Anyone running FDB in production, good luck, downloads are removed from <a href="https://foundationdb.com/" rel="nofollow">https://foundationdb.com/</a><p>I don't understand why anyone would run a closed source database, especially with the open source options available.
Nooooo!!!! FoundationDB is too special to relegate to the iCloud back-end. There's nothing else quite like it out there that's publicly available, either commercial or open-source. This just set the industry back several years. Given that Apple has virtually zero interest in server-side development tools, I highly doubt us civilians will ever see this amazing technology again. :-(
What really differentiated it was the fact that multi-key transactions allowed for you to reasonably build any number of logical data models on top of it in a linearly scalable way. It was all built to an extremely high degree of polish with an extremely good testing and simulation harness and a high degree of predictability in performance. It was basically Spanner for the rest of us, without atomic clocks (and they also shipped an F1, their SQL layer on top). As others have mentioned, the closest cousin at the moment is probably cockroach, but it relies on wall clocks which will probably lead to problems in certain cases, but gets an easier way to scale writes.<p>Here's the architecture diagram for FDB, it's pretty fun to read:
<a href="https://foundationdb.com/files/Architecture.pdf" rel="nofollow">https://foundationdb.com/files/Architecture.pdf</a>
All the company's github repositories have been pulled:<p><a href="https://github.com/FoundationDB/" rel="nofollow">https://github.com/FoundationDB/</a>
Apple is the worst acquirer in the industry, at least for users of the acquired company's products. Nobody else would dare kill a _database_ with no warning, explanation, or migration plan - not even a goodbye blog post!<p>Whenever Apple acquires anything that runs on a competing company's platform, that version is immediately killed (see any of their mobile app acquisitions).<p>Thanks for making things that much harder for every other database startup.
Do any of you know if Apple has acquired companies which are partially open-source (FDB SQL layer) with a substantial technologist user-base before?<p>Knowing their past behavior would be an interesting indicator of what is most likely to follow here.<p>According to the TC article (and their website), FDB is no longer available for download and there is a "goodbye"-esque type of message on their community site [0]. Their github [1] repos have all now been made private. This seems most unfortunate for anyone who's included FDB in their tech stack.<p>For reference, FoundationDB has been compared to Google's F1 database [2] [3], so this is Apple purchasing a pretty shiny piece of technology.<p>[0] <a href="http://community.foundationdb.com/index.html" rel="nofollow">http://community.foundationdb.com/index.html</a><p>[1] <a href="https://github.com/FoundationDB/" rel="nofollow">https://github.com/FoundationDB/</a><p>[2] <a href="http://blog.foundationdb.com/7-things-that-make-google-f1-and-the-foundationdb-sql-layer-so-strikingly-similar" rel="nofollow">http://blog.foundationdb.com/7-things-that-make-google-f1-an...</a><p>[3] F1: A Distributed SQL Database That Scales - <a href="http://research.google.com/pubs/pub41344.html" rel="nofollow">http://research.google.com/pubs/pub41344.html</a>
So, database experts , what was specific about foundationDB ? Why apple chose this one ? There is planty of [a-bA-B]+DB's out there. Why this one ? , Sry I dont know anything about databases , But I would love to know about them.
I interviewed with FoundationDB and came very close to working there. They were about as smart as you would expect the people behind that kind of technology would be--scalable distributed transactions at speeds many zeroes higher than what was thought possible--and I'm glad to see them succeed, but wondering what's going to happen to their software. Whatever secret sauce made their software I'm not convinced the market can replicate anytime soon.<p>I also can't help but wonder how much my options would have been worth.
This just lines up with what we've seen in the KV space over the last 5 years. Mutating data and key-lookup are all well and good, but without a powerful query language and real index support, it's much less interesting.<p>Quoting the Google F1 Paper:
"Features like indexes and ad hoc query are not just nice to have, but absolute requirements for our business."<p>Cassandra got ahead of this with CQL. FoundationDB saw this coming and bought Akiban to add a SQL layer. But bolting SQL onto a KV store, even a really good one, isn't trivial to do. I'm not sure it ever delivered on the promise of a real query layer.<p>Still, I hope this is a good exit for the FDB team. The KV layer is pretty cool stuff.<p>Full disclosure: VoltDB engineer here.
One of their engineers gave a great talk at Strangeloop (<a href="https://m.youtube.com/watch?v=4fFDFbi3toc" rel="nofollow">https://m.youtube.com/watch?v=4fFDFbi3toc</a>) that shows some of the amazing simulations they ran FoundationDB through. It's well worth a watch if you want to understand the punishment that they put this database through to make it stand out.
Well this is a very unusual announcement.<p>From what I know Apple is a big user of Cassandra and Teradata for iCloud, iTunes etc. Both of which are very solid databases that have been proven to scale.<p>I am not doubting FoundationDB's credentials but it's pretty extraordinary if they are having scalability issues with either.<p>My guess is that Apple plans to create an equivalent to Facebook's Parse. Either that or this is an acquihire.
About FoundationDB, they have this blog post about achieving almost 15 million writes per second<p><a href="http://blog.foundationdb.com/databases-at-14.4mhz" rel="nofollow">http://blog.foundationdb.com/databases-at-14.4mhz</a><p>It's a pretty impressive number, though as always with benchmarks, should be taken with care.
One interesting aspect FDB was offering is the multi-model approach. Fortunately, there are still true open source alternatives for this like ArangoDB and OrientDB.
Can anyone explain what's so special about FoundationDB? Why Apple would want to acquire it? Why exactly FDB and not some other *DB, which we have tons now after last… I don't know… seven years? Just why? I don't get it.
I wonder whether Apple is planning on open-sourcing the code. If they want it for internal use, then open sourcing it could make the whole project less expensive.
Reminds me a little of Apple's buyout of the compositing software Shake from Nothing Real and their brutal "end of life" process.<p><a href="http://en.wikipedia.org/wiki/Shake_%28software%29" rel="nofollow">http://en.wikipedia.org/wiki/Shake_%28software%29</a><p>"Existing maintenance program subscribers had the option to license the Shake source code for $50,000 USD."
What is needed is a middle ground between OSS and proprietary licensing. Towards this end we are attempting to define a new software license that is organized around the idea of a software cooperative. This provides the basis for a sustainable business model that OSS, frankly, does not have.<p>This also helps solve the larger problem of concentration of wealth/power within large corporations as the cooperatives would be distributing the wealth to all participants and not just to founders.<p>It also provides a better path for startups so they don't have to borrow VC money and can get the help of the community to move them toward profitability in the long term rather than taking the first available bag of money from disinterested investors who only want a return on "their" money.<p>Cooperatives are among the most stable forms of enterprise and are run via direct democracy (at least in our view.)<p>To discuss this further, join our mailing list here:<p>groups.google.com/d/forum/coopsource
Now, what I do not understand is that why would Apple wish to make an over-the-top TV set or whatever. I'd say that they have a loyal userbase--if it only consisted of the mac-using coffeeshop-goers, the business probably would still survive--and a sustainable business model, and they are not going in negative direction economically. Why install another arm to your business?<p>I also wonder if it is lawful to remove an opensource project overnight from public access, to which people outside the company might have contributed. I have not used this product before and I know neither its licence nor the amount or existence of its contributors, but if I have had made a significant contribution, I would be very annoyed in this case and have sued--if possible--Apple or whoever responsible for publication of the sources until the acquisition.
Congrats to the FoundationDB team from us at Aerospike. We respect what FoundationDB was doing with NoSQL and ACID and believe this validates the importance of reliability and enterprise-readiness in the NoSQL market.<p>I'd love to hear from anyone who was actually using FoundationDB in development or production, and what other NoSQL open source alternatives you are considering. pcorless@aerospike.com ; <a href="http://goo.gl/KVQxyq" rel="nofollow">http://goo.gl/KVQxyq</a>
The company I'm working for is looking for a way to scale our DB-layer. Anyway, FoundationDB was more or less the only candidate against MySQL.<p>Does anyone know of any good (and proven) alternatives to FoundationDB?
Interesting. I was checking out FDB earlier this month for solving some interesting scalability challenges. I'm quite glad I hadn't decided to re-architect a bunch things based on it!<p>Did any of you get bitten?
Too bad not many people read this <a href="https://twitter.com/ibobrik/status/454205142662119424" rel="nofollow">https://twitter.com/ibobrik/status/454205142662119424</a>
I feel really bad for all other database startups out there. I'm sure this is sending a negative shockwave through every devops team out there to <i>not</i> use up-and-coming databases.
They removed the npm package few hour before our very important deployment. As the result, our client didn't get a feature, we didn't get our money.
Die, FoundationDB, die!