Will Open Source Cloud be the rainmaker?

By Subraya Mallya - July 2010 | Topics - Cloud Computing, Open Source

OpenStackRackspace and NASA announced a Open Stack, an open source open standards based Cloud platform. To make it worthy of your attention, they have also enlisted Intel, Dell,  AMD, Citrix and a host of other smaller tool vendors into the program as supporters of this initiative. The OpenStack will be available in an Apache 2.0 License for any ISV, IT Organizations to build their Clouds.

Their mission statement sort of sums it up nicely what they are trying to do.

To produce the ubiquitous OpenSource Cloud Computing platform that will meet the needs of public and private clouds regardless of size, by being simple to implement and massively scalable.

As the two (currently identified) deliverables of this initiative they will deliver a OpenStack Compute and OpenStack Storage for companies to provision and manage large private cloud and storage clusters respectively and being generally available by October 2010.

This seems like a amalgamation of Rackspace’s Cloud Files filesystem that it has already been providing to its current customers and NASA’s Nebula Cloud that has been designed to managed large datasets.

Rackspace has been the primary provocateur behind this initiative. Novel as this idea is and I sincerely hope that the industry for once comes together to define a common standard that all of them can stand behind. That said, I think the true motivation behind it might be driven by some of the posture of other bigwigs in the Cloud.

  1. Primary among them would be the increasing acceptance of Amazon EC2, S3 APIs as a de-facto standard, in the absence of one. CEO of Eucalyptus, the first open source cloud vendor, Marten Mikos mentioned as much in his recent panel talk in Structure 2010.
  2. The coalition of Cisco, EMC and VMWare, coming up with their Vblock infrastructure building blocks. This coalition is causing more corner office meetings in Armonk, Palo Alto and Redwood Shores.
  3. Microsoft Azure getting ever so closer in the review mirror – as I alluded to in my earlier post.

Rackspace has been primarily a public Infrastructure-as-a-Service (IaaS) cloud provider with over 80K+ customers (according to their website) and second only to Amazon Cloud Services.

The pure-play Infrastructure-as-a-Service (IaaS) have run into some roadblocks with security and compliance concerns. Consequently the Private/Hybrid Cloud rhetoric has gained lot of momentum. Rackspace, and so also Amazon, has had to come up with a Private – Dedicated Cloud offering for enterprises that are holding back from crossing the Public Cloud bridge. Without the large enterprises moving to the Cloud, keeping up the current rate of growth is impossible. Private Clouds, while allaying the fear to a certain extent, leads to diminishing returns in the economies of scale, cloud vendors seek to establish.

The Infrastructure-as-a-Service (IaaS) vendors have also been venturing to move up the stack and build vertically integrated stacks. For e.g. Joyent a pure play IaaS player until recently has started building out its platform to support application development on top of their infrastructure.

Back to the OpenStack initiative. As a optimistic technologist, I would love for this OpenStack group to succeed atleast in forging standards for Clouds.  For that to happen they should bring under their umbrella some of the good work being done by Apache DeltaCloud, DMTF, Cloud Security Alliance et al, in driving standards for interoperability, portability and security of clouds. Given that Rackspace is a big VMWare partner (lead in the DMTF process), it should not be difficult to fold the DMTF initiative into this OpenStack process. Also bringing Red Hat, a big Open Source vendor, and a DeltaCloud supporter into the fold, would not hurt the chances.

Currently, the OpenStack is just starting with Cloud Storage and Cloud Compute. Given that this is a open source framework which seemingly will have many more vendors pledging their support would do well to include additional things like

  1. APIs to support metering, chargeback for thirdparty billing services.
  2. Support industry vertical clouds (healthcare, insurance, retail) and the related compliance mandates.
  3. Support standards based Identity Management (OpenID, OAAuth)

Given that HP, Oracle, Google, SAP, IBM, Microsoft, Amazon, VMWare, Salesforce.com have not pledged their allegiance (or not) about this initiative – the success of it is anything but guaranteed at this point.

My personal take:

  1. If the cloud computing, as it is, has caused enough of unknowns to companies, thereby holding them back from embracing it – making it open source just makes it  worse. It still does not resolve the current concerns around security, portability unless the big players come together on standards. CIOs now have to add a few more layers of due-diligence to make an educated decision.
  2. I don’t think by just being Open Source suddenly delivers any significant value to customers that are looking at Cloud Computing. At best it might appeal to the so-called “Private Cloud” pursuers.  But then again, we are talking about IT organizations trying to do their own Cloud thing. I am yet to believe that many companies are capable of creating/managing their own private cloud infrastructure.
  3. Also if Rackspace, Intel, Dell (all B players in Cloud so far) are combating competition, just making something open source does not make it a game changer without more established players pledge their support.
Back to Top
%d bloggers like this: