<div dir="ltr"><div>That is great.<br></div><div><br></div><div>Here is an <a href="http://appear.in">appear.in</a> room we can use for the meeting tomorrow:Â <a href="https://appear.in/erasurecode41">https://appear.in/erasurecode41</a></div><div><br></div><div>Looking forward to the meeting!</div><div><br></div><div>Best,</div><div>Per</div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Fri, Mar 3, 2017 at 1:15 PM, Xavier Hernandez <span dir="ltr"><<a href="mailto:xhernandez@datalab.es" target="_blank">xhernandez@datalab.es</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">March 8th also works for me.<span class=""><br>
<br>
On 03/03/17 06:18, Pranith Kumar Karampuri wrote:<br>
</span><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">
8th March works for me. 9th I am travelling. Xavi is well versed with<br>
this part of the code, so even if I am not available, you guys should<br>
pick the time which is convenient for Xavi.<br>
<br>
The following may be of help for this discussion.<br>
<a href="https://github.com/gluster/glusterfs/blob/master/doc/developer-guide/ec-implementation.md" rel="noreferrer" target="_blank">https://github.com/gluster/glu<wbr>sterfs/blob/master/doc/develop<wbr>er-guide/ec-implementation.md</a><br>
<br>
If my memory serves right, Xavi's implementation outperformed<br>
liberasure, that is the reason we didn't integrate with liberasure sooner.<br>
<br>
<br>
On Fri, Mar 3, 2017 at 7:01 AM, Per Simonsen <<a href="mailto:per.simonsen@memoscale.com" target="_blank">per.simonsen@memoscale.com</a><br></span><span class="">
<mailto:<a href="mailto:per.simonsen@memoscale.com" target="_blank">per.simonsen@memoscale<wbr>.com</a>>> wrote:<br>
<br>
  Hi,<br>
<br>
  I forgot to add the time zone: the suggested time is 10 am (GMT+1).<br>
<br>
  Best,<br>
  Per<br>
<br>
  On Thu, Mar 2, 2017 at 5:19 PM, Per Simonsen<br></span><span class="">
  <<a href="mailto:per.simonsen@memoscale.com" target="_blank">per.simonsen@memoscale.com</a> <mailto:<a href="mailto:per.simonsen@memoscale.com" target="_blank">per.simonsen@memoscale<wbr>.com</a>>> wrote:<br>
<br>
    Hi,<br>
<br>
    I suggest that we set up an online meeting next week to discuss<br>
    the erasure coding features as well as possible implementations<br>
    of a plugin architecture. We also have some experience<br>
    integrating with the liberasure library mentioned by Prashant<br>
    which we can share.<br>
<br>
    Does 10 am on Wednesday(8th of March) or Thursday(9th of March)<br>
    next week work for you guys?<br>
<br>
    Best,<br>
    Per Simonsen<br>
    CEO<br>
    MemoScale<br>
<br>
<br>
    On Thu, Mar 2, 2017 at 12:00 AM, Xavier Hernandez<br></span><div><div class="h5">
    <<a href="mailto:xhernandez@datalab.es" target="_blank">xhernandez@datalab.es</a> <mailto:<a href="mailto:xhernandez@datalab.es" target="_blank">xhernandez@datalab.es</a>><wbr>> wrote:<br>
<br>
      Hi Niels,<br>
<br>
      On 02/03/17 07:58, Niels de Vos wrote:<br>
<br>
        Hi guys,<br>
<br>
        I think this is a topic/question that has come up<br>
        before, but I can not<br>
        find any references or feature requests related to it.<br>
        Because there are<br>
        different libraries for Erasure Coding, it would be<br>
        interesting to be<br>
        able to select alternatives to the bundled<br>
        implementation that Gluster<br>
        has.<br>
<br>
<br>
      I agree.<br>
<br>
        Are there any plans to make the current Erasure Coding<br>
        implementation more pluggable?<br>
<br>
<br>
      Yes. I've had this in my todo list for a long time. Once I<br>
      even tried to implement the necessary infrastructure but<br>
      didn't finish and now the code has changed too much to reuse it.<br>
<br>
        Would this be a possible feature request,<br>
        or would it require a major rewrite of the current<br>
        interface?<br>
<br>
<br>
      At the time I tried it, it required major changes. Now that<br>
      the code has been considerably restructured to incorporate<br>
      the dynamic code generation feature, maybe it doesn't<br>
      require so many changes, though I'm not sure.<br>
<br>
<br>
        Here at FAST [0] I have briefly spoken to Per Simonsen<br>
        from MemoScale<br>
        [1]. This company offers a (proprietary) library for<br>
        Erasure Coding,<br>
        optimized for different architectures, and with some<br>
        unique(?) features<br>
        for recovering a failed fragment/disk. If Gluster allows<br>
        alternative<br>
        implementations for the encoding, it would help<br>
        organisations and<br>
        researchers to get results of their work in a<br>
        distributed filesystem.<br>
        And with that, spread the word about how easy to adapt<br>
        and extend<br>
        Gluster is :-)<br>
<br>
<br>
      That could be interesting. Is there any place where I can<br>
      find additional information about the features of this library ?<br>
<br>
      Xavi<br>
<br>
<br>
<br>
        Thanks,<br>
        Niels<br>
<br>
<br>
        0. <a href="https://www.usenix.org/conference/fast17" rel="noreferrer" target="_blank">https://www.usenix.org/confere<wbr>nce/fast17</a><br>
        <<a href="https://www.usenix.org/conference/fast17" rel="noreferrer" target="_blank">https://www.usenix.org/confer<wbr>ence/fast17</a>><br>
        1. <a href="https://memoscale.com/" rel="noreferrer" target="_blank">https://memoscale.com/</a><br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
--<br>
Pranith<br>
</div></div></blockquote>
<br>
</blockquote></div><br></div>