Anyone running FDB in production, good luck, downloads are removed from https://foundationdb.com/
I don't understand why anyone would run a closed source database, especially with the open source options available.
I don't understand why anyone would run a closed source database, especially with the open source options available.
The cow was the target all along, not the ant.
The parasites did their thing just fine for each individual organism. But that one random one that happened to cross wires in just the right way ended up doing orders of magnitude better. A rare occurrence, but very advantageous once it happens.
My theory is the parasite community for that species can survive other ways just fine, but has stumbled across a scenario that serendipidously gives it an extra boost at the change of reproducing.
I would imagine that most species that benefit from these complex interactions tend to have left-over backups for survival and reproduction. When species over-adapt to specific scenarios, it seems likely they'd become to susceptible extinction or their populations would be periodically thinned out.
That's one guess.
Berkeley's Understanding Evolution website has a nice overview of all the mechanisms of evolution. http://evolution.berkeley.edu/evolibrary/article/evo_14
I think it was Peeps.
http://www.ted.com/talks/dan_dennett_on_dangerous_memes?lang...
Good luck to all the former FDB'ers.
This is why everything you build your own product on and that can't be replaced in a matter of hours by a competing product should come with source. Worst-case scenario, you fix the issue in source yourself.
1. the new owners aren't interested because they want to product to cannibalise for something else they're doing. Original product will never see the light of day again, no new bug fixes and possibly no more security patches.
2. the support contract cost is hiked up from maybe a few hundred dollars a year to several thousands
or the possibility of:
3. one or other of above but with the product's support staff reduced, dispersed or fired.
#1+3 above happened to us after Oracle bought out a bunch of stuff we relied on heavily for our hosting platform. Was a very painful time.
Like, don't build on top of freaking closed source platforms and after 17 years of OSI, 24 years of GPLv2 and myriads of open-source alternatives available, you'd think people would learn.
Of course. Because Gimp does all the thing Photoshop does
Blender does all that Maya does I'm sure as well.
SCADA systems for industrial control? Just download the first one that shows up on Google.
EDA systems? Sure, there's a GPL alternative. People use Mentor (as an example) because they don't know better.
It actually does, for the majority of the population. Which doesn't matter. Any Photoshop alternative must be a clone of Photoshop.
For most people, yes. Unfortunately Photoshop is pirated, so it isn't being used at face value.
> Blender does all that Maya does I'm sure as well
If you want to build your own stuff, which isn't uncommon at all and if you have a movie budget of 100 million dollars, you might be better off forking Blender. Also, Autodesk's Maya is not a startup and it isn't being used necessarily as a platform, so that's building a straw-man on your part ;-)
> SCADA systems for industrial control?
You mean Eclipse SCADA?
Funny thing is that I'm working on a software system using it right now. My mission is to steer and monitor power plants and nothing on the market was suitable for my needs. Building on top of an open source stack helped tremendously.
But yeah - Pretty much agree. The fact that most oss products are best-of-class makes the choice even easier.
a) look at building your own. What's inconceivable to one person may not be to another.
b) Build support for at least two closed-source alternatives, and make sure you always have at least two current alternatives.
What if startups with high switching costs (DBMS, OS, pretty much anything else infrastructural) added a guarantee to their sales contracts?
For example, it could be that, if they get acquired, they'll open-source the technology, sell it to another for-profit entity that will maintain it, or provide a migration tool.
Even for something like FoundationDB, it'd hardly be any skin off Apple's back to have a few employees spend a few months ensuring that previous customers have some sort of support.
That would be like them putting a huge paper hat on their head, saying "I'm not a good target for acquisition".
Companies buy them so they (also) get their product/IP etc. If those startups have promised to give it away in such a case, then they are not that good of a buy.
That scenario won't hurt companies that are being acquired with the intent of keeping the product running, for obvious reasons. It also shouldn't hurt acqui-hire scenarios.
The only time it might hurt is when the acquirer wants to use the technology internally, but not offer it to anyone else. That's fairly rare, and the downside (potentially scaring away that tiny fraction of acquirers) is much smaller than the upside (making potential customers feel safe).
The typical large closed-source codebase is full of undocumented things and hidden dependencies on random chunks of proprietary environment. Making it usable as open source is a ton of work.
And you can't guarantee that work will happen without putting up some kind of bond or buying insurance, given that the vendor could simply go bankrupt and not honor the contract.
This is what we are now doing for a project with Berkeley DB XML, which hadn't seen updates for five years. When there finally was an update, it was buggy and moved to the Affero GPL 3, which conflicts with other open source licenses used in the project. So, we continue to use the five year old iteration with a small set of patches.
(Lesson learned: once a product is owned by Oracle, prepare your evacuation plans.)
Fortunately it is stable (for my use cases), and it doesn't actually seem ton be that big a deal that it isn't being worked on. Django_REST_framework is a lot nicer though.
In the case of tastypie, I think all of the maintainers have switched jobs at least once in the last few years and at the same time the general Django community has been moving in the direction of simplicity rather than complex generic frameworks. Daniel's list of things he's not interested in implementing in restless is a good list of things which have been painful in tastypie: https://github.com/toastdriven/restless#anti-features
I've only been a minor contributor but I've increasingly found myself favoring really simple views – roughly https://docs.djangoproject.com/en/1.7/topics/class-based-vie... – since I work almost entirely on read-only public data.
If you or someone you know would like to work on tastypie, we're looking for new maintainers:
http://toastdriven.com/blog/2014/may/23/state-tastypie/
If you have an urgent patch, let me know & I'll see about merging it.
I have started using REST framework for other parts of the project, and it seems a lot more consistent with Django's other components (serializers are similar to forms, APIViews are similar to the generic views). In the end that just makes things easier. Its less context switching essentially, which is really useful when I don't touch that part of the project for a few months, then need to update something.
Thanks for the offer though (and thanks for the framework, it has been useful). And unfortunately I don't have the time to help with maintenance.
Often what is done is source escrow. The source is given to a third party, and if anything goes wrong the third party releases the source to the one who purchased the product.
Is it simply a case of the acquirer wanting the technology and maybe the personnel while the acquired company is losing money hand over fist so they just shut the company down?
The acquirer may love the technology and use it internally, but it's expensive to keep it in a public marketplace. It requires support staff, marketing, sales, etc.
There are also acquisitions that are purely about customers, so the startup's product is shut down or rolled into the acquirer's existing products.
So there are a lot of good reasons you'd want to company other than its product, but we can't always tell which reason it is right away.
"Technology" is rarely re-used per se.
I wonder how often this actually works. In most cases where I've seen this (or heard of it happening first hand from people I know) the vast majority of the people you'd want to keep were out the door of the new place pretty close to as soon as possible (meaning, as soon as the contracts allow, or the golden handcuffs are mostly off, or whatever is relevant to the specific situation).
On the inside there tends to be a pretty predictable path: New management tells everyone nothing will change materially, everything inevitably changes very quickly, people get disgruntled and take off for other opportunities (at a quickly accelerating pace as the old guard sees all their former colleagues from the old place leaving the new).
That is to say, people use Oracle, SQL server, Teradata, etc and they are not worried they will go out of business or be sold or otherwise that they will be left in the dark by a sudden shift in business practices and product availability.
The problem is almost entirely with startups, which are an easy target for bigger companies interested in their technology and team skills set. This is even more so an issue because the majority of startups are VC funded and are under pressure to sell or comply with VCs interests.
So in practice you have three ‘safe’ ways to build your infrastructure, which are not mutually exclusive. Choose OSS software, buy from companies such as Oracle, IBM, HANA, etc, and build it yourself. Depending on the expertise of your team and funds available for purchases, as well as qualifies of the available solutions, OSS is probably the safest way, followed by purchased from large corps. Rolling your own infrastructure is expensive, time consuming, requires committing your developers to building infrastructure instead of well, building a product, and may not work in the end at all.
Even the companies that can afford to do it everything in house (Facebook, Yahoo, Twitter, etc) choose OSS for the majority of their needs and build on top of that. Google is Google. However, if that infrastructure is your selling point and what differentiates you from the rest, and/or you have very specific needs and makes more sense to do it this way, doing it yourself can be a great alternative.
We rarely use OSS here, and we don’t use any proprietary infrastructure technology. We have built everything ourselves, and it has worked out great so far, but we have put a lot more effort and resources into that, whereas we could have instead invested on the Product. If we had to make a choice again, in retrospect, we ‘d most likely have gone the OSS way. It’s all about tradeoffs.
FOSS or proprietary to yourself are the only valid options.
So, you don't use a libc, or a web framework, or any library that implements common logic that other programmers have spent ages refining?
If yes, erk, I don't want to be anywhere near that codebase. If not, I'm reminded of https://www.youtube.com/watch?v=YKjPI6no5ng
It happens more often than you'd think. There are still many things in the database world that open source does relatively poorly compared to alternatives.
Though it's not usually write throughput that most of these technologies are worried about. It's usually compression using dsp methods, aggregate stream folding computations, etc... that matters.
From an IoT or sensor network standpoint, 7000 writes/second is an idle server.
If you want a database for blazing fast data-storage and retrieval, there are many options available. You start seeing the real benefits of kdb+/q when you use q to simplify very complex operations that aren't easily done in SQL. Also, the high level operators that q makes your code extremely terse. I've written complex backtesting systems that perform data mining on massive datasets - all in one page of very tight q code!
-- sad production ex-user of Datomic
Would love to hear some honest feedback. Maybe your struggles were because of the tech, earlier versions, bad hardware config, or mis-applied use case?
[0] http://stardog.com/ [1] They've started calling it a graph database, though I think triplestore is the most correct name
Clark & Parsia had a history of open source (eg. Pellet) which was the best in-memory reasoner for a long time IMO... but not a lot of luck getting sustainable business subscription revenue. This led to the switch to dual-license AGPL in 2008 and now closed-source Stardog...
SPARQL should really be everyone first technology to investigate before heading off to anything else. i.e. when you are still pivoting every week you should have the most generic database tech possible. Only when you scale you should specialize.
Presumably, because you have a business, which has a product, which has a nascent feature, which requires some particular set of time-and-space-and-distribution guarantees that no current database on the market makes. This is why, for example, Cassandra was developed.
I would love to find a fast, scalable open source db that implements Foundationdb's features.
How about FoundationDB? ;-)
EDIT: Also, I'm not sure there's production ready free software column-store DB.
Postgresql had it long before Oracle, but it was dropped as being too much of a hassle to maintain somewhere in the 7->8 transition IIRC.
Oracle and Teradata for example are proven databases with official support available worldwide and a talent pool you can draw from almost immediately. You don't get that with most open source databases (at least those that don't have a parent e.g. Datastax, Mongo, MySQL).
I've watched Oracle try to strangle more than one company once they were dependent on their database. One they succeeded, one managed to migrate to PostgreSQL just in time. If you build your business to be dependent on Oracle they have you by the balls; don't think they're not going to squeeze. And IME the worldwide talent pool is much more available and... well, talented, for PostgreSQL or MySQL or any of the major open-source options.
And I never said that people shouldn't use small startup technologies. Only that when you do you take the risk of the company not being around in a few years. And the people who will take that risk are really other startups or early adopters.
Correct. But you are betting the success of your nascent start up on another nascent start up. This is straight up wrong.
For a large company its different. They have all the resources to go into months long migration projects. As a start up, you can't afford time for migrations when you are busy doing the real work.
Enterprise companies will trade technology for stability and supportability. Most of us will flip the other way.
I strongly disagree your stance on Oracle in particular. I had two arrays at a medium size academic library 8 years ago. Anything I had to call about the data base meant a line item for business review due to cost if it wasn't covered by Oracle's service agreement.
PostgreSQL is amazing and I much rather work with that and hire whoever I want with what I want to do either per instance or annual contracts.
I didn't have the time to delve through all the options out there for this purpose, and evaluate each of them, when there are out of the box solutions that were closer to my needs, though not strictly SQL based (Mongo, Rethink, ElasticSearch, Cassandra all come to mind). There is ~6k/month allocated to hosting costs, and ~$40k/month to the handful of people in the IT team... there isn't much wiggle room there for a small company, and everyone wears a couple of hats. The current application is using MS-SQL (hosted in Azure without redundancy) and MongoDB mirrored data for searching against... licensing to get a replicated MS-SQL setup for better availability would be more than our entire next generation hosting budget... If we could have actually talked to someone who wasn't a sales person at EnterpriseDB that could do more than send you a PDF sheet targetted at managers that might have swayed me.
Sorry, will end my mini rant.. in the end, what support I do have from MongoDB (using their backup service), and my experience with actually just using ElasticSearch and Cassandra has been far better for setting up for something resembling high availability/distributed configuration has been easier than even getting a proof of concept PostgreSQL setup working.
I really hope that PostgreSQL gets it together within the next year or so, it would have been my first choice had I been able to actually get some support within a reasonable budget for my needs, or if I actually had the equivalent of a DBAs salary or more to throw at the problem, which I didn't/don't.
I don't know the features of PostgreSQL that you want to use, but I'm totally willing to learn if somebody is going to cover my living costs. But I'm not even going to respond to your job ad if you put "PostgreSQL plv8 REQUIRED" in it.
For that matter, if you think it's simple enough for a part-time DBA, then why don't you just assign one of your existing IT people to learning and implementing the RDBMS that you need? Surely not all of them want to do the exact same job forever. PostgreSQL has excellent documentation.
If I was hiring a full time DBA, I would have put POSTGRESQL DBA as the job title, and made plv8 a feature requirement that I needed/wanted. As it is, there's no budget for that.
Which still sucks, since downvotes can lead to shadowbans.
PS. Just double checked the Guidelines, and this codification is no longer there. However, there's also no guidance to suggest an appropriate reason to downvote.
I believe I have a few invites.
I generally share this opinion. However, in this case no open source solution came close to the features offered by FoundationDB. There are a couple of attempts (like CockroachDB) which could achieve something similar in the future.
One more reason for me to hate Apple now.
This organization has no public repositories.
if you think you've got the secret sauce, and you've actually had to put it into action and still hit five nines, awesome. But IME doing that with something like PostgreSQL is non-trivial (read: costly). That's why FoundationDB looked so appealing (to me).
It's been used in production by big ad houses like AppNexus as well as retailers like SnapDeal. Lots of miles on the code.
Was closed source for years, but went open source about 9 months ago.
VoltDB offers full transactions / open-source. Lots of differences between us and Aerospike / FDB.
You might be surprised how many apps deploy on VoltDB with cross-partition transactions making up a solid chunk of their workload. Yes, they're slower than partitioned operations, but they're still faster than MySQL much of the time.
Most of the apps we see partition very well, especially for writes. The fact that they can run 10k distributed aggregates a second to get a global view is something few other systems can touch.
You pay me and the licensing question goes away. You still get to poke around the source.
MongoDB, RethinkDB, ElasticSearch, Cassandra and I'm sure a number of others.. each of them have HA options (though RethinkDB is still a few months out for auto-failover iirc), and Cassandra has pretty close to linear scalability in production for some very large data loads. I really like each for different reasons, and would lean towards one or another depending on load.
Not to mention, my speculation is PostgreSQL will likely have an in-the-box replication with failover and/or multi-master solution in place within a few versions.
But yeah, what made FoundationDB's SQL-Layer exciting (for me) was:
- For small clusters it was free. - Automatic HA - Operationally Inexpensive (talking admin time/effort/training; far cheaper than PostgreSQL) - Horizontal Scalability
The things that didn't matter for the 99%:
- It wasn't very fast.
I don't have "big data" problems (I could invent some). Most small shops (I suspect) don't.
The problem I do have is 3AM pagers, availability, wearing too many hats, putting dozens of hours into learning and experimentation to get PostgreSQL to use the hardware it's put on effectively, coming up with complex CARP+REPLICATION+FAILOVER plans, ZFS snapshotting because PostgreSQL still can't match the backup/restore process any commercial database had nailed down two decades ago, backing up the snapshots, figuring out how to partition clients into different table-spaces, blah blah blah.
You sacrifice some single-client performance with FoundationDB, but you solve almost every other problem you've got. And you now have the option of deploying a couple extra nodes to exceed your previously fairly intractable TPS milestones.
It's so easy in fact, you can now autoscale your database with your application servers.
And for your 80% of smaller clients it's absolutely free.
Such a bummer. :-(
A natural fit is something like dropbox, evernote, feedly, etc. Every user has his own actor and it is a full SQL database. Which makes the kind of queries dropbox would need very easy to do.
It also has a KV store option. This is basically a large table across all your servers. This table can have foreign key sub tables. From my understanding of FoundationDB this part is closer to what they were offering.
This is not to rubbish it - I've not used it after all - but the claims being made for ActorDB are pretty far away from the claims made for Foundation.
ActorDB fits a specific data model extremely well. Some less so. But that is the case with all databases.
Agreed. And what I'm saying is that it appears that ActorDB's per-shard area concurrency is limited to one writer. And that means that SQLite's concurrency support is (contrary to your earlier post) extremely relevant: not just in terms of pure performance, but also ability to perform concurrent operations. If you need more concurrency, your only choice is to shard extremely heavily (which might mean you require more cross-shard operations, which are apparently slow).
As you say, some data models fit the actor model well, but this is still a far cry from the capabilities that were promised by FoundationDB.
The reason why I said sqlite concurrency support is irrelevant is because ActorDB serializes requests anyway. It must do so for safe replication.
Interesting - I didn't know that. Even so, it depends on what kind of writer concurrency we're talking about, I guess - I presume that ActorDB is limited not just by having to run requests one at a time per-process (which is a legitimate tactic to avoid latching overheads and so on), but by also not being able to run any new transactions against an actor that's received a write until that write commits?
> The reason why I said sqlite concurrency support is irrelevant is because ActorDB serializes requests anyway. It must do so for safe replication.
Do you mean by this that the entire cluster can only perform one request at a time? Or am I misreading you?
Read/write calls to an actor are sequential. I'm quite sure this is how other KV stores like Riak do as well. They have X units per server and those process requests sequentially. Their actual concurrency is basically how many units per server are running. They may interleave reads/writes per node or they may not.
ActorDB does not allow reads while a write is in progress. It is quite possible we will optimize this part in the future as it is quite doable.
As a thought experiment, if the company just went bankrupt and had no one working for it any more, you contract won't help right? Second part, what happens if someone just buys the assets (IP) and closes the company, then hires the old team? There are many possible variants here.