Basho Relaunches to Deliver a Cohesive Big Data Platform(forbes.com)
forbes.com
Basho Relaunches to Deliver a Cohesive Big Data Platform
http://www.forbes.com/sites/benkepes/2015/05/27/basho-relaunches-to-deliver-a-cohesive-big-data-platform/
8 comments
Basho has experienced some MAJOR brain drain in the last year. I wonder if this re-positioning is related to that.
To be honest, the repositioning isn't really a response to the brain drain -- it's a reflection of a changing culture at Basho that was the cause of the brain drain in the first place.
The primary reason I left Basho is that I felt the company was turning into a "don't invent here" culture, the polar opposite of "not invented here". The goal was no longer to solve hard distributed systems problems and build amazing technology, but to just integrate various "trendy" technologies and make an enterprise simplification play.
The problem is that most of these other technologies have major failings and/or punt on the corner cases that us old time Basho engineers obsessed over.
The whole point of Riak was to be the most highly available, fault tolerant, trusted database you could use. You can't just integrate Riak with arbitrary products X, Y, and Z without compromising on those core tenants.
I really hope Basho can deliver on the promises they're making about this data platform product. I really do. But, for various reasons I can't really talk about, I'd say I'm extremely skeptical.
(BTW, for those that don't know me: I'm a former Principal Engineer at Basho and was the lead developer on a variety of sub-systems in Riak over the last four years)
The primary reason I left Basho is that I felt the company was turning into a "don't invent here" culture, the polar opposite of "not invented here". The goal was no longer to solve hard distributed systems problems and build amazing technology, but to just integrate various "trendy" technologies and make an enterprise simplification play.
The problem is that most of these other technologies have major failings and/or punt on the corner cases that us old time Basho engineers obsessed over.
The whole point of Riak was to be the most highly available, fault tolerant, trusted database you could use. You can't just integrate Riak with arbitrary products X, Y, and Z without compromising on those core tenants.
I really hope Basho can deliver on the promises they're making about this data platform product. I really do. But, for various reasons I can't really talk about, I'd say I'm extremely skeptical.
(BTW, for those that don't know me: I'm a former Principal Engineer at Basho and was the lead developer on a variety of sub-systems in Riak over the last four years)
Thanks for the context - I'm not sure if we've met, but I definitely know and respect your work. Of course there are compromises with this approach, but unfortunately I need a set of capabilities that are not currently available from a single source. Ultimately this just another entrant in a field of options, some of which have evolved in a similarly ad hoc fashion (though one I'm looking at has been under development for 25 years). I hope it can at least hold its own.
Would you still recommend using Riak?
Riak is a great product/technology. I have no idea what the future of Riak looks like, but Riak in it's current form is something I still highly recommend. Just keep in mind that Riak works best for pure key/value workloads -- most of Riak's other features are (as they've been for years) hit/miss.
I used Riak for several years and was a big fan (still am of that version) but I migrated to CouchBase because it had some key features that were important to me that Riak lacked (Riak was just a KV store at that point).
I'm very happy with CouchBase and I feel like they are way ahead of everyone else in terms of skating to where there puck is going to be. In fact, at the time I was using Riak they were moving very fast engineering wise and building great stuff- and CouchBase was still figuring out what it wanted to be when it grew up (with the merger of Membase and CouchDB/CouchIO). Now the roles are reversed. CouchBase is the one that's moving fast without breaking stuff.
To be honest, I think CouchBase is the secret sauce of what I'm doing. I laugh at all the MongoDB posts on HN especially given so rarely that CouchBase gets mentioned.
I'm not bashing Riak- it's solidly engineered.
I'm very happy with CouchBase and I feel like they are way ahead of everyone else in terms of skating to where there puck is going to be. In fact, at the time I was using Riak they were moving very fast engineering wise and building great stuff- and CouchBase was still figuring out what it wanted to be when it grew up (with the merger of Membase and CouchDB/CouchIO). Now the roles are reversed. CouchBase is the one that's moving fast without breaking stuff.
To be honest, I think CouchBase is the secret sauce of what I'm doing. I laugh at all the MongoDB posts on HN especially given so rarely that CouchBase gets mentioned.
I'm not bashing Riak- it's solidly engineered.
Good info. Currently considering CouchBase vs Riak for pure K/V use case. Had been leaning CouchBase and this reinforces that.
I would also like to know the answer to this. Specifically, riak CS. It seems to have been passed by other technologies in the space. It also has some really bizarre requirements (like 1 disk per node).
Unfortunately, Riak CS (as of today now rebranded Riak S2, apparently), is something I don't have enough personal experience with to really have an opinion on.
New CEO and CTO a year ago are a likely source of repositioning: http://basho.com/basho-names-new-ceo-and-cto/
I can't speculate on how that all relates to brain drain.
I can't speculate on how that all relates to brain drain.
Given tweets like this one (https://twitter.com/jtuple/status/598230043093897216), there doesn't seem to be much to speculate about.
I've heard a bit from a friend there, so what I really mean is "I think I'll stay out of this one" :P
But yeah, @jtuple, @Roach_B_me, and others have some choice tweets on the matter.
Off topic- your name is a fantastic book!
But yeah, @jtuple, @Roach_B_me, and others have some choice tweets on the matter.
Off topic- your name is a fantastic book!
"Data gravity" ?
It's why data wants to be together!
http://datagravity.org/
http://datagravity.org/
Sounds like Terry Pratchett's idea of L-Space translated to Enterpriselish:
"L-space, short for library-space, is the ultimate portrayal of Pratchett's concept that the written word has powerful magical properties on the Discworld, and that in large quantities all books warp space and time around them. The principle of L-space revolves around a seemingly logical equation; it is an extension of the 'Knowledge is Power':
Books = Knowledge = Power = (Force x Distance^2) ÷ Time"
http://wiki.lspace.org/mediawiki/L-space
"L-space, short for library-space, is the ultimate portrayal of Pratchett's concept that the written word has powerful magical properties on the Discworld, and that in large quantities all books warp space and time around them. The principle of L-space revolves around a seemingly logical equation; it is an extension of the 'Knowledge is Power':
Books = Knowledge = Power = (Force x Distance^2) ÷ Time"
http://wiki.lspace.org/mediawiki/L-space
Well they have to monitze effectively if they expect to retain talent, and so it's definitely related in that way.
It's a nice database, but not well understood. I've found it difficult to communicate how everything fits together to people do don't use this stuff everyday, so maybe this will help? At least it makes it clear that a seemingly complex configuration is perfectly valid and workable in practice.
It's a nice database, but not well understood. I've found it difficult to communicate how everything fits together to people do don't use this stuff everyday, so maybe this will help? At least it makes it clear that a seemingly complex configuration is perfectly valid and workable in practice.
Is anyone able to parse the explanation below for "Data Gravity"? Even with a great product like Riak, it's difficult to build a hypergrowth standalone business. So apparently they got Cloudera-envy, and decided to.. uh.. synergize new paradigms, and reintermediate plug-and-play partnerships, or something.
> Data gravity describes the effect that as data accumulates, there is a greater likelihood that additional services and applications will be attracted to this data, essentially having the same effect gravity has on objects around a planet. As the mass and density increases, so does the strength of the gravitational pull and as things get closer to the mass, they accelerate towards it at increasing velocity. Although services and applications have their own gravity, data is the most massive and dense, meaning it has the most gravity. If data becomes large enough it can become virtually impossible to move. Usually as services and applications interact with data, they cause even more rapid growth of the data itself, creating a continuous cycle of data growth.
> Data gravity describes the effect that as data accumulates, there is a greater likelihood that additional services and applications will be attracted to this data, essentially having the same effect gravity has on objects around a planet. As the mass and density increases, so does the strength of the gravitational pull and as things get closer to the mass, they accelerate towards it at increasing velocity. Although services and applications have their own gravity, data is the most massive and dense, meaning it has the most gravity. If data becomes large enough it can become virtually impossible to move. Usually as services and applications interact with data, they cause even more rapid growth of the data itself, creating a continuous cycle of data growth.
I'm really excited by RiakCS (now rebranded as RiakS2) because when it works, it offers a powerful on-premise version of S3. I've been integrating it with Cloud Foundry for about a year now, repackaging it to deploy with BOSH, but it's been a constant source of frustration.
From an operator perspective it's incredibly hard to install and configure - the onboarding process to create a cluster from scratch is terrible and the outputted logs are often little more than erlang stack dumps. As an example, I've already spent days trying to get riak-cs 2.0 running on a stock ubuntu 14.04 machine. It should not be this difficult to stand up a product out of the box.
Basho's support engineers are generally well intentioned but often their response is "run the following erlang command, and reply to us with the response (also in erlang)". As an operator, I have no idea what effect the provided commands have on my system, now what the output should tell me. Similarly, it took months for their support/engineering team to answer a support ticket about garbage collection - eventually providing us some defaults such that RiakCS would perform garbage collection out of the box. They could not tell us why these settings would work, nor any side-effects of changing low-level parameters.
From a product perspective, Basho do not appear willing or able to support the open-source community around their products - typically every answer we've received boils down to "it depends on your use case" and effectively "it should work, I don't believe you are seeing the issues you are raising".
I really hope this 're-positioning' results in more support for operators and the community in general. RiakS2 has a lot of potential and I hope Basho are able to realize this.
From an operator perspective it's incredibly hard to install and configure - the onboarding process to create a cluster from scratch is terrible and the outputted logs are often little more than erlang stack dumps. As an example, I've already spent days trying to get riak-cs 2.0 running on a stock ubuntu 14.04 machine. It should not be this difficult to stand up a product out of the box.
Basho's support engineers are generally well intentioned but often their response is "run the following erlang command, and reply to us with the response (also in erlang)". As an operator, I have no idea what effect the provided commands have on my system, now what the output should tell me. Similarly, it took months for their support/engineering team to answer a support ticket about garbage collection - eventually providing us some defaults such that RiakCS would perform garbage collection out of the box. They could not tell us why these settings would work, nor any side-effects of changing low-level parameters.
From a product perspective, Basho do not appear willing or able to support the open-source community around their products - typically every answer we've received boils down to "it depends on your use case" and effectively "it should work, I don't believe you are seeing the issues you are raising".
I really hope this 're-positioning' results in more support for operators and the community in general. RiakS2 has a lot of potential and I hope Basho are able to realize this.
Do I understand correctly that Basho is now a platform-as-service company, like Compose or Heroku, rather than an enterprise database product company?
Think data model capabilities like k/v (original riak) and document (riak 2.x integration with solr) soon to include redis and spark capabilities. But you run this in your own datacenter or cloud like aws or containers like docker. We don't offer a hosted solution (although a ton of people ask for it...)
(I work for basho)
(I work for basho)
Looks like a nice product, but I can't seem to find details on the internal message routing... Is it built on Redis pub/sub, Riak itself, or something else altogether?
And I wish the branding was toned down. As it is, I'm reluctant to direct coworkers or friends to the site despite my interest.
And I wish the branding was toned down. As it is, I'm reluctant to direct coworkers or friends to the site despite my interest.
The value add on riak is riak_core[0][1][2] which handles data distribution/replication and the like.
[0] https://github.com/basho/riak_core [1] http://basho.com/understanding-riak_core-handoff/ [2] http://basho.com/understanding-riak_core-building-handoff/
(I work for basho)
[0] https://github.com/basho/riak_core [1] http://basho.com/understanding-riak_core-handoff/ [2] http://basho.com/understanding-riak_core-building-handoff/
(I work for basho)
It's Riak Core.
Biggest thing I can't stand about the new overhaul is the barrier for the open source version - I want to evaluate for my purposes, not give you my email.
Been following Riak for some time now, and I'm hopeful this will do good things for Basho.
Been following Riak for some time now, and I'm hopeful this will do good things for Basho.
Pull the source from github[0]...
[0] https://github.com/basho/riak
(I work for basho)
[0] https://github.com/basho/riak
(I work for basho)
edit: (Although I think Orchestrate used HBase behind the scenes instead of Riak, which is probably a more appealing stack for Enterprises since Hadoop is more accepted than Riak.)