Performance Testing

Since version 2.7, Aspera has offered some special performance testing capabilities. The major bottlenecks in performance are:

1) Network bandwidth
2) Receiving system’s storage write speed
3) Sending system’s storage read speed
4) CPU speed of the end systems

The network bandwidth capacity can be isolated from the other performance limiting factors by using the “--no-read --no-write” command line options. These options cause ascp not to read from the storage, nor to write to the storage, but rather to simulate reading and writing. There must actually be a source file existing and specified for this process to work.
 

The syntax is as follows for a 500Mbps transfer, testing only the network (replace 500 with your network bandwidth) and not the storage:

ascp -QT -l 500M --no-read --no-write bigfile user@destination:/

If the transfer completes at close to 500Mbps, with little packet loss, then you can assume that the transfers can proceed at 500Mbps, without being constrained by network limitations. If your transfer rate is less than the target rate, then the transfer is either limited by network congestion, or endpoint CPU speed, you should contact Aspera Technical Support for assistance.

To test the endpoints' read/write speed. If the previous test transferred at our near the target rate, then you should enable reading to the previous test:

ascp -QT -l 500M --no-write bigfile user@destination:/

 Once again, if the transfer rate is close to the target rate, then you do not have a sender side storage limitation. If the transfer rate is considerably slower, you do have a sender side storage limitation, which may be responsive to tuning. In either case, you would need to contact Aspera Technical Support for assistance.

To test the receiving side's ability to write to storage with this command:

ascp -QT -l 500M --no-read bigfile user@destination:/

As in the sender side storage test, if the transfer rate is considerably less than the target rate, then you have a storage write speed limitation, which you may be able to improve with some tuning, but you would need to contact Aspera Technical support for assistance.

To test the CPU speed of the end systems, you can run the initial test by enabling encryption (no '-T' and no reading from or writing to disk):

ascp -Q -l 500M --no-read --no-write bigfile user@destination:/

If the transfer rate is noticeably slower than that attained with the '-T' option (disabling encryption), then you can be sure that your CPU will be a bottleneck with encrypted transfers, and possibly even unencrypted transfers, if the original no-read/no-write test was considerably less than your network bandwidth. 

 

Have more questions? Submit a request

0 Comments

Please sign in to leave a comment.
Powered by Zendesk