5 Double-Edged Things You Should Know about OpenStack for Cloud Computing
Want to build your own cloud company or even launch a public cloud service? OpenStack is an open source software resource that aims to let you do all of that. It competes with other notable software offerings such as vCloud (paying) from VMware and CloudStack (also open). At another level it also competes with cloud service providers such as Amazon, Google and Microsoft that offer instant cloud computing. But if nonetheless you’re chewing over open and home-grown as the way to go, here are some points to ponder while you plan.
OpenStack is Open
Ho ho, sounds obvious, right? As open source software, anyone can obtain and use the source code, modify it and feed those modifications back into the OpenStack community. This is good for innovation and market-driven development. This is not so good for getting consistent, coherent information about OpenStack. The Linux community contained this potential drawback of openness by having a central figure (Linus Torvalds) drive and coordinate development. Who will step up to the plate for OpenStack?
It Has Considerable Modularity
OpenStack offers a basis for IaaS working/computing (Nova), redundant object storage (Swift), networking (Neutron), a management and automation dashboard (Horizon), user directory services (Keystone), backup management (Glance), and billing information (Ceilometer). Additional modules provide database-as-a-service, cloud application integration (via APIs), cloud messaging, disk-oriented storage and Big Data handling (elastic map reduce). On the other hand, vendors have sliced and diced to make commercial offerings that are confused and lack market focus. See paragraph above too.
There are APIs
With the OpenStack-native REST API and a CloudFormation-compatible Query API, OpenStack opens its arms to ‘multiple composite cloud application orchestration’. Unfortunately, not everybody finds the welcome a very warm one. Microsoft reported that ‘almost no-one’ among their customers used the APIs directly, and that significant difficulties had been experienced in setting up and maintaining OpenStack. Now, that is a competitor talking, but it does lead us into our next point…
OpenStack has Enterprise References
Some. But perhaps not as many as the market would like to see. In addition, OpenStack references have sometimes been alluded to by profile, but never named, remaining cloaked in mystery. The OpenStack website shows the company logos of Bloomberg, Best Buy, Comcast, PayPal and Wells Fargo, which helps increase credibility. And vendors like Hewlett-Packard and Red Hat have made commitments to using OpenStack too; as in the HP Helion cloud service, for example. But it’s a telling comment when Red Hat OpenStack director Radhesh Balakrishnan says that the coming year (starting only from now) will be the one in which OpenStack goes beyond the IT, research and university environment to reach other vertical sectors.
With OpenStack You’re Mobile
OpenStack presents its APIs as being compatible with Amazon EC2 and Amazon S3. Client applications can therefore move from one to the other with ‘minimal porting effort’. The job market for OpenStack expertise appears to be mobile too. Competent systems engineers are able to move easily between different employers. Companies who have found the OpenStack talent they need (no small feat in itself) will need to make sure they can hang on to it.
Fingers Crossed
Whatever the gripes, OpenStack still seems promising as a cloud technology of the near-ish future. Its open competitor CloudStack has waned in popularity recently, making OpenStack the main open-source choice. With the federating power and industrial-strength culture of Red Hat and others alongside, we’ll keep our fingers crossed for a more successful, mainstream OpenStack evolution.
