[Foundation-l] Wolf Mountain MediaWiki Appliances Released

Wed Jul 26 21:37:49 UTC 2006

Jeffrey V. Merkey wrote:

>geni wrote:
>
>  
>
>>On 7/26/06, Gregory Maxwell <gmaxwell at gmail.com> wrote:
>> 
>>
>>    
>>
>>>Why so much storage?
>>>
>>>Uncompressed enwiki text is 680GB. All images in total are about
>>>300GB. Loaded into postgresql with a normal set of indexes enwiki
>>>takes about 400GB, I can imagine that a mysql load would be any larger
>>>(in mysql mediawiki uses zlib batch compression).
>>>
>>>I guess the 3TB makes sense to have some growth room.. but 18TB?
>>>
>>>:)
>>>   
>>>
>>>      
>>>
>>If you are doing mass computerised translations you could burn through
>>it pretty fast.
>>
>> 
>>
>>    
>>
>I have to host three instances of Wikipedia, so my storage for 
>WikiGadugi is at 2 TB at present. I use the a shared images
>setup just like the Commons, and an automated syncrhonizer that checks 
>for new images in a non-obstrusive way.
>
>Wikimedia may want to consider setting up a subscription service through 
>cogento for folks using these appliances to
>increase their revenues by offering a private mirror link for images and 
>content. Folks could buy appliances then purchase
>a subscription to Wikimedia's images and commons setup via rsync or http 
>in a similair setup to Red Hat's setup for
>support. Would allow a lot of MediaWiki appliances to move and also 
>provide Wikimedia additional sources of revenue
>with little to no investment on their end.
>
>Jeff
>_______________________________________________
>foundation-l mailing list
>foundation-l at wikimedia.org
>http://mail.wikipedia.org/mailman/listinfo/foundation-l
>
>  
>
And one more thing. If we are truly going to challenge google and yahoo 
at some point for internet search engine
supremecy, this is the path to go. Google got there by seeding the 
internet with appliances which skulk and harvest
content from the web with massive storage. I am currently adding search 
engine capabilities combined with
automated translation and wiki formatting so we can construct a search 
engine over time. It will take us about two
years to get where google is, but I have more powerful technology than 
they use, so over time, I think we can
eventually get there and wikify the internet. That's where this is going 
long term. Native Language preservation is
a huge deal right now -- one step at a time.

Jeff