Slashdot Mirror
MS Releases Open Source Alternative To BigTable
gollito writes in with news that Microsoft has released an open source alternative to Google's BigTable file system, which is used on large distributed computer clusters. Matt Asay writes for CNet: "I also believe that Microsoft's fear-mongering around open source cost it years of productivity and quality gains that it could have been delivering to customers through open source. I hope that reign of ignorance is over."
So this means pigs CAN fly?
They open sourced the surface?
So... the linked article says the Kumo search team (the ones who develop the FS) USE open source. But I can nowhere see that the FS is released as open source. A citation would be good, especially since the used license would be quit important.
Google doesn't sell/license BigTable in any way. It's used internally. I fail to see how it's possible to release an alternative to something which can't be acquired in any form.
Not completely correct. You can use BigTable right now. There are Google AppEngine APIs that can access BigTable. You just can't use it without using Google's servers, that's all.
If, at this point, you still can't see why it's completely obvious why Microsoft would write an alternative to BigTable and open source it, all I can say you haven't been paying attention.
My blog
Using Google's AppEngine, you can use BigTable.. so while you can't install it on your own servers, you can still write software that uses it.
Clearly Microsoft is using open source as a tactical weapon here, the way companies often do against entrenched competitors.
But is this a new tactic for them? No. Back in the '90s, they competed against Netscape in the browser wars by giving away IE for free; unlike Netscape, which was hoping to eventually start charging for Navigator, Microsoft made IE part of Windows (so it was effectively free for anyone who already paid for the PC).
And Microsoft released an "Open Letter to Netscape", asking its rival to cooperate with the W3C and avoid making proprietary extensions to web protocols. As if anything else about Windows desktop development at the time was based on open standards!
Going back even further, at one point Borland International was the leading PC software tools vendor. Microsoft wanted this title for itself (remember "developers developers developers developers"), so to compete against Borland's Object Windows C++ framework, they came up with MFC. And following Borland's lead, they made MFC open source (or "shared source" or whatever. Source available).
So no, they aren't having a change of heart. They will do whatever it takes to get control of this hot market segment.
Microsoft has allowed two of the primary HBase developers, who work at Powerset, to continue their open-source work on HBase, which is definitely cool. But to say that Microsoft is releasing this is just flat out wrong.
(Full disclosure: I am a non-Microsoft-employed HBase committer.)
Actually, you're not even close.
A company called Powerset developed the open-source alternative to BigTable called HBase. This was developed as an Apache Software Foundation project under the Apache license.
Microsoft bought Powerset for a bucket of money because their search technology based of Hbase was pretty damned good. This was last year. This year, the folks behind powerset - as Microsoft employees - were given the go-ahead to continue committing to the ASF project and they continue to make it better. For what I can see, they aren't keeping anything juicy in-house.
It's honest-to-goodness MS committing to an Apache project.
Rubbish
emacs is clearly superior
"Physics is to math as sex is to masturbation." -R. Feynman
Which means that your appliance that uses BigTable needs continuous access to the Internet.
What!? This is absolutely outrageous~! None of my servers have internet access!
You'll note that Google aren't opening up their crown jewels: you can't just download their raw web page index and do your own thing with it. Since they're not in the software business, they can afford to give away or open their software tools. Since Microsoft are in the software business, that hurts them.
Now there's an interesting symmetry here. Being (primarily) in the software business should mean that actual content and databases isn't too important for Microsoft. If they wanted to hurt Google, they would open up their raw msnsearch indexes and other useful content databases. That would hurt Google, because people could download massive competing data collections and create their own competing search engines without the huge resource investment in crawler farms etc.