Cloud computing has evolved from the hype cycle of the last few years, to being an integral part of the Enterprise IT strategy as well as a fundamental service provider offering.  The types of cloud constructs have evolved as well – public, private, hybrid and community clouds are all the basic variants, with more sophisticated application-specific cloud offerings continuing to evolve.

While the journey to the private cloud has been continuing and relatively maturing, at least in the more developed countries, and public cloud services offerings are becoming relatively ubiquitous, adoption and deployment of hybrid cloud offerings have had a relatively modest uptake.

The reason for this is not because the allure of hybrid clouds is unappealing, or that it has few use-cases. It is quite the opposite.  There are several use-cases all of which are applicable to real-world IT deployments today:

  • Workload migration:  Seamless migration of workloads from the data center or private cloud to the public cloud for better capacity utilization.
  • Dev/QA operations:  Testing of new applications can induce requirement for additional temporary capacity and having an extensible hybrid cloud is quite appealing, instead of investing in on-premise infrastructure.
  • Cloud-bursting: To handle the needs of bursty applications, temporary capacity allocation in public cloud environments can be extremely cost-effective, providing the convenience of “infrastructure-on-demand”
  • Disaster recovery: Providing data resiliency in case of failure of on-premise resources

If the use-cases are real and the benefits are so apparent, why have Enterprise not gone all out to deploy more robust hybrid clouds? Why have only few Enterprise and selective applications followed this model?

I can think of a few. To make it real, let’s consider the use-case of migrating a virtual machine (VM) from the private cloud to a provider cloud, as an example to illustrate some of the challenges:Shangri-la

  • Operational complexity:   Consider the image above. Once a VM is selected to migrate, it can be quite the task to handle the right documentation, including all the L4-7 policies, convert it to a provider relevant format (which could vary depending on provider), initiate a separate VM instance from template and start it in the cloud. Now this has to be done in a secure way with reconfiguration of the policies and security profiles in most cases.  Bottom line – it is not simple.
  • Inconsistency: The different formats and opaque private and public cloud environments can induce inconsistencies, and lack of consistent policies can further erode transparency
  • Visibility and management: Can the operator responsible for the private cloud have the same degree of visibility and management as the workloads migrate from one environment to another. Can they still be accountable for “hard SLAs”, or do these become best effort.

Reducing any (or all) of the above challenges would go a long way in lowering the barrier to the hybrid cloud deployment. It can bring in operational simplicity, rapid provisioning and an accelerated time-to-market enabling IT and business agility.

As we tackle the problems associated with building high-performance, scalable cloud infrastructure, solving these challenges of extending the private cloud into public cloud environments while preserving consistency and doing it in a simple enough manner is an equally important consideration.

For those of you wanting to learn more about some of these challenges, what Cisco is doing to address them, I invite you to tune into a webcast I’m hosting with some of  Cisco’s top data center and cloud executives  on February 5th, 2013. Joining the webcast to lend their perspectives, will be Cisco’s CTO and Chief Strategy Offier Padmasree Warrior, SVP of DataCenter Group David Yen and VP of Cisco’s Network Operating Systems Group Ayman Sayed. We will also be joined by Andre Kindness from Forrester Research to provide industry perspectives on this topic, and by two special guests providing a customer view point – Kerby Lyons, VP of Global Network Engineering  from SunGard Availability services, and Greg Sanchez, CTO of General Dynamics IT.

The webcast is titled “Fabric Innovations to the world of many clouds”. I invite you to register here. 

I can assure you that it will be an hour well spent.


Shashi Kiran

Senior Director, Market Management

Data Center, Cloud and Open Networking