One of the oft cited reasons in surveys that enterprises aren’t flocking to the cloud like lemmings off a cliff is “lack of control”. Problem is that articles and pundits quoting this reason never really define what that means.

After all, cloud providers appear to be cognizant of the need for users (IT) to be able to define thresholds, reserve instances, deploy a variety of “infrastructure”, and manage their cloud deployment themselves. The lack of control, however, is at least partially about control over the infrastructure itself and, perhaps, complicated by the shallow definition of “infrastructure” by cloud providers.

cloudcontrol Perusing the options available from Amazon can be overwhelming and yet underwhelming at the same time, depending on what you’re looking for. The list of operating systems and application environments that can be deployed is extensive, a veritable potpourii of options. But that’s where it ends – at the application infrastructure. If there’s an option for deploying or managing the rest of the supporting infrastructure I’ve yet to stumble across it. Load balancing. Acceleration. Optimization. Identity and Access Management. Security. Quality of Service. Rate Shaping. Compression. Secure remote access? Certificate management? LDAP? XML gateway? The list would grow quite extensive if you listed out all the components that go into a well-oiled network and application network infrastructure. You know, the infrastructure that’s necessary to deliver applications.

Suffice it to say this is where the shallow definition of infrastructure in the cloud meets lack of control and forces C-level executives to make a decision on which way to go: with the cloud or with control.

THE ARCHITECTURAL DESTINY of IT

When IT is tasked with delivering an application it must necessarily consider available, security, and performance of that application as it pertains to traversing the data center network. When applications perform poorly IT is tasked with addressing the issue in order to adhere to organizational service-level agreements. This is the reason so many solutions and options exist in the network and application network infrastructure: to provide the means by which IT can architect a well-performing, reliable, and secure environment in which applications can be delivered.

When the ability of IT to choose and deploy those integral pieces of the network and application network infrastructure they lose control over the delivery of that application. They no longer have the means by which they can deploy solutions designed to augment or enhance or fix issues commonly occurring with the deployment of applications. They lose control over their own architectural destiny, trusting implicitly in cloud providers to offer the choices necessary to the successful deployment and delivery of applications.

The shallow definition of “infrastructure” used today is impeding adoption of the cloud because users lack choices and options. They lack control over that environment in an architectural sense. It isn’t the turning over of the reins for management and maintenance that’s problematic; it’s the architectural control that’s lost that gives rise to these concerns.

CONTROL IMPLIES CHOICE

Cloud providers are doing the term “infrastructure” a disservice when they focus solely on the application infrastructure – databases, operating systems, web and application servers, server hardware – and ignore that a real infrastructure includes network and application infrastructure. And that without the ability to easily choose and deploy network and application-network focused services in the cloud that organizations are necessarily going to be leery of the ability to meet service-level agreements and improve performance and security levels in the cloud without the control they are typically afforded by a physical data center.

Increased capacity does not always result in improved performance of applications. It is but one way in which solutions attempt to address the problem of application performance. In the cloud, this is the only way to address the problem, unless you’re using Amazon and are keen on adding yet another line-item to your monthly cloud provider bill by signing up for CloudFront. That’s certainly an option, but unfortunately for organizations it’s currently the only option for many in the cloud.

Cloud providers need to recognize the need for alternative solutions addressing performance and security in the cloud and offer their potential enterprise customers not only choice but control over those solutions. If providers are going to offer Infrastructure as a Service then they need to support more than just application infrastructure as a service. They need to broaden their support and inclusion of the same types of options available to the CTO and CIO in their own data center, which includes a much wider array of solutions than is currently available today.

THE PARADOX FOR CLOUD PROVIDERS

Dave Rosenberg of C|NET’s Negative Approach hit upon the reason for the reluctance we’re likely to continue seeing in the cloud around choice and control:

Disaster recovery, compliance, and enterprisey [sic] features are where the growth is in the near term. You have to have an SLA and support for true enterprise-class applications. Amazon will probably do this over time, but right now, you have no real option. Rowell pointed out that for Amazon to offer these services would add significantly overhead and likely cause the price point to rise significantly.

Cost. Right now cloud is selling, albeit slowly in the enterprise market, based on perceived initial and long term cost reductions. Offering choice and therefore the control necessary to implement the infrastructure necessary to ensure compliance with SLAs and “true enterprise-class applications” would require additional investments by cloud providers that would most certainly be passed on to the customer by way of price increases. Increasing the cost of cloud computing would certainly decrease its already lackluster appeal to many enterprises, thus effectively shooting providers in their own feet.

It’s a Catch-22 situation; giving control to customers would make it more appealing technically but less appealing from a financial perspective. This means cloud providers are going to need to evaluate options that keep the costs down while affording enterprise markets the control and choice they need to adopt cloud computing as a viable alternative to in-house data centers.

HOW INFRASTRUCTURE 2.0 CAN HELP

Vendors already understand the evolutionary need for a more dynamic infrastructure that addresses the issues cropping up with cloud and virtualization and emerging data center models. Hence the drive toward Infrastructure 2.0. In addition to its core capabilities around connectivity intelligence (collaboration), elasticity, and intelligence is elasticity.

That elasticity needs to expand beyond the traditional view of “capacity on demand” and include “management on demand”. Support for multi-tenant environments like cloud computing and virtualized architectures is a must. Coupled with a flexible architecture and collaborative intelligence, multi-tenancy allows providers the ability to deploy new functionality “on-demand” on existing infrastructure platforms and extend the ability to control that new functionality to customers in an isolated management structure.

By enabling cloud providers to extend the functionality of existing investments to include many of the solution choices enterprises would normally make in the course of deploying applications and their supporting infrastructure, they can remove “lack of control” and “limited choice” from the list of reasons why an organization rejects cloud computing as an option while ensuring that the “cost” borne by the provider is significantly reduced, thus making it a feasible option to offer these options to customers who may require it.

Providers will, of course, charge for such options but if the investment is not so heavy a burden on the part of the provider then the cost passed on to the organization will certainly be (or at least certainly can be) lessened to the point it is comparable to the investment the enterprise would make internally.

 

Follow me on Twitter View Lori's profile on SlideShare friendfeedicon_facebook AddThis Feed Button Bookmark and Share