‘The Future Global Network’
Have you ever wondered how the future Internet will affect our photonics field – or how our photonics field can impact the future Internet? Have you ever wanted to learn more about the computer science networking field, which provided the foundation for the Internet? This summer at the IEEE Photonics Society summer topical meeting on The Future Global Network we brought together leading researchers from the photonics and networking communities to share their visions of the future network from their very different technical perspectives. Leading figures were invited from academe and industry, with diverse backgrounds and interests covering the range from photonic devices and systems to Mega-Data-Centers and time critical networking, providing excellent presentations that both introduced topics and provided visions for the future. To reach a broader audience the presentations have been placed online and can be downloaded from:
Historically, telecommunications networks, e.g. telephone, and data communications networks, e.g. the Internet, developed separately while maintaining their own requirements reflected in architectures, technologies, network management, protocols, and services. We are currently seeing the rapid convergence of telecom and datacom worlds into a new, lower cost network, using IP (Internet Protocol) to support all applications (communications, information, entertainment) on any physical layer technology (wireless, optical, wired). While IP based convergence is taking place at a rapid pace, there are issues due to the fact that IP was originally designed as a ‘best-effort’ protocol without concern for quality-of-service, security, and global management support.
Worldwide research communities are now focused on the future, beyond this converged network, taking the opportunity to step back and imagine how a future global network might be created – using a ‘clean slate’ mentality, i.e. how would I design the future network from scratch, to support current and future applications and services, taking advantage of new and emerging physical technologies, creating new network architectures and paradigms? Future global network research requires disparate communities to work closely together to create a vision, taking advantage of new concepts and technology advances in all fields.
This meeting brought together researchers from the optical communications and computer science networking fields to discuss areas key to the future global network. It provided a forum for sharing essential concepts for the future network from both communities, to reach a mutual understanding of the true problems in each research arena, and to then move collectively in the right direction. Invited talks from leaders of each community provided an excellent opportunity to learn about key concepts and technologies that will lead to a future global network. Another goal of this meeting was to bridge the ‘cultural divide’ between the two communities, building a common language and relationships on which cooperation and collaboration could occur.
Topical Meeting Presentations
The meeting ran for three days, with 10 technical sessions addressing major topics from photonics and networking, relating to the future global network. The meeting also included two ‘hot’ topics as panel sessions in the evenings, at which full rooms and spirited discussions were found. The organizers would like to thank Stu Elby and Verizon for sponsoring these panel sessions, providing good food and drink for all attendees. A decision was made early on to include mostly invited presentations, in light of the goals for the meeting to bring together key photonics and computer science networking people to both educate each other and to start to collaborate with each other – these invited presentations are briefly described in the following paragraphs. A majority of these presentations have been provided by the authors for distribution on the website referenced earlier – we would like to thank the presenters for providing their presentations, and Guru Parulkar and Saurav Das at Stanford University for hosting this site.
The meeting opened with a session on Future Global Network Visions, with two talks looking at the big picture of how these new networks may look. Ty Znati (National Science Foundation) led off the meeting with an excellent presentation from the computer science networking viewpoint of research towards a future network, in which he described complexity and some of the scientific approaches to modeling and understanding complex networks. Ty described some of the rethinking underway of the architecture for a future Internet, and also the need for an experimental infrastructure to try out technologies that may become part of the future Internet. One such infrastructure being funded by NSF is part of the GENI program (www.geni.net); other opportunities for photonics researchers to obtain funding from the NSF CISE Directorate (www.nsf.gov/dir/index.jsp?org=CISE) were also described. The second talk by Hisaya Hadama explained the design concepts behind NTT’s future carrier networks, based on the paradigm shift that in the future network every user can obtain customized services to meet their needs – which will be dynamically reflected in the network.
The 2nd session, on Optical/IP Networking Architectures, included two talks covering the integration of networking and optical technologies. Kristin Rauschenbach described three projects, GENI, CIAN and CORONET, and explained the higher-layer network requirements for optical innovation. Ken-ichi Sato then provided a view of the future optical network, focusing on video and multimedia communications with end devices operating at higher and higher resolutions. He described how rapidly increasing bandwidth requirements provide the need for hierarchical optical path networks and waveband processing. This was followed by enlightening talks (especially for photonics folks) covering Impacts of Applications and Cloud Computing on the Future Global Network in Session 3. These two networking talks described first what cloud computing and Mega-Data-Center’s are (Amin Vahdat), and also Donn Lee described the infrastructure of a Facebook Data-Center; they both described issues with networking large data centers and also how these facilities and the cloud computing they perform will impact the shape of the future Internet. Session 4 on Photonic Integration and Impacts on the Future Global Network provided excellent talks on photonic integration by Dan Blumenthal and Tom Koch. Dan described how power dissipation is a key driver in how we build networks, and he showed the tremendous progress made using photonic integration for optical switching and routing components that may some day replace power hungry electronic switches and routers.
Day one finished with a panel session on Energy Usage and “Green” Networking, a hot topic at present, both from a technical and political viewpoint – which led to some excellent discussions. Thought provoking talks at this session came from a government agency (Bruce Nordman), a University (Ben Yoo), and two major industrial companies (Jan Soderstrom and Andreas Gladisch).
Day two started with Session 5 on Network Control and Management, with an interesting talk describing network virtualization from Aki Nakao, followed by a talk on optical network control challenges by Souichirou Araki, which focused on GMPLS in current and future networks. Session 6 provided very interesting viewpoints, Commercial Views of the Future Internet, from a service provider (Stu Elby – Verizon) and from a major networking equipment supplier (Flavio Bonomi – Cisco). Stu explained that the network is becoming too complex and costly, suggesting solutions such as collapsing functions into fewer boxes, automating the network, and where possible using technologies that track Moore’s Law. Flavio’s talk focused on dynamic and flexible WDM transport, looking towards more automation and adaptive WDM networks.
A strong session on Optical Transmission Technologies included a talk focused on the evolution of optical transport by Peter Winzer, in which he described why optical networking should be considered an integral part of the future Internet. The following talk by Andrew Ellis described adaptive modulation schemes, with a focus on multi-carrier systems such as Optical OFDM. Session 8, on Cross Layer Research, started with a talk by Guru Parulkar describing ‘Openflow’, a novel architecture that unifies circuit and packet networks over wired, wireless and optical technologies. Guru described how many of the limitations of the current Internet are set within the infrastructure, and must be changed to allow innovation of the network – which Openflow provides. Ilya Baldine talked about cross-layer networking, explaining how cross layer support is required to make the optical layer dynamic, and describing BEN – the Breakable Experimental Network. At the end of the second day, the panel session (When) Will Commercial Packet Switching Happen? provided spirited discussions between proponents and non-believers of optical packet switching, including enlightening presentations by Masataka Ohta, Keren Bergman and Loukas Paraschis.
The final day started with a session on Time Critical Networking and Applications. Edward Lee discussed time critical networking, describing cyber physical systems in which networked computational resources are orchestrated with physical systems to meet an application requirement – e.g. factory automation. He explained that in order for time critical systems to work correctly, timing needs to be included as part of the network semantics (synchronization, latency) rather than try to add it afterwards. Larry Smarr described the OptIPuter project, which enables user-controlled 10 Gbps dedicated lightpaths to access global data repositories, scientific instruments and computational resources from OptIPortals – PC clusters providing large scale visualization and collaboration facilities around the globe. Larry described applications of this technology in a number of different realms. The final session of the meeting focused on Global Research Initiatives and Test-Beds. Dimitra Simeonidou discussed large scale optical networking test-beds in Europe, including the JANET Aurora UK Network; describing how this network will be used to capture and multicast 8K content from the 2012 Olympics around the UK and Europe. Dimitra described the operation of the JANET Aurora network, as well as the PHOSPHORUS European Network Test-bed for high-performance applications. The final talk, by Hiroaki Harai, described NICT’s ‘New Generation Network’ based on the AKARI architecture design; using a ‘clean-slate’ design with novel network technologies to provide a new infrastructure for the next two or three decades.
The Future Global Network meeting was a great success thanks to the hard work of the co-chairs, committee members, and the superb presentations provided by invited speakers and contributed papers. We hope you enjoy the online presentations, use them to learn about research areas outside of your own, work to build collaborations with researchers from these other fields, and become part of the global effort to define the future global network.