service modeling unit 21) derives a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests.

The virtual network mapping unit 502 (for example, the virtual network mapping unit 23) finds an optimal mapping for the virtual network components on a physical infrastructure based on the virtual network request. Wherein the virtual network mapping unit 502 sets reliability goals to each of the components by distributing the reliability expectation for the service among them based on their relative importance in the virtual network , and finds the optimal mapping based on the reliability goals, the resource requirements and the reliability expectation for the service. When sets the reliability goals, the virtual network mapping unit may ensure higher reliabilities for the relatively important virtual network components of virtual network minimizes the adverse effect on the running services in the event.

A part or the whole of the embodiments may also be described as the following supplementary notes, but is not limited to the followings.

(Supplementary note 1)
A resource allocation system for communication service requests comprising:
a service modeling unit which derives a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and
a virtual network mapping unit which finds an optimal mapping for the virtual network components on a physical infrastructure based on the virtual network request;
wherein the virtual network mapping unit sets reliability goals to each of the components by distributing the reliability expectation for the service among them based on their relative importance in the virtual network , and finds the optimal mapping based on the reliability goals, the resource requirements and the reliability expectation for the service.

(Supplementary note 2)
The resource allocation system according to supplementary note 1 further comprising:
a deployment unit which launches the service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in a deployment configuration template received from the virtual network mapping unit;
wherein the service modelling unit receives the service request, derives the virtual network request and pass on to the virtual network mapping unit, which then pre-process it to set reliability goals to the virtual network components, finds optimal mapping, ensures the reliability expectation for the service and forwards the deployment configuration template describing the mapping result and the resource requirement of each of the virtual network components to the deployment unit.

(Supplementary note 3)
The resource allocation system according to

supplementary note

1 or 2 further comprising:
a database which stores template informations about various virtual network templates suitable for different types of services and information about physical infrastructure;
wherein each of the template information includes knowledge about one or more network functions belong to the virtual network indicated the template, such as their interconnection, order, operational behavior, failure characteristic, recovery mechanism, and reliability requirement, and the information about physical infrastructure includes network topology, location, available capacities, reliability or availability values at least, and
the database is accessible by the service modelling unit and the virtual network mapping unit.

(Supplementary note 4)
The resource allocation system according to any one of supplementary notes 1 to 3, wherein the virtual network mapping unit further comprising:
a pre-processing unit which distributes the expected reliability for the service among the virtual network components of the virtual network based on the relative importance of the virtual network components;
wherein the pre-processing unit decides the relative importance by calculating the weight value for each of the virtual network components by considering a plurality of factors, and
the factors comprises position in the topology, type of the virtual network component and operational feature.

(Supplementary note 5)
The resource allocation system according to supplementary note 4, wherein the pre-processing unit further comprising:
a virtual network analysis module which analyzes the position of the virtual network components in the virtual network topology;
a data collection module which gathers the information associated with the factors crucial for calculating the weight value for the virtual network components;
a reliability estimation module which estimates the reliability equation describing the relationship between set of the reliabilities for virtual network components and reliability for the service in the virtual network;
a priority assignment module which calculates the weight values for the virtual network components based on the factors; and
a reliability goal setting module which distributes the expected reliability for the service among the virtual network components using the weight values.

(Supplementary note 6)
The resource allocation system according to any one of supplementary notes 1 to 5, wherein the virtual network mapping unit further comprising:
a mapping unit which performs optimal mapping of the virtual network components on physical infrastructure such that all the requirements mentioned in the virtual network request received from the service modelling unit are satisfied;
wherein the mapping unit performs a node mapping and a link mapping, and ensures minimal reliability gap between the reliability goals for the virtual network components and available reliability values of resources from physical infrastructure.

(Supplementary note 7)
The resource allocation system according to supplementary note 6, wherein the mapping unit further comprising:
an optimization module which solves the optimization problem as per predefined objectives and constraints;
a node mapping module which decides the optimal placement locations for the virtual nodes of the virtual network on the physical nodes of physical infrastructure with the support of the optimization module;
a link mapping module which decides the optimal placement locations for the virtual links of the virtual network on the physical links of physical infrastructure with the support of the optimization module;
a reliability validation and negotiation module which validates the realizable reliability for the service with the reliability expectation for the service and is responsible for associated negotiation with user interface layer; and
a criticality analysis module which identifies the optimal location for improving overall reliability for the virtual network.

(Supplementary note 8)
A resource allocation method for communication service requests comprising:
deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests;
finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.

(Supplementary note 9)
The resource allocation method according to supplementary note 8 further comprising:
launching the service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in a deployment configuration template describing the mapping result and the resource requirement of each of the virtual network components received from the virtual network mapping unit.

(Supplementary note 10)
A resource allocation program for communication service requests for causing a computer to execute:
a service modeling processing for deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and
a virtual network mapping processing for finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.

(Supplementary note 11)
A reliability aware resource allocation system for cost efficiently satisfying reliability guarantees for communication service requests, comprising;
a service modelling unit which derives virtual network request comprising computing & networking resource demands, service quality demands which can ensure the expected performance for the requested service;
a virtual network mapping unit which finds the optimal mapping for the virtual network request on the physical infrastructure, wherein the virtual network mapping unit performs the pre-processing to allocate reliability goals to a plurality of virtual network components, finds the optimal mapping, takes necessary steps to satisfy the reliability guarantee for the virtual network request and generates deployment configuration template; and
a deployment unit which launches the service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in the deployment configuration template received from the virtual network mapping unit;
wherein, the service modelling unit receives the service request, derives the virtual network request and pass on to the virtual network mapping unit, which then pre-process it, finds optimal mapping, ensures reliability guarantee and forwards the deployment configuration template to the deployment means to finally provision the requested services.

Whereby the reliability aware resource allocation system by ensuring higher reliabilities for the relatively important virtual network components of virtual network minimizes the adverse effect on the running services in the event of failures. The reliability aware resource allocation system also ensure minimal mean time to recovery with minimal associated resource costs, and improves the acceptance ratio of the virtual network requests.

(Supplementary note 12)
A reliability aware resource allocation system according to supplementary note 11, further comprising;
a database for storing information about various virtual network templates suitable for different types of services. Wherein the information comprises of knowledge about different network functions, their interconnection and order, operational behavior, failure characteristics, recovery mechanisms, reliability. The database also stores the information about physical infrastructure, network topology, location, available capacities, reliability and availability values gathered by resource monitoring unit and resource discovery unit;
wherein the database is accessible by the service modelling unit and the virtual network mapping unit.

(Supplementary note 13)
A reliability aware resource allocation system according to

supplementary note

11 or 12, wherein the virtual network mapping unit further comprising;
a pre-processing unit that distributes expected reliability for the service among the virtual network components of the virtual network based on the relative importance of the virtual network components. Wherein the pre-processing unit decides the relative importance by calculating the weight value for each of the virtual network components by considering a plurality of factors. Wherein the factors comprises position in the topology such as shared among multiple flows or not, parallel or series, type of the virtual network component such as workload type, stateful or stateless, core or supplementary function, operational features such as cost for recovery, upgrades support, mean time to recovery, migration support.

(Supplementary note 14)
A reliability aware resource allocation system according to supplementary note 13, wherein the pre-processing unit further comprising;
a virtual network analysis module for analyzing the position of the virtual network components in the virtual network topology;
a data collection module to gather the information associated with the factors crucial for calculating the weight value for the virtual network components;
a reliability estimation module to estimate the reliability equation for the virtual network;
a priority assignment module to calculate the weight values for the virtual network components based on the factors; and
a reliability goal setting module distributes the expected reliability for the service among the virtual network components using the weight values.

(Supplementary note 15)
A reliability aware resource allocation system according to

supplementary note

11 or 12, wherein the virtual network mapping unit further comprising;
a mapping unit that performs optimal mapping of the virtual network on physical infrastructure such that all the requirements mentioned in the virtual network request received from the service modelling unit or the pre-processing unit are satisfied,
wherein the mapping unit performs node mapping and link mapping, and ensures minimal reliability gap between the expected reliability for the virtual network components and available reliability values of resources from physical infrastructure.

Whereby the mapping unit achieves optimal utilization of reliable resources by allocating just enough reliability and improves overall reliability for the virtual network by minimum additional resources and costs.

(Supplementary note 16)
A reliability aware resource allocation system according to supplementary note 15, wherein the mapping unit further comprising;
an optimization module to solve the optimization problem as per predefined objectives and constraints;
a node mapping module decides the optimal placement locations for the virtual nodes of the virtual network on the physical nodes of physical infrastructure with the support of the optimization module;
a link mapping module decides the optimal placement locations for the virtual links of the virtual network on the physical links of physical infrastructure with the support of the optimization module;
a reliability validation and negotiation module validates realizable reliability with the expected reliability and is responsible for associated negotiation with user interface layer; and
a criticality analysis module identifies the optimal location for improving the overall reliability for the virtual network.

(Supplementary note 17)
A reliability aware resource allocation method for communication service requests, comprising;
deriving a virtual network request for the communication service request comprising computing & networking resource demands, service quality demands which can ensure the expected performance for the requested service;
finding the optimal placement for the virtual network request on the physical infrastructure, comprising pre-processing to allocate reliability goals to a plurality of virtual network components, finding the optimal mapping, taking necessary steps to satisfy the reliability guarantee for the virtual network request and generating deployment configuration template; and
deploying the communication service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in the deployment configuration template.

(Supplementary note 18)
A reliability aware resource allocation method according to supplementary note 17, further comprising;
pre-processing the virtual network request to distribute expected reliability for the service among the virtual network components of the virtual network based on the relative importance of the virtual network components. Wherein the pre-processing step decides the relative importance by calculating a weight value for each of the virtual network components by considering a plurality of factors. Wherein the factors comprises position in the topology such as shared among multiple flows or not, parallel or series, type of the virtual network component such as workload type, stateful or stateless, core or supplementary function, operational features such as cost for recovery, upgrades support, mean time to recovery, migration support.

(Supplementary note 19)
A reliability aware resource allocation method according to

supplementary note

17 or 18, further comprising;
performing optimal mapping of the virtual network on physical infrastructure such that all the requirements mentioned in the virtual network request. Wherein the mapping stage involves node mapping and link mapping, and ensuring minimal reliability gap between the expected reliability for the virtual network components and available reliability values of resources from physical infrastructure.

(Supplementary note 20)
A resource allocation method according to supplementary note 19, further comprising;
sorting virtual nodes in the descending order of reliability expectations, finding optimal mapping for the virtual node with highest reliability expectation, recalculating reliability expectation for remaining the virtual nodes and iteratively perform the process until all the virtual nodes are mapped; and
assigning the weights to the virtual links connecting the virtual nodes of the virtual network, sorting virtual links in descending order of the reliability expectations, finding the optimal mapping and applying means to improve reliability to satisfy the expected reliability for the virtual link.

(Supplementary note 21)
A resource allocation method according to supplementary note 20, further comprising;
applying anti-affinity consideration to the node mapping step as per the predefined rules to avoid the failure of multiple virtual nodes of one virtual network.

(Supplementary note 22)
A resource allocation method according to supplementary note 21, further comprising;
estimating realizable reliability for the virtual network and comparing against the reliability expectations of a the virtual network and updating the virtual network configuration template with the optimized mapping that satisfies the reliability expectations.

(Supplementary note 23)
A resource allocation method according to supplementary note 22, further comprising;
modifying the optimized mapping by adding additional minimal resources for fault tolerance mechanism for bottleneck virtual network components so as to satisfy the reliability requirement for the virtual network;

(Supplementary note 24)
A resource allocation method according to supplementary note 23, further comprising;
applying criticality importance measure to identify the optimal location for improving the overall reliability for the virtual network.

Whereby the reliability expectations will be satisfied by utilizing minimum the physical resources and optimally assigning as per the reliability goals leads to accommodating more such plurality of virtual networks.

200 resource allocation system
21, 501 service modelling unit
22 database
23, 502 virtual network mapping unit
24 pre-processing unit
25 mapping unit
26 deployment unit
27 physical infrastructure
900 information processing apparatus
901 CPU
902 ROM
903 RAM
904 hard disk
905 communication interface
906 reader/writer
907 bus
908 reader/writer
909 Input/output interface

Claims

A resource allocation system for communication service requests comprising:
a service modeling unit which derives a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and
a virtual network mapping unit which finds an optimal mapping for the virtual network components on a physical infrastructure based on the virtual network request;
wherein the virtual network mapping unit sets reliability goals to each of the components by distributing the reliability expectation for the service among them based on their relative importance in the virtual network , and finds the optimal mapping based on the reliability goals, the resource requirements and the reliability expectation for the service.
The resource allocation system according to claim 1 further comprising:
a deployment unit which launches the service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in a deployment configuration template received from the virtual network mapping unit;
wherein the service modelling unit receives the service request, derives the virtual network request and pass on to the virtual network mapping unit, which then pre-process it to set reliability goals to the virtual network components, finds optimal mapping, ensures the reliability expectation for the service and forwards the deployment configuration template describing the mapping result and the resource requirement of each of the virtual network components to the deployment unit.
The resource allocation system according to claim 1 or 2 further comprising:
a database which stores template informations about various virtual network templates suitable for different types of services and information about physical infrastructure;
wherein each of the template information includes knowledge about one or more network functions belong to the virtual network indicated the template, such as their interconnection, order, operational behavior, failure characteristic, recovery mechanism, and reliability requirement, and the information about physical infrastructure includes network topology, location, available capacities, reliability or availability values at least, and
the database is accessible by the service modelling unit and the virtual network mapping unit.
The resource allocation system according to any one of claims 1 to 3, wherein the virtual network mapping unit further comprising:
a pre-processing unit which distributes the reliability expectation for the service among the virtual network components of the virtual network based on the relative importance of the virtual network components;
wherein the pre-processing unit decides the relative importance by calculating the weight value for each of the virtual network components by considering a plurality of factors, and
the factors comprises position in the topology, type of the virtual network component and operational feature.
The resource allocation system according to claim 4, wherein the pre-processing unit further comprising:
a virtual network analysis module which analyzes the position of the virtual network components in the virtual network topology;
a data collection module which gathers the information associated with the factors crucial for calculating the weight value for the virtual network components;
a reliability estimation module which estimates the reliability equation describing the relationship between set of the reliabilities for virtual network components and reliability for the service in the virtual network;
a priority assignment module which calculates the weight values for the virtual network components based on the factors; and
a reliability goal setting module which distributes the reliability expectation for the service among the virtual network components using the weight values.
The resource allocation system according to any one of claims 1 to 5, wherein the virtual network mapping unit further comprising:
a mapping unit which performs optimal mapping of the virtual network components on physical infrastructure such that all the requirements mentioned in the virtual network request received from the service modelling unit are satisfied;
wherein the mapping unit performs a node mapping and a link mapping, and ensures minimal reliability gap between the reliability goals for the virtual network components and available reliability values of resources from physical infrastructure.
The resource allocation system according to claim 6, wherein the mapping unit further comprising:
an optimization module which solves the optimization problem as per predefined objectives and constraints;
a node mapping module which decides the optimal placement locations for the virtual nodes of the virtual network on the physical nodes of physical infrastructure with the support of the optimization module;
a link mapping module which decides the optimal placement locations for the virtual links of the virtual network on the physical links of physical infrastructure with the support of the optimization module;
a reliability validation and negotiation module which validates the realizable reliability for the service with the reliability expectation for the service and is responsible for associated negotiation with user interface layer; and
a criticality analysis module which identifies the optimal location for improving overall reliability for the virtual network.
A resource allocation method for communication service requests comprising:
deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests;
finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.
The resource allocation method according to claim 8 further comprising:
launching the service comprising the steps such as instantiating servers, reserving bandwidth, setting up connections on physical infrastructure as per the specifications mentioned in a deployment configuration template describing the mapping result and the resource requirement of each of the virtual network components received from the virtual network mapping unit.
A resource allocation program for communication service requests for causing a computer to execute:
a service modeling processing for deriving a virtual network request that includes one or more resource requirements for each of a plurality of virtual network components and one or more service quality requirements which indicate a reliability expectation for the service at least, based on the service requests; and
a virtual network mapping processing for finding an optimal mapping for the virtual network components on a physical infrastructure based on the reliability goals set by distributing the reliability expectation for the service among them based on their relative importance in the virtual network, the resource requirements and the reliability expectation for the service.