Currently, network engineers are faced with the challenge of building networks which support multiple classes of traffic, with differing service guarantees, while optimizing network resource utilization. While some technologies, such as ATM, hold the promise of supporting applications with a wide range of requirements, the interaction of applications with these new emerging technologies is not well understood and warrants serious investigation through careful research. Today, scientific applications are typically run over packet networks, offering best effort service. In order to enhance end-to-end services for these applications by applying traffic shaping capabilities, comprehensive measurements of traffic characteristics, performance bounds, and service parameters are needed. The nature of this traffic in the local, metropolitan, and wide-area cell-switched environments requires study to determine how to effectively provision, police (on input) and shape (on output) future networks in support of scientific applications. MREN has undertaken a wide range of research projects that will investigate these areas as well as develop new techniques for connecting local network infrastructure to the "next generation" Internet, for example, by using innovative traffic segmentation methods. MREN also supports special innovative networking research projects, such as the I-WAY project at the Supercomputing '95 conference.
Several MREN projects involve basic research. One such MREN project will investigate the use of traffic descriptors (i.e., variable bit rate, latency, multipoint distribution) to exploit the advantages of ATM in order to manage effectively network traffic to achieve high levels of resource utilization, and to provision the necessary quality-of-service and other network services. Some of the objectives we propose are to:
This project will generate valuable data on performance optimization for a variety of scientific research applications utilizing high speed networks. The data will benefit application developers as well as network designers in configuration, management, and tuning for the types of traffic generated by varied applications. This data will be particularly useful in developing optimal network configuration rules for the specific types of traffic-shaping attributes required by such applications. For example, the project will generate data, and perhaps techniques, that could assist in developing application-specific descriptor information that can be signaled to the network in advance of its utilization in order to optimize it for that particular application. Such data would be an assistance in the development of a type of signaling process whereby associative links can be formed between requests for particular types of flows and related types of performance requirements.
Furthermore, such information will provide a base for generating rules required for optimal network performance while the application is running, for example, with regard to traffic governance, monitoring and regulating. Information developed through this project will also be useful for designing rules on priority queuing techniques for specific types of scientific research applications (e.g., remote use of VR environments, such as the CAVE, interactivity with instrumentation at the Advanced Photon Source, dynamic scientific visualization, etc.). Applications requiring multicasting will provide opportunities to explore a particularly complex set of these issues.
In order to begin enhanced planning for, and additional implementation of a systematic means of technically managing the MREN network, the MREN community has agreed to establish a pooled fund to accomplish a variety of initial tasks. MREN's advanced network application projects and technical infrastructure projects can be categorized into four quadrants: near-term and long term leading-edge technologies and applications (pre-production) and near-term and long term advanced technologies and applications (production). Some of these projects are directed at creating network management tools for next generation networks. In addition to the development of a range of network management tools for advanced high performance networks, MREN has undertaken a wide-range of R&D projects.
EMERGE: ESnet/MREN Regional Experimental NGI Testbed
This EMERGE effort means to achieve and demonstrate Differentiated Services (DiffServ) over an IP/ATM GigaPoP regional network as a representative of the second model for DoE/University connectivity. This effort will support DoE-specific next generation internet (NGI) applications and attempt to motivate interoperability across GigaPoPs, the UCAID/Internet2 Abilene network (the third model) and ESnet (the first model). This testbed will establish a common suite of advanced networking services and the applications-friendly Grid Sevices Package for use by DoE laboratories and university applications, ultimately nationwide.
The partners are the Electronic Visualization Laboratory (EVL) at the University of Illinois at Chicago (UIC), the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign, and the International Center for Advance Internet Research (iCAIR) at Northwestern University, with assistance of the Metropolitan Research and Education Network (MREN). In addition, the funded university partners are The University of Chicago ASCI FLASH Center, The University of Wisconsin Engine Research Center, The University of Illiois at Chicago Electronic Visualization Laboratory, The University of Illinois at Urbana-Champaign ASCI CSAR Center, and the iCAIR Center at Northwestern University.
Joint Projects with Ameritech:
MREN has undertaken a number of joint research and development projects, including several projects with Ameritech, that will focus on designing, developing, and implementing advanced innovative networking technologies.
Although MREN relies on ATM, its researchers have investigated various aspects of ATM provisioning. MREN has explored the issue of ATM QoS support in the ISP clouds as well as access to these capabilities by the end user. For example, one option has been to treat the ATM cloud as only a raw bit pipe and to rely on techniques such as RSVP to provide end-to-end QoS across not only the non-ATM LAN technologies, but also the carrier's ATM clouds. However, this approach defeats one of the major reasons an end user would consider deploying ATM on the campus or explicitly request it for WAN services. It is possible to argue that RSVP QoS is not the absolute QoS some applications require, and therefore MREN should utilize ATM QoS whenever possible. In either case, the ability of the end users to use ATM QoS signaling in a dynamic fashion to satisfy their dynamic application requirements is dependent on the availability of standards-based signaling implementations and APIs in the switches and end host systems, as well as admission control capabilities for both ATM and RSVP. The current state of deployment for ATM equipment that can support applications dynamically signaling and managing QoS in regional and WAN networks is fairly poor; this may impede the adoption of native ATM by the end user community. The lack of RSVP admission control tools available for use by the end user and network manager, as well as the lack of admission policies based on the application and campus network managers' perspective, may also impede the adoption of RSVP.
MREN intends to extend the concept of regional sharing of infrastructure one step further by defining a network peering point where multiple local entities and institutions can connect and peer with each other and by providing a common aggregation point and peering point with WAN ISPs, such as Sprint, MCI, the vBNS and ESnet. The Network Access Points (NAPs) were originally designed to support this model, but the implementations failed in this regard because they only provided ISP-to-ISP peering. The Gigapop is the latest iterative concept and attempt to support a communal sharing of infrastructure to peer local institutions with advanced production services and ISPs. MREN contends that the Multimode GigaPOP (M-GigaPOP) extends the GigaPOP and NAP concept, because it can concurrently support both production and R&D traffic on as much of the same infrastructure as possible and hand off the traffic to the appropriate commercial or R&D ISP, depending on the type of traffic.
The research challenges are how to ensure that one type of traffic does not adversely affect the other at the MREN-GigaPOP and how to provide for distributed network management of the peering point(s) (i.e., what end user tools and management capabilities are required in the switches, routers, and multiplexors).
MREN has begun to address issues of optimizing performance over shared WAN infrastructure. When providing shared wide-area infrastructure, the telecommunications service providers (e.g., MCI) and ISPs (e.g., ESnet and vBNS) will face many of the same issues as the traditional regional carriers (e.g., Ameritech ) and ISPs. The major issues center on what access interface and capabilities are provided to the end user and how experimental traffic is supported on the same or separate infrastructure as production traffic. For example, all experimental traffic may be provided over physically separate circuits and switches within the WAN ISPs cloud.
The ability of the telecommunications carriers to provide multi-modal infrastructure may be hindered by the fact that some of their customers do not like to assume any risk. The federally funded private WAN ISPs (e.g., ESnet, NSI, vBNS, DREN) may have a little more latitude in supporting some experimental network traffic and capabilities within their clouds, but they are also reluctant to assume much risk because some members of the application research community they support expect absolute production-level services.
However, the challenge still facing all ISPs who expect to be solvent and viable service providers in the future will be how to support multiple varied policy (e.g., production versus experimental or guaranteed versus best-effort services) virtual networks, because it is too costly for both the end user and the provider to support duplicative infrastructures (for the reasons already outlined in this report). Small amounts of calculated risk are critical in the evolution of networks and must be assumed by the end user and the service providers. Even when one tests router or undertakes switch upgrades in a bounded environment prior to deploying these changes into production networks, one still assumes some risk when the upgrades are finally deployed, because any change to the running system or network in effect changes it from a production to an experimental network, albeit a controlled one. On many occasions seemingly small upgrades or modifications have caused far-reaching problems. Networks must be developed that are more resilient and fault tolerant (i.e., can support experimental as well as production traffic and be dynamically configured to compensate for problems) on both a macroscopic and microscopic level. The on-demand use and re-use of network infrastructure components and segments will further enable the service providers to support both the production and experimental requirements, as well as the other varied and sometimes conflicting policy-based network requirements of its customers more efficiently and cost-effectively.
Many of the objectives of the various MREN R&D projects are captured in the concepts behind the proposed MORPHnet project which will allow MREN to become THE prototypical Gigapop and next generation NAP. Recent conversations with the NSF's DNCRI have verified that NSF is interested in establishing interconnection and peering on multiple dimensions. The first dimension is the Gigapop, which is envisioned as an aggregation point for institutions and networks. The second dimension focuses on the international STAR -TAP Science, Technology and Research Testbed Access Point, for which NCSA has the program lead, and both NCSA and ANL have the technical lead for design and implementation. The final dimension speaks directly to those ideas professed in the ANL's MORPHnet document and that is the development of peering points (aka SUPERNAPs) that address the varied and complex requirements of both research and production applications and research activities.
MREN campus research network infrastructures are currently being migrated to newer technologies. These new developments tend to be focused on ATM backbones, and most new links are to such backbones. Also, FDDI-based research networks are being converted to ATM. Over twelve months, the standard campus research network may be ATM based. Plans have been made to ensure a high performance connection to all qualified users and not be restricted to the proposed applications requiring the high-performance network. Planning also calls for the Internet to be interconnected with the high performance, local network infrastructure.
The MREN consortium works to ensure that particular quality of service guarantees are available for the applications of its constituent community. This requires an end-to-end architectural approach, which requires a fairly complete knowledge of the total architecture and of all management parameters. It also requires the development of guidelines for numerous options: a) addressing schemes, b) selection of standards, especially for interoperability and routing, c) MAC layer options, d) performance benchmarks, e) performance monitoring, analysis, and adjustment techniques, and f) security.
This issue also involves protocol selection and management, especially decisions related to IP and ATM. One general issue is the integration of MREN as an ATM-based regional network with existing on-campus high performance networks and general purpose networks, especially the common campus Ethernet with an FDDI backbone, but also including network segments with enhanced components such as switched Ethernet. Another issue that must be addressed is the development of guidelines for integrating research workgroup facilities and individual workstations, including suggestions on application tuning and requirements signaling, feedback, and adjustment, operating system and system utility tuning, system configuration, NICs, NIC drivers, and bus structures.
Key issues being addressed include:
MREN expects to become increasingly involved in the Globus project, an innovative project designing a system (the Globus System) intended to achieve a vertically integrated treatment of high performance applications, middleware, and networks. This project is being led by Ian Foster, the Mathematics and Computer Science Division of Argonne National Laboratory, and Carl Kesselman, at the University of California, Information Sciences Institute.
[Home] [Next Page] [Previous Page]
Copyright 2000 by MREN.ORG. All rights reserved.