Cloud computing building and management technologies
Network Functions Virtualization (NFV) is the concept of separation of network node function from hardware which implements it, using virtualization technology. NFV allows software-based deployment and management of network functions on a standart data center's hardware through the virtualization of physical resources (computing, network and storage). This ensures independence of the network service logic from the hardware on which it operates. The "Virtual Network Function" (VNF) engineering depends on the goals for which the virtualization infrastructure of network functions is built, on who builds this infrastructure and for what. Examples of NFV are services for the network traffic analysis, management and engineering. The sample of a virtual network function for telecom operators is a software that provides the functionality of appliance for a traffic switching, routing, filtering, balancing, etc. Other examples might be Voice over IP, video conference, EPC, billing, DPI (Deep Package Inspection), traffic engineering and monitoring, and the like.
Advantages of the NFV concept are the possibilities of dynamic scaling, automatic configuration of virtual network functions, efficient use of physical resources and their redistribution.
This research area covers a wide range of issues related to the cloud infrastructures organization:
- Virtual function's life cycle maintenance in the cloud infrastructure;
- Tools development for creating and editing of virtual functions;
- Validation and verification of virtual functions software;
- Planning of the cloud infrastructure resources according to the VNF orchestration policy;
- Integration of real-time monitoring systems for the implementation of VNF orchestration policy;
- Development of management algorithms for distributed cloud environments;
- Development of algorithms for synchronizing the state of resources between components of a distributed cloud;
- Building of a blockchain-based chargeback system for distributed cloud environments;
- Development of a micro-service cloud platform architecture based on lightweight virtualization;
- Development of Dark Data Analysis systems for analysis of cloud monitoring logs data in order to form a recommendation for the cloud operator (Self-Service Analytics);
- High-Performance Computing as a Service development (HPCaaS);
- Development of a system for automatic deployment of cloud software on a high performance infrastructure (HPC Cloud).
The main research directions in the NFV area are as follows:
- Development and implementation of the interdomain interaction management system in a distributed NFV platform. The goal is to organize cloud's components interaction in a distributed NFV platform;
- Verification and validation algorithms for virtual network functions. The research’s aim is in developing methods and tools that allow to accelerate, simplify and make more secure the process of network functions transfer to the NFV platform;
- Efficient use of physical resources by lightweight containers in cloud systems. The aim is to build a monitoring system that allows efficient planning and replanning (in the case of migration) of the containers' location in the cloud.
Romanov A.R. Antonenko V.A. Development of the system for provisioning of reliable and scalable virtual network service in the cloud environment (in Russian). https://istina.msu.ru/publications/article/28174140/