NASA successfully launched the ACTS on September 12, 1993. For various information regarding ACTS, see http://kronos.lerc.nasa.gov. A particular capability of ACTS is its Gigabit Satellite Network (GSN) capability (see http://www.cgrg.ohio-state.edu/other/actsgsn/gsnhome.html) which among other features allows ACTS to concurrently support up to three separate 622 Mbps digital data streams.
To utilize ACTS' GSN capability, NASA partnered with DARPA in the development of five High Data Rate (HDR) Terminals (HDRT's). Each of the HDRT's can support either one OC-12c (622 Mbps) full duplex connection or up to four concurrent OC-3c (155 Mbps) full duplex connection.
Of particular importance to this experiment plan, one HDRT recently has been located at NASA's Goddard Space Flight Center (GSFC) which is a core site in the ATDNet, and another HDRT recently has been located at US Sprint's Technical Integration and Operations Center (TIOC) which is a key site in the MAGIC network.
To interconnect between the GSFC and TIOC sites via ACTS at 622 Mbps, a change will need to be made in the ACTS-related default beams used in data transmissions with the HDRT at GSFC. Without such a change, the interconnection via ACTS would be limited to a maximum of 311 Mbps. Such a change is presently thought to be technically feasible. If, however, it should not be, then this experiment plan will apply for 311 Mbps network tests between the ATDNet and MAGIC networks via ACTS.
1.2 Goals
This experiment plan identifies a number of special ATM-based network tests planned to be conducted at 622 Mbps rates over ACTS between sites on the ATDNet and sites on the MAGIC network. The network tests range from assessment of ATM signaling at ISO layers 1-2, "tuning" of IP/TCP protocols at network/transport layers 3-4, evaluation of new network management/monitoring techniques involving several layers, through "everyday" and special purpose presentation/applications at layers 6-7. Each of these tests is planned to be conducted so as to evaluate and maximize performance in the unique 622 Mbps high bandwidth*delay network created by the hybrid satellite/terrestrial ATDNet-ACTS-MAGIC network.
ATM-based network tests at 622 Mbps are rare today even in terrestrial-based networks. The inclusion of the ACTS-based 622 Mbps interconnection between such terrestrial-based networks creates an additional challenge, but may be indicative of many future high performance networks. The ACTS-based interconnection creates the opportunity to conduct and evaluate many experiments which were not previously possible.
1.3 Co-I's and other Participants
The Co-I's and other participants identified below largely are already either DARPA-funded (or other DoD-funded) investigators working in/for the ATDNet and/or MAGIC projects (or the ACTS ATM Internetwork (AAI); see http://info.arl.army.mil/HPCMP/DREN/testbed/aai.html) or NASA-funded investigators working in/for the ACTS project; and the network tests to be conducted under this experiment plan are relatively straightforward extensions to on-going research being conducted within those projects.
In each of the Network Test Suites below, a single Co-I has been identified who will serve as either the lead researcher for a relatively integrated group of researchers working closely together for that type of network test or as a facilitator for a number of researchers/research groups who are working relatively independently of one another.
2.0 Network Test Suites
2.1 Assessment of Satellite Links on ATM Signaling
Co-I: Rich Verjinski/Fore
Objective: Identify and measure effects of satellite-link-induced
high latency transmissions on new ATM signaling options. Approach: Exploit special hardware monitoring tools in joint collaboration involving FORE, NRL, and US Sprint in KC.
2.2 Tuning TCP over High Speed Satellite Links
Co-I: Pat Gary/GSFC
Objective: Assess effects of new TCP protocols with selective
retransmissions/acknowledgments over high speed satellite links
Approach: Utilize "drop-in" UNIX codes for selective
retransmissions/acknowledgements now being beta-tested at the Pittsburgh Supercomputer Center
2.3 Evaluation of ATM Flow Control and Traffic Monitoring Techniques in a
622 Mbps Hybrid Satellite/Terrestrial Network
Co-I: Victor Frost/KU
Objective: Produce new network measurement tools applicable to high
speed hybrid satellite/terrestrial networks. Approach: Extend the ATM network monitoring techniques developed in
MAGIC and apply over the ATDNet-ACTS-MAGIC network.
2.4 Demonstration and Evaluation of Everyday Internet Applications
across the ATDNet-ACTS-MAGIC Network at 622 Mbps
Co-I: Pat Gary/GSFC
Objective: Demonstrate that high speed hybrid satellite/terrestrial
networks can be applicable to typical Internet use. Approach: Install and assess performance of client/server daemon
implementations for everyday-use protocols such as ftp, http, nfs, mpi, pvm, etc., but which have been modified to utilize the extended windows needed for effective use in very high bandwidth*delay networks.
2.5 Demonstration and Evaluation of TerraVision/ISS Operating over the
ATDNet-ACTS-MAGIC Network
Co-I: Jay Feuquay/EDC
Objective: Use TerraVision/ISS as an application that can stress the
622 Mbps the ATDNet-ACTS-MAGIC network.
Approach: Create and utilize ISS servers running concurrently on
the ATDNet and MAGIC networks.
2.6 Multimedia Telemedicine Applications Operating over the
ATDNet-ACTS-MAGIC Network
Co-I: Kenneth Kempner/NIH
Objective: Evaluate 155/622 Mbps networks for multimedia telemedicine
consultation applications.
Approach: Conduct multimedia telemedicine consultation sessions, in
support of medical imaging applications, between the NIH and GUMC on ATDNet, and between the NIH and Washington University in St. Louis on ATDnet and MAGIC, via ACTS.
2.7 Telemedicine-enabling R&D Testbed Experiments Operating over the
ATDNet-ACTS-MAGIC/NTON Network
Co-I: Mike Gill/NLM
Objective: Evaluate 155/622 Mbps networks for telemedicine
enabling technologies.
Approach: Conduct telemedicine-enabling R&D testbed experiments,
typically involving large databases, between NLM on ATDnet and the Laboratory for Radiological Informatics at the University of California San Francisco on the National Transparent Optical Network (NTON) via the HDRT at JPL.
2.8 Testbedding of New Applications at 622 Mbps
Co-I: Pat Gary/GSFC
Objective: Provide opportunity for demonstrations of new
applications that can effectively utilize 622 Mbps, including applications running directly over ATM. Approach: Allocate and coordinate satellite time and host computer
time for developers of new applications that require high speed networks.
3.0 Coordination of Network Test Times with ACTS Operations
PI J. Patrick Gary will serve as overall coordinator and primary point-of-contact of this experiment's network test time with ACTS Operations. Mr. Gary will provide this coordination based on network test time needs provided by the above Network Test Suite Co-I's who will be responsible for identifying and organizing requests for ACTS time to support their suite of network tests.
Until further refined based on Co-I inputs, an initial request for ACTS time consisting of four hours contiguous duration in the afternoon's of Monday, Wednesday, and Friday will be submitted to the ACTS Project.
4.0 Overall Schedule
Overall this experiment plans to operate from the beginning of ACTS HDRT functional readiness at GSFC and TIOC, which is essentially May 1996, until at least September 1997.
However, at present not all of the terrestrial-based 622 Mbps network infrastructure is in place. In particular, DARPA and the rest of the ATDNet's Steering Committee only recently approved that the ATDNet's SONET service provider Bell Atlantic is to provide GSFC with a pair of 622 Mbps SONET drops off the 2.4 Gbps ATDNet backbone. This will allow the ACTS HDRT at GSFC to be accessed at 622 Mbps rates by NSA, NRL, and/or DIA which already have such speed drops in ATDNet. Furthermore, GSFC must acquire and install a new ATM core switch with three new 622 Mbps netmods - two to sustain the 622 Mbps ATM ring with NSA, NRL, and DIA, and one to interface with the ACTS HDRT.
These upgrades at GSFC are expected to be completed by mid-summer 1996.
Nevertheless, this experiment will start requesting ACTS time at 155 Mbps right away to facilitate ramping up of the various network test suites.