<p>An important question is whether Hcoop wishes to own hardware infrastructure. Ownership of resources is important, especially at large scales.</p>
<p>If we were to go the VPS route -- largely for the simplified managment -- I would recommend we purchase our VPS from a cooperative provider such as Cernio. Despite the slightly higher cost, our members payments would go toward supporting a similarly-aligned effort and we would have some say in how the VPS organization was run.</p>
<p>I hope that some day our internet transit, power, colocation space and software-level organizations are all democratically run and accountable.</p>
<p>--philip</p>
<p><blockquote>On May 17, 2009 7:00 AM, "Adam Chlipala" <<a href="mailto:adamc@hcoop.net">adamc@hcoop.net</a>> wrote:<br><br>This discussion seems to have sputtered out. Personally, I'm thinking<br>
of 100% VPS hosting (probably spread among multiple providers to some<br>
extent) as the best option for HCoop going forward. Rob Gubler<br>
mentioned some file system/IO failures he's experienced with<br>
virtualization. Shaun Empie said he doesn't think moving to 100% VPS is<br>
a good idea, but he didn't give any reasons, and he hasn't responded yet<br>
to a query about what they are. Nathan Kennedy has been saying for a<br>
while that we should own our own servers. Besides that, I think we've<br>
only seen positive opinions about the VPS option.<br>
<br>
Here's my summary of the relevant issues, in decreasing order of importance:<br>
<br>
VPS Pros:<br>
- Virtual servers can be accessed from anywhere, as long as they're in<br>
good enough shape to be accessible via console. We can always at least<br>
reboot them remotely for free. This means that we lose _all_ geographic<br>
constraints on where admins need to be to handle any possible failure.<br>
Thus, we can have 100% time coverage for emergency response, without<br>
needing to find anyone willing to work the admin graveyard shift, since<br>
we can spread admins across time zones. Also, even if our servers live<br>
in a relatively high cost-of-living area which has good 'net<br>
connectivity, we can still hire admins anywhere in the world.<br>
- We don't have to purchase, diagnose, or replace hardware. The VPS<br>
provider will do that for us. We just need to be ready to move<br>
functionality from broken nodes reasonably quickly, probably using the<br>
provider's control panel interface.<br>
- We can add new machines with a few control panel clicks, in response<br>
to load increases or anything else.<br>
- For underutilized virtual machines, we probably save at least a little<br>
bit of money compared to dedicated servers or colocation of HCoop-owned<br>
servers.<br>
<br>
VPS Cons:<br>
- Crazy stuff might happen with simulated hardware. I don't know how<br>
likely it is that we'd experience anything like this with one of the<br>
major providers, though, since they spend a lot of time tweaking kernels<br>
to avoid such trouble.<br>
- We wouldn't own the machines. (Is this really a problem? Maybe it's<br>
just old-fashioned to think of owning machines instead of data.)<br>
- We would be sharing machines with other customers, which could lead to<br>
not enough computational resources being available for us, in periods of<br>
high load.<br>
- We would be sharing machines, so new potential security problems<br>
emerge. (Interesting note on this and the last point: Linode apparently<br>
has an unlisted VPS kind big enough that it actually takes up its own<br>
8-core machine, for $800/mo..)<br>
<br>
Did I miss anything?<br>
<p><font color="#500050">
_______________________________________________
HCoop-Discuss mailing list
HCoop-Discuss@lists.hcoo...</font></p></blockquote></p>