PRISTINE’s deliverable D2.1 is already available from the website’s deliverables section.
The goal of this document is to analyze the use cases that will shape the requirements of the research, development and experimentation activities that will be carried out by the PRISTINE partners. Each use case provides an opportunity to explore and understand how RINA can be applied in a real-world scenario, as well as to assess the advantages it provides over existing alternatives. The main outcome of D2.1 is a set of draft designs of RINA-based solutions for the Distributed Cloud, Datacenter Networking and Network Service Provider use cases. Each design specification includes the type of DIFs in the scenario, the requirements each DIF has to fulfill and the relationships between DIFs; the requirements for the Network Management System and the requirements for supporting tools and subsystems that allow RINA to interoperate with current technologies and applications.
PRISTINE use cases are based on three well-defined scenarios, targeting different realistic deployments of RINA interoperating with some of the current computer networking technologies. The following paragraphs provide a short summary of these scenarios.
Distributed Cloud. SlapOS is a decentralized cloud technology used to build a physically distributed cloud. Customer’s applications are run in traditional datacenters, but also in servers from offices and home users. SlapOS is in charge of managing the overall cloud from a logically centralized location: the SlapOS master (a distributed approach is currently under development). The SlapOS master controls the different computers running SlapOS slaves. In terms of networking, the master and the nodes at different locations are interconnected through multiple IPv6 providers. In order to guarantee a high reliability (99.999%), SlapOS uses an overlay called re6st, which creates a mesh network of OpenVPN tunnels on top of several IPv6 providers and uses the Babel protocol for choosing the best routes between nodes. PRISTINE will provide an alternative to the re6st overlay, by using RINA.
Datacentre Networking. The datacenter space is one of the areas that has seen more virtual networking innovations during the last few years, fueled by the flexibility requirements of cloud computing. A myriad of SDN-based virtual network solutions, usually providing L2 over L3 or L4 tunnels and a control plane, are available in the market (VXLAN, NVGRE, STT, etc). PRISTINE will investigate and trial the use of RINA-based solutions for datacenter networking. Important issues to be addressed in a datacenter environment are the mobility of Virtual Machines to allow an efficient utilization of datacenter resources as well as high reliability; multi-homing support; guaranteeing the level of service in inter-data center communications and flexible allocation of flows supporting computer and storage resources. RINA provides an excellent framework to tackle these issues, and the PRISTINE project will exploit them as explained in task T2.1.
Network Service Provider. The goals of this scenario are to investigate and trial the benefits of the use of the RINA technology by a Network Service Provider (NSP), and to analyze RINA as a essential component of the Network Functions Virtualization (NFV) concept within an operator network. It is obvious that a disruptive, clean-slate technology like RINA would have a difficult way within a NSP environment, very much oriented towards service provisioning and stability, and the phased incorporation of thoroughly tested technologies. But the recent advent of the NFV proposal opens a very interesting window of opportunity for demonstrating RINA in the NSP environment, and showcase the advantages that the combination of both approaches (NFV and RINA) can bring to network service design, management, and operation.