Talk:Emerging strategic priorities/ESP 5 key questions/Will Wikimedia move away from a centralised server? How about distributing server loads across different volunteers?
- do you mean something like this? Proposal:Distributed_Wikipedia Lkcl 13:42, 23 August 2009 (UTC)
- The wiki way would definitely be decentralization to the max, as is teached in en:The starfish and the spider. Dedalus 16:14, 21 October 2009 (UTC)
- Cloud computing, I have done some basic work on the topic and find that having a cloud computing infrastructure in place would greatly augment our ability to leverage existing infrastructure and handle server loads in a very efficient and cost effective manner. Josemanimala 12:00, 22 October 2009 (UTC)
- I'm not convinced the benefits would be worth the effort required to develop and manage it. Its easy to find developers and sysadmins familiar with a LAMP system, probably a lot harder to find ones with experience with a distributed infrastructure. The only things such an infrastructure can really be useful for are tasks that are 1) Processor-intensive, 2) Take a relatively long time to complete, 3) Can be clearly defined, and 4) Are not time-critical. Currently we don't have very many things that meet those requirements. The main issue is number 4; being a website, most things need to be generated and served on-demand. Mr.Z-man 05:50, 25 October 2009 (UTC)
- w:Category:Cloud storage isn't exactly processor-intensive. --Fasten 10:33, 15 November 2009 (UTC)
- You want to "distribute load across different volunteers"? I'm not sure we have any volunteers that can serve as well as a server.
- More to the point, there is little evidence that truly distributed computing solves any problems we actually have. Mike.lifeguard | @meta 01:38, 15 November 2009 (UTC)