Internet Technologies of the New Generation

researches in the area of computer networks and internet technologies in Russia, developement of the most breakthrough computer networks technologies of the new generation

Cloud computing in a heterogeneous data center environment

The main purpose of this direction is the development of special tools to build virtual information infrastructure on the resources of geographically distributed data centers network, based on SDN technologies, and to support highly specialized staff activities, in order to significantly improve the effectiveness of those activities.

In the framework of this direction, ARCCN specialists work on the development of the following solutions:
  • a software platform, which provides basic functions to collect information on the state of data center network resources, control applications performance, transfer of control actions onto network switches;
  • set of control applications, responsible for routing, load balancing, resource allocation optimization, compliance with security policies, ensuring a given level of quality of service (QoS), and docking with networks of traditional architecture;
  • logically centralized control platform, which provides a consistent understanding of the state of data center applications network resources on a single node, as well as on several geographically distributed nodes;
  • tools for constructing virtual information infrastructure, including computing, networking resources and storages, for research in various domains of knowledge.

The development of these solutions will enable substitution of imported network equipment (more than 90% of the equipment today is foreign). Instead of expensive proprietary routers for network management, we would use conventional servers with the widespread use of domestic software.

The main research topics under this direction:

  • Development of a system for network functions debugging
Purpose: to expand the specification of the virtual network function template adding debugging mechanisms.

Relevance:
The number of virtualized network functions will continue to grow. It is expected that the complexity of these functions will also increase. Assuming that VNFs cease to be independent of each other, the developers will need to build VNFs based on other VNFs. In this case, it is relevant to build VNF debugging mechanisms, similar to throw-catch blocks in OO programming languages.

  • Development of a system for network functions upgrades
Purpose: Extend the specification of the virtual network function template by adding mechanisms to upgrade virtual networking functions. 

Relevance:
Following the logic of NFV platforms development, we can assume that VNFs will be developed by third-party companies. NFV platform serves as a platform for services, while such a platform must have upgrade mechanisms for already deployed services, offering developers a variety of upgrade options.

  • Development of a hybrid network services orchestrator
Purpose: suggest an algorithm for hybrid orchestration of network services

Relevance:
Virtualization of network services means providing a service as a software, operating on one or more virtual devices: a virtual machine, a virtual network device, a virtual container, or various combinations of these devices.
One of the most important areas to apply service-oriented infrastructure is the technology of virtual network services.
Monolithic applications are replaced by systems, consisting of loosely connected virtual network services that can run anywhere and in isolation from each other. It is necessary to support the full lifecycle of a virtual network service from the stage of service registration to the stage of data center resources release, supporting different virtualization systems.

  • Development of an adaptive scheduler for network services
Purpose: Development of adaptive scheduling algorithms based on network services orchestration policy

Relevance:
With the development of cloud computing, emerged a question of building cloud recourses management systems. One of their purposes is to make the use of resources more effective. This problem can be solved using different resource allocation algorithms. This project is focused on development of a resource scheduling algorithm to schedule all types of resources (virtual machines, virtual data storages, virtual communication channels) based on network services orchestration.

  • Development of a distributed storage system and system components of the NFV cloud platform
Purpose: Suggest the NFV platform architecture based on one of the open source implementations of a distributed file system.