JANET Optical Briefing
The JANET Optical briefing took place at the Institute of Physics in London on February 8, 2010. It brought together experts in optical networking from both the academic and commercial sectors to provide a fascinating peek at the future for the education and research community. Two recurring themes emerged during the presentations: First was the control plane, in particular the importance of measurement and monitoring to provide clear evidence that the automated systems are doing what they are supposed to do; Secondly the potential of ROADM technology. This report doesn’t cover in detail each of the talk but instead tries to capture the theme and the discussion. The slides will be available on the JANET web site.
Jeremy Sharp (JANET) opened the meeting with some background. He highlighted the recent tests with 100 Gbps connectivity. He noted that the contract for Super JANET 5 will be extended to October 2013 and significant effort has been to ensure enough capacity and additional engineering to grow. Jeremy also spoke about future opportunities for JANET, namely the JANET 6 network. JANET are working with JISC and the funding bodies to secure funding. He also mentioned the research requirements workshop in June. Jeremy chaired the morning session.
David Salmon (JANET) spoke about the requirements for an advanced optical infrastructure from an NREN perspective. David’s presentation explored why JANET, as an NREN, and the education and research community in general are interested in the latest advances in optical infrastructure. He looked at our customer profile compared to the drivers for commercial providers. David mentioned that Super JANET 4 was built on leased fiber but Super JANET 5 provided access to the fiber itself, as will JANET 6. While this level of control brings flexibility to the services that JANET can offer, David highlighted the significant cost reduction that this transition has also brought. David concluded that there are a wide range of interesting possibilities and practical concerns. A question was asked regarding ROADMS. The observation was made that a few years ago, alien wavelengths were discounted, but now it seems we are saying it is the future. David responded that the technology is simply better than before. He went on to say that it might be possible to reserve (say) 60 wavelengths at the bottom for production, and top 20 for research, perhaps exotic research. It is a speculative idea, and maybe expensive, but it can be done. ROADMS have features that can be advantageous. What has happened as ROADMS have evolved, is that they have become sophisticated and flexible devices.
Nigel Baker (NEOS Networks) spoke about the requirements for an optical infrastructure from a telecommunications provider perspective. He said their thinking is dominated by costs. It is a highly competitive industry and they are driven by need to make a profit. Nigel expanded on some of David’s comments. He spoke about keeping an eye on major data centers for 40G and 100G connections. He focused on protection and restoration. Fast restoration depends on utilizing the control plane. Networks are coming to a situation where costs and delays are due to human intervention, there is a clear need for autonomous operation. A key point was the observation that network engineers are loathed to give up manual control. However, when their engineers began t make use of control plane technologies, it worked well. Nigel said the key is measurement and monitoring, i.e. engineers need to see it is doing what it is supposed to do.
Mark Gibbon (Nortel Networks) gave an interesting talk about High transmission rates – a review of 100 G. He explored some of the issues involved in moving from 40G to 100G. He spoke about agile photonic networking and the photonic domain control. He spoke about PMD distribution and work published at the Optical Fiber Communication workshop. There was a question about the limitations on number of ROADMS. Mark responded that the specification would suggest the limit is 20 but realistically it network isn’t going to be limited by this.
After lunch, the afternoon session was chaired by David Salmon.
David Boyle (Ciena) spoke about the state of the art optical networking. He talked about switching at 100G line rate, both packet and TDM. He covered some details regarding OTN and WDM and spoke about Ciena layer 1 control plane for restoration. He added that there has been lots of work at layer 2 but work needs to be done at layer 0. He spoke about domain management planning and design tools and how this can be used as feed-back for the control plane. He went on to review operational issues with control planes. He spoke about the real efficiencies and benefits and echoed Nigel’s comments about the concerns of network engineers to give up manual control. David discussed the generic functions of the control plane, namely Discovery, connection/teardown, maintenance/operation, protection and restoration. He again highlighted that Layer 1 had GMPLS but Layer 0 has nothing. The concluded by mentioning that the control plane needs to understand optical characteristics of the fiber. This is where the work is going on.
Eduard Escalona (University of Essex) spoke about network provisioning in optical networks. He complemented Nigel and David’s perspective with a review of activities within the research and education community. He highlighted operational costs again. He reviewed the activities of the DICE group, consisting of Dante, Internet2, Canarie, ESNET – the research and education networks for Europe, the US and Canada. He also reviewed the requirements for control plane system. He listed: Fast provisioning of bandwidth on demand, scheduled reservations, automatic resilience, and automatic topology discovery. Eduard went on to describe a goal for the Autobahn software to be deployed on 4 or 5 European NRENs by March 2011. Autobahn is the multi domain glue for local provisioning systems. He also discussed the Phosphorus test-bed for Grid network services, and the related Harmony and now GEYSERS projects. He also mentioned the OGF NSI WG as the organization overseeing the development.
John Colton (Lucid Optical Services) spoke about fault finding and trouble shooting on optical networks. John’s talk was a fascinating insight into challenges that most users wouldn’t think about. He extended some of the points mentioned in David Boyle’s talk regarding characterization of fiber. He added that careful testing is critical. The danger of poor testing is that it misses real faults, and find faults that don’t exist. People cause problems so keep people away from the fiber as much as possible. John reviewed traditional characterization (<1991), the current best practices, and how characterizations might be conducted in the future. He laid out a comprehensive plan for accurate and detailed characterization but pondered if there was real demand. John also mentioned his other hat with the Fiber Optic industry Association. They are developing a statistical approach to characteristics, partly in a response to unrealistic requirements.
Alwyn Seeds (University College London) concluded the presentations. He spoke about new technologies for optical networks. He covered the advances made by several UK research groups in the area of fast optical switching. He highlighted research published at several conferences and mentioned AURORA, the photonics research network across JANET.
In conclusion, the briefing consisted of a fascinating series of presentations which provided an overview of the current status. The complementary views from the education, research and commercial sectors gave the participants a peek into the direction of advanced optical networking that will aid their planning and decision making. David closed the meeting. He thanked the speakers and participants and said he looked forward to seeing everyone again.
Warren Matthews - JANET Optical event - 8th February, London