Anand Iyengar

Hybrid Cloud Requirements


Last week Greg blogged about the hybrid cloud barrier:

Perhaps the most significant barrier to a true hybrid cloud operating model is a wide gulf of manual processes required to run existing multi-tier apps and services seamlessly across data centers, colocation facilities and public clouds. That gulf is perhaps the largest barrier to the enterprise adoption of public cloud.

A few years ago, through a series of challenges faced after Neopath / Cisco and while at Mayfield Fund, our team ran into how complex this gulf really was, and spent considerable time developing the CloudVelox cloud migration solutions. Before I start talking about what we did to automate the processes and bridge the gulf, let’s first discuss the key hybrid cloud operating requirements. I would like to post what I think are at least the first three:

  1. New and existing apps need to be able to run seamlessly across data centers and clouds.
  2. Critical data center services (including data bases, file servers, authentication, and firewall settings) need to be available to those apps regardless of location.
  3. Hybrid cloud operating models should be free of any lock-in traps, including having to standardize on a single virtualization platform, or modify your apps only with a specific vendor.

I think these three requirements get us to the early business case benefits of hybrid cloud. That is, those deploying hybrid models will have more efficient, scalable and agile delivery of IT services. They will be able to optimize their apps and services across sites and providers in the same way that server virtualization enabled IT teams to optimize their apps and operating systems across multiple servers.

Of course, hybrid cloud operating models are potentially more powerful than just virtualization, which is a technology. We are early, but already see cloud migration, cloud cloning and cloud failover solutions emerging that can transform the way that IT manages, deploys, protects and delivers apps.