Global Network Architecture Technical Group
Greg Bell (ESnet), Erik-Jan Bos (NORDUnet), Rob Evans (JANET), Gustavo Garcia (RedCLARA), Xing Li (CERNET), Inder Monga (ESnet), Edward Moynihan (Internet2), Bram Peeters (SURFnet), Michael Stanton (RNP), Mian Usman (DANTE), Rob Vietzke (Internet2), David Wilde (AARNet), James Williams (Internet2)
The charge of the Global Network Architecture group from the NREN CEO Forum is to draft a blueprint for the intercontinental R&E Network interconnects, based on the latest technologies and promising developments, with a ten-year horizon. This blueprint will enable participating R&E Networks to align their spending for intercontinental bandwidth. The blueprint is all-inclusive, and has the potential to take along all regions of the world if and when they are ready to embark, bringing intercontinental network capacity as much as possible in line with what is available within the continents.
By joining forces the R&E Networks can be bigger players in relevant markets -e.g. the intercontinental transmission market- than today where spending is less directed. Hence, the R&E Networks can be much more ambitious and create a more far reaching result, yielding a sustainable high-performance global interconnect for Research & Education. By joining forces, the leading R and E Networks have the potential to bring this Global R&E Network infrastructure within reach.
An important piece of the vision is that the R&E Networks that now subscribe to the GNA strategy agree to bring their procurements in line with the overall architecture, as much as possible. First, this makes it possible to bring pieces of the infrastructure together early on. Second, it allows the NRENs -and their funders- collectively to understand and contribute to a bigger picture.
The GNA embraces open standards, such as those ratified by the IETF, the OGF, the IEEE, and other globally recognized standards bodies, as much as possible. Also, the GNA adheres to the resource management policies for IPv4 and IPv6 addresses, autonomous system numbers, domain names, etc., and collaborates with the authorities of these resources.
The move to this Global Network Architecture, once adopted and embraced, will of course not happen immediately. We expect that we will see the first instances of the outlined architectural elements emerge in the areas where critical ingredients such as bandwidth are most affordable and available. We expect other regions to follow, when feasible, and to leverage the results and experiences of the first implementations.
The starting point for the GNA group was the document prepared for the NREN CEO Forum titled “Global R&E Network Architecture Program” authored by Prof. Jianping Wu (CERNET) and H. David Lambert (Internet2), dated April 30, 2013. In this white paper, the authors presented the rationale for R&E Networks collaborating on the intercontinental R&E network infrastructure and argued that by collaborating the R&E Networks can move themselves into a much stronger position for realizing a next generation global network infrastructure for R&E.
As directed by the NREN CEO Forum, the GNA challenge group has, starting with a face-to-face meeting in Maastricht, The Netherlands in June 2013 and a subsequent series of video conferences, further sharpened and detailed this starting point, resulting in a high-level description of a Global Network Architecture for R&E.
The GNA Technical Group has recently agreed to create three subgroups:
- Network Services
- Operations and Security
- Technology and Topology
Each subgroup has a clear mission to compile a white paper that contributes to the overall vision of the GNA. Additionally, the group is working on a questionnaire soliciting further input from the R & E Networks around the globe.
The discussions inside the GNA group have led to a global network architecture model that consists of a powerful intercontinental transmission substrate, consisting of
- Global Open Exchange Points ("GXPs"),
- High-bandwidth transmission pipes (running between GXPs).
The GXPs in this architecture model have two major functions:
- Attachment points for the high-bandwidth circuits of the substrate
- Connection points for the Regional and National R&E Networks that are part of and make use of the GNA infrastructure.
We envision at least two GXPs per continent or region that are placed in consultation with the region's R&E Network organizations, taking the intercontinental fiber paths into account.
The GNA substrate supports the creation, modification, and deletion of overlay networks. In the first instance, this is expected to be manual. Over time, it is expected that this can be fully automated.
Examples of permanent overlay networks include:
- General purpose IP routed interconnect, supporting IPv4 and IPv6, both for unicast and multicas
- Commercial Peering Services
- LHCOPN and other Optical Private Networks
- LHCONE and other Open Network Environments
Examples of ad hoc overlay networks include:
- Science instrument support networks like for e-VLBI that are only needed during a science run
- Virtual Organization networks for large multi-continent science collaborations likeLHC, ITER, Climate, or agencies like NIH.
Three things are being achieved by this architecture: Connectivity, isolation, and resource allocation. Figure 1 depicts the substrate and overlay architecture.
Figure 1. The GNA Overlay Network Architecture
GNA Operations must provide a very highly reliable and secure global production network. At the same time, GNA Operations must provide a flexible operations environment to meet the changing needs of the research community. Operations for the GNA is envisioned as a federated model, with the world's leading R & E network operations centers playing an active role.
Towards a next gen intercontinental R&E infrastructure
The GNA network infrastructure is expected to evolve over time. On intercontinental routes that have more competition and more availability of bandwidth we expect to see the first instances of the GNA infrastructure start and grow. Ultimately, all continents and all regions will connect to the expanding GNA infrastructure once development has started and the model has proven itself as a technically, financially and politically sound model.