Adaptive Trust Management In NFV

In the era of mobile Internet,in addition to the competition with the Internet companies,the closed and rigid telecom network infrastructure adopted by the operators is one of the important factors which restrict the further development of business model innovation and technology update of the operators.The basic network elements adopted by the operators are the devices that couple with hardware and software,which reduces the flexibility of upgrading network functions or hardware devices and causes expansive maintenance cost in the future.A wide variety of and poor compatibility of network equipments make the operators hard to choose or update single or some specific equipments.As a result,the operators need to replace the whole set of equipment to upgrade network equipment.Besides,in the original infrastructure,underlying network resources and business providing are isolated.Without efficient coordination between them,business cannot make good use of underlying network resources on demand,wasting plenty of network equipment resources.In order to solve the above problems,a new technology Network Function Virtualization(NFV)emerges.NFV implements many network functions of network hardware devices in a way of software by leveraging standard IT virtualization technology.NFV decouples hardware and software and changes the way of constructing a network by network operators and provides the opportunity to create a new type of eco-system.NFV encourages and supports the rapid innovation,reduces cost and overhead for the operators.When introducing flexibility,NFV also brings some problems and challenges: such as Quality of Service(Qo S)of VNFs(Virtualized Network Functions),allocation of underlying resources,security of VNFs,security of NFV architecture,stability of NFV network and so on.At present,some researches focused on effective allocation of underlying resources and Qo S of VNFs in NFV.But most researches ignored the interaction relationship between Qo S of VNFs and allocation of underlying resources in NFV and cannot ensure good Qo S of VNFs and effective allocation of underlying resources at the same time.The improvement of Qo S of VNFs will make VNFs take up more NFV underlying resources,affecting Qo S of the other VNFs,and the limited NFV underlying resources will affect the Qo S of VNFs inevitably.How to allocate resources to VNFs with different priorities reasonably and dynamically to ensure that Qo S of VNFs satisfy the Qo S threshold in the condition of limited NFV underlying resources is an important problem.This thesis proposes an adaptive trust management mechanism to allocate underlying network resources effectively and ensure the Qo S of VNFs in the meantime.It mainly includes three aspects: Qo S evaluation of VNFs,adaptive trust management procedure,resource allocation algorithms.Qo S evaluation of VNFs illustrates the relationship between Qo S of VNF and QAs(Quality Attributes)that affect Qo S of VNF,the resources assigned to VNF and evaluates the Qo S of VNFs according to collected data about VNF QAs performance.Adaptive trust management procedure describes the execution relationship and procedure between Qo S evaluation of VNFs and the resource allocation algorithms.The resource allocation algorithms allocate underlying resources to VNFs with different priorities effectively and dynamically.We implemented four VNFs with the Click element components,a QAs monitoring module,a Qo S evaluation module and a resource allocation module in Linux.Furthermore,we tested the QAs monitoring module,the Qo S evaluation module and the resource allocation module in four different conditions of VNFs running environment.The results show that our proposed mechanism can allocate underlying resources to VNFs reasonably and dynamically.The Qo S evaluation module can evaluate the Qo S of VNFs in real-time,the resource allocation algorithms can allocate underlying resources to VNFs reasonably and dynamically to ensure that Qo S of each VNF can meet the requirements of the threshold.