Some interesting news lately from dotCloud – firstly – docker.io could well be more important to PaaS and linux containers than things like CloudFoundry – it is a way to package up apps and dependencies and have them run on any (linux) platform – this is really about making LXC (linux containers) easily accessible to developers. Something like this could be the missing link for PaaS.
Also – dotCloud are retiring their “sandbox” free hosting tier – but the sweetener is that they are open souring the bits needed to make it run on your own servers
Read about docker.io here and dotCloud open sourcing here.
Over at cloudscaling Randy Bias writes how VM Ware are sounding worried.
This is a recording of our first Google Hangout – recorded live – this is just the audio. You can watch the video here.
1. Framework popularity stats (http://developer-blog.cloudbees.com/)
1.5: Clickstarts and Clickstacks !
2. Recap of citcon conference http://developer-blog.cloudbees.com/2013/02/citcon-conference-2013.html
3. GUI testing (web apps) on DEV@cloud http://developer-blog.cloudbees.com/2013/02/web-app-guibrowser-testing-on-cloudbees.html
4 AngularJS clickstart – now available. Test driven web app UIs. http://developer-blog.cloudbees.com/2013/01/angularjs-html-for-apps-try-it-now-as.html
5. Play 2.1 (one of our most popular frameworks) new release http://developer-blog.cloudbees.com/2013/02/play21-is-out-and-how-amazing-are.html
6. Websocket in preview release
PHP on CloudBees with
Updates to CloudBees SDK 1.3.1 (need to see what’s new here)
8. Jenkins Enterprise
Upcoming bundle 1.480.2.10
Update on the past security advisory
Recap on Jenkins Survey
Heroku have posted a blog explaining the situation, and how things have changed over time.
It seems that some of the blame lays with stale documentation, but also rails:
On the Cedar stack, the root cause is the fact that Cedar is optimized for concurrent request routing, while some frameworks, like Rails, are not concurrent in their default configurations.
In some ways these issues are a result of the huge scale of Heroku – however in this case it appears that end users are paying the penalty for that scaling challenge (rather than benefitting from it).
It wasn’t too many years ago people were complaining about surprising bills and issues of scale with Google App Engine …
It seems VMW has sued 4 ex-VMW employees over “unfair practices“. Read: staff poaching.
Where this gets interesting is that these 4 employees are working for Hortonworks, one of the top leading companies in the Big Data space, and that Paul Maritz announced a few days ago, that one of Pivotal Labs’ first announcement, in a week (25th of Feb 2013), would be related to … Hadoop!
This leaves a number of questions open: Is this the real launch of Pivotal? Since Hadoop will be part of Pivotal, why isn’t Pivotal suing (instead of VMW, since VMW doesn’t own any Hadoop business of its own anymore)? Is this the “counter strike” for a failed/refused acquisition?
Any other thoughts?
GigaOM had an article talking about contextual (work within the contraints and context of a framework or language) vs composable paases, an interesting attempt to document the evolution of paas styles over the years:
Modern versions of Heroku, EngineYard, CloudBees and other also exhibit more of this approach than “first-generation” PaaS systems.