Monday, July 2, 2012

Evolution of "Hybrid Cloud"


For the past couple of years, we've heard the term "Hybrid Cloud" discussed as the nirvana that Enterprise IT is truly looking for as they consider new operating paradigms to better align business needs and evolving technology. "Hybrid Cloud" would be the bridge between the known entity ("Private Cloud", "resources within your Data Center", "virtualized legacy applications", etc.) and the new, dynamic entity ("Public Cloud"). There was just one small problem - this reality was a lot easier said than done (previously written about herehere, and here) . The list of challenges included:
  • Limited consistency between Public Clouds and Private Clouds (hypervisors, networking models, security models, authentication, APIs, etc.).
  • Data Gravity - the challenge of moving huge amounts of data (terabytes, petabytes) across the network in a reasonable amount of time, or at a reasonable cost.
  • Network Mobility - the challenge of trying to move a workload (VMs, applications) between different networking models.
Some of these challenges has been partially addressed by early technologies such as:
But none of these have achieved significant market penetration to date. This is partially because the initial users of Private Cloud vs. Public Cloud were different groups (IT vs. developers). Additionally, most legacy applications weren't designed for the dynamic scaling or availability advantages a hybrid cloud might offer. Furthermore, many IT organizations didn't want to be locked into a single Cloud Computing model that didn't give them price flexibility, geographic flexibility or application flexibility. In spite of all of this, the appeal of "Hybrid Cloud" is still there for many Enterprises that wish to find a balance between control and flexibility, between internal and external resources.

In recent months, this space has become somewhat more complicated (new Cloud offerings from GoogleMicrosoft, Rackspace, Tier 3 [podcast] and Virtustream [podcast]) and even fragmented as some experts suggest that "Enterprise Cloud" could actually be an internal-hybrid (web-scale and virtualization segmentation, within the same Enterprise). In addition, we've seen:
  • The introduction of several public and private OpenStack offerings 
  • The introduction of several PaaS platforms based on VMware CloudFoundry (AppFog, Tier 3, Uhuru [podcast], as well as several Microsoft and non-Microsoft .NET environments (Azure, Apprenda [podcast], Tier 3 - IronFoundry, Uhuru)
  • The introduction of several *-as-a-Service offerings to augment both public and private cloud offerings (eg. CloudPassage [podcast], CloudAccess [podcast], Cloudability [podcast]
The great news for IT organizations is that all of these new cloud offerings will drive competition, hence better pricing and flexibility for business needs. But the choices also create some new challenges - how to potentially manage many application across multiple clouds?

[UPDATE - July 2, 2012] George Reese wrote this post with his opinion about VMware's overall Cloud strategy, some of which was altered by their acquisition of DynamicOps.

What had been called "Hybrid Cloud" in the past was probably better characterized as "Hybrid Cloud v0.1". In some cases it worked, and it solved some challenges for the earliest of adopters, but it wasn't ready for mass adoption. But we've now moved into cloud maturity levels that should be called "Hybrid Cloud v1.0", where IT organizations (or lines of business) should be expecting greater flexibility:

  • Companies should be able to deploy applications (DevTest, Production, various Tier levels) to any cloud through a fairly consistent management framework. 
  • Companies should be able to automate the deployment of these applications, whether they are legacy (deployed in VMs), or modern applications (may or may not require VMs; eg. PaaS or SaaS environments).
  • Companies should be able to have visibility and transparency of application costs, application/data geography, and application status/performance across any cloud through a fairly consistent management framework.
  • Companies should have the flexibility to treat IT resources similar to other parts of the business, where different business demands require different resources (costs, regulatory, time-to-market, etc.). They should be able to describe the IT need in business terms, not be bound by cloud-specific capabilities.
  • Companies may opt to evolve their Hybrid Cloud strategy to not be a mix of internal and external resources, but entirely external Cloud resources, across multiple cloud providers and deployment models (IaaS, PaaS, SaaS). 
  • Companies should realize that "the journey to Cloud" may be more complicated from a people and process perspective, so it may make sense to create a fence around the legacy environments and look at more modern process and technologies to move IT forward. 
  • Relying on a single "open API" to drive you Cloud strategy may sound good today, but may be cause for concern down the road.
What will be interesting to watch for the rest of 2012 is if "Hybrid Cloud" customers become more confused by all the new options and constrain their Cloud strategies to single-vendor driven approaches; or if these new offerings drive new demands from management and deployment frameworks to deliver the Hybrid Cloud 1.0 functionality that could deliver new business flexibility.

4 comments:

  1. This comment has been removed by the author.

    ReplyDelete
  2. Excellent article. And just as "...all of these new cloud offerings will drive competition, hence better pricing and flexibility for business needs.", the new challenges mean more opportunities for consultants.

    The old "systems integrator" type of IT consulting firm must evolve into a "cloud integrator" in order to guide organizations in the design and implementation of the many facets of the hybrid cloud. Ray Noorda of Novell used to say, "In mystery, there's margin." Hypbrid cloud provides plenty of mystery.

    One of the areas that cloud integrators can help shed light upon is an accurate financial analysis not only on different hybrid cloud options, but also of an organization's true cost of providing IT services (something that many organizations don't have a good handle on). This will enable design of a much more cost-effective hybrid cloud architecture as well as for implementing a charge back system that optimizes resource utilization.

    ReplyDelete
  3. Steve,

    Thanks for the comment. "Integrators" now face an interesting crossroads. With vendors pushing "converged solutions" and cloud providers offering "on-demand solutions", it forces the integrators to decide if they invest in people (architects, consultants) or software (development) expertise. Both can be profitable, but both are also considerably different business models that traditional "systems integrators" delivered in the past.

    But as you said, "in mystery, ...". These are definitely confusing times, and will probably get more confusing (than less) as consolidation happens thru 2012-2013

    ReplyDelete
  4. The problem I see here is that digital scanning services still have this problem with synchronizing between the local and enterprise servers, that's why they still can't integrate systems fully.

    ReplyDelete