MIT eVLBI Tests, October 2002

Fastest eVLBI Transfer Rate to Date

Test-Results-1.pdf
Test-Results-2.pdf

On 24 October 2002, a successful direct e-VLBI data transfer from the GSFC/GGAO antenna in Maryland to Haystack Observatory in Massachusetts at a data rate of 288 Mbps was completed over a shared network. In this experiment, the data were transferred in real-time from the GGAO antenna to Haystack Observatory, where they were recorded on disk. The data were then correlated with data from the Westford antenna on the Mark 4 correlator.

Mark 5 data systems were used at all sites. Due to scheduling complications and an imminent and potentially lengthy fiber changeover in the Washington, D.C. area, this experiment had to be done as a piggyback to the ongoing 15-day continuous CONT02 experiment in which Westford is currently participating, without disturbing the CONT02 experiment. This severely limited the sources and integration times available, and we were required to observe in the same 256 Mbps mode as CONT02. Nevertheless, we were able to co-observe on a scan of the strong source 4C39.25 for a 4 minute period. During this period, data at 288 Mbps (with parity) was transmitted in real-time from GGAO and recorded on disks at Haystack. Correlation with disk recordings from Westford was done on the Mark 4 VLBI correlator. Correlation results were normal (see attached fringe plots) in all respects except for an unrelated problem at GGAO which caused the loss of 3 of 16 channels. The GGAO to Haystack link and the Mark 5 will support higher data rates, but in this experiment we were limited to 288 Mbps for compatibility with the ongoing CONT02 experiment.

Direct-transfer e-VLBI is another step in our progression of e-VLBI experiments. The next step will be real-time correlation of directly-transferred data. A fair amount of correlator software work remains to be done before this can be achieved. I note with respect that our Japanese colleagues have already accomplished this several years ago, but on dedicated links.

This work was supported by DARPA and NASA with generous cooperation from MIT Lincoln Lab, ISI-E, MAX, UMCP, and Supernet. A lot of people scrambled to make this experiment happen when we received notification of the likelihood of the network going down the next day. Steve Bernstein at Lincoln Lab secured the time for us on Glownet/Bossnet on extremely short notice. Mike Poirier at Westford and Jay Redmond and Chuck Kodak at GGAO scrambled to get antenna and data systems coordinated and set up. Nancy Vandenberg at GSFC quickly got schedules together for GGAO. Kevin Dudevoir and John Ball at Haystack oversaw the data transfer operation from GGAO, and Mike Titus oversaw the correlation. We were fortunate to accomplish all of this less than two hours before the network plug was pulled!

Submitted by Alan Whitney, Haystack Observatory