Specifications

3 Data Recording
According to the station architecture do cument [1], LWA stations are to have a data aggregation and
communication (DAC) subsystem that is responsible for the routing of data from digital processing,
and further says “The DAC includes separate output to facilitate local recording of output directly
to disk. For example, this would allow continued operation of the station should data path to the
LWA central processing facility be interrupted, or not yet implemented.” However, DAC is currently
an “orphan” – i.e., no institution is working on it, and no ONR FY07 and 08 funds are available
for its development. In its place, University of New Mexico agreed to provide a network interface
allowing off-site network communication, and Virginia Tech agreed to implement a rudimentary local
data recording capability as part of the LWA-1 MCS.
The MCS interim data recording capability will accept output from DP via 10 Gb/s ethernet
connections. Each connection nominally operates at less than 100 MB/s (800 Mb/s), although
transfer rates up to 200 MB/s (1600 Mb/s) should be possible. The maximum duration of sustained
recording depends on the transfer rate as well as the amount of disk drive storage. Also, the fraction
of drive space that is available for sustained transfer diminishes with increasing transfer rate.
1
An
educated guess is that at least 500 GB will be available per ethernet input, assuming 100 MB/s
transfer per ethernet input. At 200 MB/s, the available storage for sustained transfer is difficult to
predict but is likely to be significantly less.
It is important to note that MCS-DR will only record data from DP; it will not necessarily refor-
mat the data, nor will it necessarily reorder data packets from DP should they arrive out of order.
Thus, the data available to users will be in the same form and the same order received from DP by
MCS-DR, at least in the initial realization of the system.
There will ultimately be four ways in which data can be recovered from the data recording com-
puters. (1) Through an LTO-type tape drive which is part of MCS, (2) Through an internal station
network directly into a user-provided PC, (3) Through an MCS-provided USB port directly to a
user-provided external hard drive, and (4) Through the external internet connection to a remote
computer. Option (3) will be implemented first, and is intended primarily as a stopgap measure.
Option (4) may be restricted based on availability of the connection and priorities for its use.
It should be noted that it will probably not be possible to offload data – using any of the four
techniques described above – while an observation is underway. It may be possible to “array” the
data recording PCs so that some can b e offloading while others are recording.
It is anticipated that the total amount of storage provided by data recording PCs will be limited
to a few terabytes. Thus, a protocol will be established to automatically delete data from previous
observations in order to make room for upcoming observations. The time required to offload data
could result in significant “down time” for certain observing modes, and will need to be considered
when planning observations.
4 User Interface
1. The primary physical interface with users will be via a network connection. Users which are
on-site may access MCS using personally-owned computers connected to an internal ethernet
connection. Users which are off-site may access MCS through the external internet connection.
In this sense, there will be not normally be a distinction between users which are physically
present at the station, and users which are accessing MCS via an external internet connection.
1
This is because transfer to hard drives tends to be most efficient when the drive is empty of data, and tends to
degrade as more data is placed on the drive.
3