HEAnet sees value in BoD trial
5 December 2011 | First published in Connect
Irish NREN a key participant in the 12 month pilot, seeing great value in automatic provisioning
Multi-domain Bandwidth on Demand (BoD) in September entered the second stage of its 12 month pilot that originally started in June 2011. During this phase, pilot participants will be investigating the operational aspects of providing a BoD service and testing the multi-domain BoD service process and procedures. Connect spoke to Brian Nisbet of HEAnet to better understand how they see involvement in the BoD pilot to be of such importance.
HEAnet is participating in the GÉANT community's multi-domain bandwidth on demand service pilot programme. As NOC Manager, what are your reasons for doing this?
National Research and Education Networks must, by their very nature, explore new technologies and new ways of connecting those who use the network. Bandwidth on demand services offer a way of better connecting users across networks within a controlled and managed framework, allowing for potentially better use of network resources and more straight-forward troubleshooting should things go wrong. This is particularly true in relation to easy to provision Layer2 circuits.
What advantages do you think it gives your team?
Managing dedicated connectivity across multiple domains is no easy feat. Building a BoD system that uses provisioning systems has several benefits. If the links are built automatically it greatly reduces the chance of manual configuration errors and potentially allows for circuits to be provisioned without any interaction from the Operations team. Manual provisioning also hits scaling errors very quickly, for instance, we'd never be able to keep up with the Internet if we used static routes instead of BGP. Finally, it allows for easier troubleshooting and the ability to quickly remove BoD circuits should they impact on general network usage.
What do you see as the challenges to deployment?
Our biggest challenge in all of this was trusting the provisioning system to build circuits on orders from another network. This is a challenge that HEAnet has surmounted, but other networks will face the same issues and building that trust will be a very large piece of work.
Operations teams are very busy - why are you willing to invest time in this type of service?
We see a long term benefit in the BoD service, both in directly providing the service and in improvements it will drive in provisioning systems and operational procedures in HEAnet and across the GÉANT network. However ultimately any service to which time is devoted needs to be one that our clients and end users will avail of. At this point in time there is sufficient interest in the service to justify the time investment. The BoD service alone is not the only reason behind our provisioning system, of course. Once we had that system in place we reaped the mentioned benefits immediately and the BoD service piggy-backs on it in quite a lightweight manner.
How do you envisage your users benefitting from this type of service?
The ability to quickly and easily set up dedicated links across networks should enable research projects that need to exchange data, either in a particular fashion (securely, using experimental protocols etc.) or in a way that would not be compatible with other network users.
That said, most of the demand we've encountered seems to benefit from the direct layer 2 connectivity rather than explicit bandwidth reservation, so we'd like to see effort put into a convenient and cheap circuit reservation system which doesn't require expensive reserved blocks of bandwidth.
Are there advantages to early involvement in this project?
Early adoption of or involvement in any project allows HEAnet to influence the design of the service and the operational procedures. It also gives us a head-start in learning how to operate the service, leading, hopefully, to less ramp-up work when it moves into more general production.
For more information, see:
Bandwidth on Demand