On March 7th, Miguel Ponce de Leon, John Day and myself gave an overview of RINA, the motivations behind it and an status of the prototyping efforts at the University of Kaiserslautern, as part of a Ph.D course on “Future Network Architectures and Experimentation“. Miguel went through an introduction to RINA, similar to the one given by John at Barcelona, and John discussed about the misconceptions about layering in networks, exposing all the troubles these misconceptions have caused.
I presented the plans for RINA prototyping and the current status of the Java prototype that is being developed based on the TINOS network experimentation framework. The objective of the prototype is twofold:
- Allow the validation of the draft RINA specifications that are currently being worked out. Provide feedback to the specifications about details that are hard to sort out without the level of detailed thinking that usually requires an implementation.
- Once it is complete enough, enable interested researchers to try their policies for the specific parts of RINA they are interested in (authentication, routing, resource allocation, data transfer, flow control, …)
Our initial prototyping goal is to implement a DIF that can provide an equivalent IPC service to what TCP/UDP + IP do together; with all the advantages that RINA provides. The initial prototype is an overlay to IP. UDP/IP can be seen as an underlying DIF, with the advantage of having a well defined, common interface (sockets) with SDKs available for most of the programming languages. IP is also basically universal nowadays, enabling a lot of different deployment scenarios that are easy to setup. Commercial RINA products
over UDP/IP (such as completely customizable VPNs) can be interesting and increase the incentives for RINA deployments at the lower layers. The RINA prototype treats IP as what it is: a network interface technology – not an internetworking technology -, essentially playing the same role as Ethernet. Summing up, although a RINA deployment over IP over Ethernet has unnecessary overhead, it is an effort-effective way for the initial prototyping efforts to develop the DIF machinery and create a valuable result that can encourage the investment of further RINA deployments in the lower layers, as shown in the next figure.
If you want to follow the evolution of the prototype and/or know more about it, you can take a look to the prototype wiki at github. We expect to have a first complete version of it by the last quarter of 2012. Follow the DANA blog for future information!