<br><br><div class="gmail_quote">On Tue, Aug 31, 2010 at 1:28 AM, Ai Austin <span dir="ltr"><<a href="mailto:ai.ai.austin@gmail.com">ai.ai.austin@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
Justincc wrote:<br>
<br>
<blockquote class="gmail_quote" style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;">
...though more to handle the case where the lookup URL had gone away or no network was available.<br>
</blockquote>
<br>
Yes, that has not been mentioned before. We have some use cases where OpenSim is run disconnected from the Internet completely. OpenSim needs to always work without depending on an hosts to a closed or isolated network. Firewalled situations and strong corporate or institutional controls on external access are also likely for many applications anyway. <br>
</blockquote><div><br>I think the cache model handles this very well - you update the cache whenever and however you can and must accept that while the cache is "dirty" it will diverge from the "master" by some unknown amount. <br>
<br>I don't see this as any different that the standard memory cache model.<br><br>The cache can only have the best data that is available, no matter how corrupt it is.<br><br>When the cache gets updated then there is a potentially huge problem of correcting the bad data, but it is all resolvable.<br>
<br>If you couldn't figure out what the data SHOULD be at any point in time then you have a fatal problem!<br><br>Karen<br></div></div>