API/Python/Open-Source Solutions

Expand all | Collapse all

is the snapshot using an encrypted channel?

Jump to Best Answer
  • 1.  is the snapshot using an encrypted channel?

    Posted 10-29-2015 11:11:00 AM
    For Oracle snapshot, I know that the Engine is using SSH to connect to the source host and to launch scripts contained by the toolkit directory. But what I'm wondering is if the trasmission of data during a snapshot (wether it is first or incremental) is over an encrypted channel. I don't know the processes and network architecture Delphix is using for an RMAP API invocation...

    Sorry for my ignorance on Oracle processes
    Thank you.
    Gianpiero


  • 2.  RE: is the snapshot using an encrypted channel?
    Best Answer

    Posted 10-29-2015 11:20:00 AM
    Hi Gianpiero,

    Delphix uses SSH to initiate the process, but the actual data transfer is performed using DSP (Delphix Session Protocol). This data transfer can indeed be encrypted, but that will obviously incur some performance overhead, the default is not encrypted.

    See also this documentation link:

    https://docs.delphix.com/display/DOCS43/Advanced+Data+Management+Settings+for+Oracle+dSources

    It can be seen/changed on the back of the dSource card in the GUI, on the right column of the card (the "Data Management" section) labelled "Encrypted Linking".

    Best Regards,

    Scott.


  • 3.  RE: is the snapshot using an encrypted channel?

    Posted 10-29-2015 03:46:00 PM
    Thank you Scott! your reply is very satisfying. And I did learn another thing: the DSP Protocol...



  • 4.  RE: is the snapshot using an encrypted channel?

    Posted 10-29-2015 03:49:00 PM
    Hi Gianpiero,

    You may also be interested in this then:

    http://blog.delphix.com/peng/2013/06/14/delphix-session-protocol/

    :-)


  • 5.  RE: is the snapshot using an encrypted channel?

    Posted 11-13-2015 12:38:00 PM
    Hi Scott,

    which is the difference between the dsp test and throughput test in the network CLI section?

    It's clearly on the protocol used during the tests. But I expect that the results should be quite the same. Or can the used protocol influence the performance over the wire?

    While I'm writing I guessed that because I can test all together (including the encryption overhead), isn't it?

    Thank you.