3 ms·
I worry about Joyent as a company. They have very interesting talent, but end up doing some sketchy things. For years, they successfully positioned themselves
by BraveNewCurency 11y ago
I worry about Joyent as a company. They have very interesting talent, but end up doing some sketchy things.
For years, they successfully positioned themselves as a "competitor" to AWS in the press. But at the time, they were a simple hosting provider with no API and no self-provisioning. Of course, the press didn't know any better and kept printing that they were a competitor. But even today, Gartner doesn't even consider them a competitor to AWS.
http://www.gartner.com/technology/reprints.do?id=1-2IH2LGI&ct=150626&st=sb http://www.gartner.com/technology/reprints.do?id=1-2IH2LGI&c...
Triton seems to be a _very_ limited port of "Docker" to their custom version of Solaris. For example, it supports only one "--volumes-from" argument per container. And I'd be worried running on an OS supported by basically one company.
https://www.joyent.com/developers/triton-faq https://www.joyent.com/developers/triton-faq
It's also important to note that their S3 knock-off has had an embarrassing 8-day outage. Oops. They literally didn't have any backups, because they didn't think they needed them with ZFS!
http://www.datacenterknowledge.com/archives/2008/01/21/joyent-services-back-after-8-day-outage/ http://www.datacenterknowledge.com/archives/2008/01/21/joyen...
- bcantrill 11y agoDisclaimer: I'm the CTO for Joyent. Normally I wouldn't respond to something that is so clearly a troll, but I wouldn't want someone to stumble on this and confuse what you've said for the truth. So at the risk of feeding the trolls... First, the Gartner magic quadrant that you quoted is for cloud storage, not for cloud infrastructure as a service. We are in the MQ for cloud infrastructure as a service[1], in which Gartner has this to say about us: Joyent has long been an advocate of OS containers, and the growing popularity of container technologies, especially Docker, has been beneficial both for its service and software businesses. Joyent's architecture is well-suited to running containers securely within a multitenant environment, and it is one of the first providers to launch a Docker-based container service. Because Joyent maintains a single codebase across its own service offerings and the cloud infrastructure framework software that it sells, customers can more easily consume Joyent technology in a hybrid fashion. Now, to be fair, we're by no means AWS-sized and we're not trying to be everything to all people -- which Gartner also points out, under "Cautions": Joyent focuses on developing its own technology, and has a track record of releasing innovative capabilities. However, it faces a long-term challenge to compete against providers with greater development resources. Joyent has a capable basic cloud IaaS offering, but its feature set is strongly oriented toward cloud-native use cases, and it is highly developer-centric. On the particular point of "worry about Joyent as a company", Gartner adds: Joyent needs an ecosystem of third-party tools that support its platform, along with a software marketplace, and managed and professional service partners. Although Joyent can take advantage of the growing Docker ecosystem, it must find ways to bring that ecosystem to its platform, and that ecosystem will not by itself be sufficient to support customer needs. I feel that that's a pretty balanced view: we have been longtime advocates of containers; we're focused on hybrid-cloud and Docker use cases; and we have the challenges endemic to being a much smaller player relative to AWS, Google and Microsoft. As for the limitations of Triton and the divergences from Docker, we are very explicit about them[2]. Yes, there are limitations (e.g., missing support for "docker build"), but what is there is production-grade -- and we're busily working on filling in the gaps. (Which, given the degree to which the entire ecosystem is in flux, is better than par for the course.) Further, if you look at the enthusiasm in the community[3][4], it's clear that the choices we've made in Triton solve real production problems. Finally, in terms of that 8-day outage from over seven years ago: Joyent divested from that service six years ago.[5] In its stead, we built Manta, a container-oriented storage service that has been in production since 2013.[6] I'm not seeking to defend the outage that you refer to (though it pre-dates me personally at the company and I know very few of the technical details surrounding it), but it's important to put it in its context -- and to make clear that this hasn't been Joyent's storage offering for the better half of a decade. [1] http://www.gartner.com/technology/reprints.do?id=1-2G2O5FC&ct=150519&st=sb http://www.gartner.com/technology/reprints.do?id=1-2G2O5FC&c... [2] https://docs.joyent.com/public-cloud/containers/docker/divergence https://docs.joyent.com/public-cloud/containers/docker/diver... [3] https://twitter.com/ahl/status/626801451663294465 https://twitter.com/ahl/status/626801451663294465 [4] https://twitter.com/Sirupsen/status/626840507235311617 https://twitter.com/Sirupsen/status/626840507235311617 [5] https://www.joyent.com/blog/joyent-sells-strongspace-and-bingodisk-to-expandrive https://www.joyent.com/blog/joyent-sells-strongspace-and-bin... [6] https://news.ycombinator.com/item?id=5939340 https://news.ycombinator.com/item?id=5939340