Friday, February 11, 2011

A new generation of Research and Education network architectures

[The future of R&E networks around the world is always under constant review.
Some argue that the role of R&E networks should only be as a low cost Internet service provider to the R&E community. Others counter that R&E networks should focus on providing services to support eScience and perhaps even integrate with other eInfrastructure providers such as HPC and grid. And yet still others argue that R&E networks should be working closely with industry by providing testbeds to develop and or improve industry products and enable commercialization of university R&D.

Although helping industry develop new products or enabling commercialization through academic R&D testbeds is a wonderful concept it is very difficult to implement in practice. Industry works on very short time scales and is very focused on developing relationships with future or current customers. Unfortunately they rarely see higher-ed academia as an important customer or an early adopter. Similarly merging R&E networks with other eInfrastructure facilities such as grids and HPC poses its own set of problems, as generally networks are much more production oriented and serve  the much broader community of “education” as well as research. It is also important to note that HPC and grids generally represent only a small fraction (but high end fraction) of the research computing on campus. The vast majority of research computing (about 80%) is done on small clusters in a variety of disciplines and is increasingly moving to clouds. R&E networks have an equal responsibility to both of these eInfrastructure communities.

In my opinion R&E networks I think can play far more important role, first in supporting eScience, but also in helping industry and creating a knowledge society by being an innovative “disrupter”. This is where R&E networks have been hugely successful in the past: first in the build out of the original Internet, next in deployment of low cost user owned fiber networks, and more recently in areas of new architecture for low carbon Internet networks and global authentication schemes. These disruptive developments were first intended, in many cases, to support the needs of science, but also had a beneficial effect of creating new network business models and enabling knowledge transformation of society as a whole.
I am pleased to see that we are now on the verge of another disruptive change with respect to R&E networks. SURFnet and the global high energy physics are once again at the forefront of these developments through the deployment of open lightpath exchanges and federated networks. For a perspective on this future vision of R&E networks  I encourage the reader to look at this presentation by Kees Neggers of SURFnet.

Another excellent perspective is by Harvey Newman from Caltech in this recent presentation he gave at CHEP 2010.

Once again while these network transformations are first being driven by the needs of eScience the network architectures are starting already to have a beneficial effect on broadband architectures in general. A good example, of course, is the unique facilitation role that Internet 2 role is playing in the roll out of national broadband through its partnership in UCAN. Other examples include the deployment of community transit exchange points by BCnet and peering points by KAREN in New Zealand.

When I was offered the chance to do some consulting work with SURFnet I jumped at the opportunity, as SURFnet is not only at the forefront of many these architecture developments that I also championed, but has been pushing the disruptive innovation envelop in a host of other fields such as national wireless Internet, an exciting new program in Green IT, collaborative platforms, enabling commercial clouds, etc. Indeed they have an entire team dedicated to innovation called SURFnet / Kennisnet Innovation Although SURFnet only receives a small amount of government funding for their innovation program ( the bulk of the funding comes from their members though the innovative services they offer) they continue to set the global agenda for many of these developments in R&E networking –BSA]

Green Internet Consultant. Practical solutions to reducing GHG emissions such as free broadband and electric highways.
twitter: BillStArnaud
skype: Pocketpro