Icehouse, the next release of the OpenStack open source cloud, will be coming in April 2014. Icehouse will bring automated installation (TripleO), bare metal provisioning (Ironic), a queuing and notification service (Marconi) and database as a service (Trove) as integrated services. Better support for Hadoop (Savanna) will be put into incubation.
The roadmap for Icehouse was determined in a design summit alongside the main OpenStack summit held 5-8 November in Hong Kong. The release features and status, and design summit discussion notes were made available from the roadmap page in mid January.
The lineup of OpenStack core components have remained the same since the Folsom release in October 2012, but improvements and tidying up continues. Icehouse should see the deprecation of networking features in Nova (the compute service) in favour of Neutron (the networking service).
The service lifecycle changed with the April 2013 Grizzly release by the introduction of the integration stage (between incubation and core). This has allowed new services to be added without further expansion of the core, which is important to adopters who just want to build a basic infrastructure as a service. The October 2013 Havana release was the first to take advantage of this with orchestration (Heat) and metering (Ceilometer) being delivered as integrated services.
In an updated ‘State of the Stack’ presentation Randy Bias (founder of OpenStack vendor Cloudscaling) notes that OpenStack is now the fastest growing open source community, with over 12,000 individuals involved. 910 contributors are credited in the Havana release, and the overall developer community has now surpassed a thousand. Despite its exotic location the OpenStack summit in Hong Kong managed to bring in over a thousand extra attendees compared to the previous summit in Portland.
There is however some dissent within the OpenStack ranks, with former Rackspace and Cloudscaling engineer Andrew Shafer declaring ‘OpenStack will leave a wake of bad projects that people unwittingly mistake for operable software solutions’. Shafer is particularly critical of OpenStack fracturing into multiple projects for the sake of prestige amongst contributors. The Ceilometer metering project is called out as particularly weak, ‘the idea that monitoring/metering is something that should be bolted on the side of a cloud is almost as nonsensical as adding on reliability and scalability’.
Gartner analyst Alessandro Perilli has also written, ‘Why vendors can’t sell OpenStack to enterprises’. He cites a lack of clarity about what OpenStack does and does not do, a lack of transparency about the business model around OpenStack, a lack of vision and long term differentiation and a lack of pragmatism. Further questions on OpenStack’s future and viability are rounded up by Brocade architect Geoff Arnold in ‘Whither OpenStack?’. This provoked responses from large OpenStack supporters with Brian Che of Red Hat saying, "open source projects are not the same as products" and Scott Sanchez of Rackspace adding, "innovative enterprises" adopt OpenStack "because they've learned that they can move faster when they build their 'enterprise features' at the application layer instead of requiring them to be at the infrastructure layer."