A vital question for companies considering cloud adoption is “What makes a good cloud provider?” Basic answers tend to include low costs, a fit for the buyer’s application models and a realistic SLA. But these answers have become table stakes in the cloud market. They’re not enough to separate “good” or even “great” cloud providers from the rest. Enterprises gambling IT budgets on cloud need concrete ways to differentiate cloud providers...
before upping the ante.
Expertise in a similar industry and size. One of the biggest indicators cloud buyers list is they want a cloud provider with demonstrated success in the market. Enterprises look for references, documented applications and sometimes even blueprints that describe how a given set of applications could be built for or migrated to the cloud. Small and medium businesses (SMBs) that often lack the in-house cloud skills of large enterprises want cloud providers with support in planning, pilot testing and cloud deployment. SMBs look for the provider to take a heavier hand on the cloud project, claiming cloud providers can be less expensive and more trustworthy than integrators.
One of the biggest indicators cloud buyers list is they want a cloud provider with demonstrated success in the market.
Support for cloud modeling and testing. Companies of all sizes need to make a business case for cloud adoption; they should understand how differences in application usage and data storage requirements would affect cloud pricing. Businesses with more complex cloud applications -- involving elastic resource deployments, shifting geographic focus or failover and load-balancing with internal IT -- will want to know how a cloud will perform in all likely loads and failure modes. This is beyond what pilot testing will tell you, so modeling is critical. While buyers don’t insist cloud providers have their own modeling tools in-house, they do want the provider to have experience working with cloud-modeling firms and tools.
Proficiency in hybrid cloud application construction and deployment. Interest in failover and load-balancing applications demonstrates the need for the next differentiator -- hybrid cloud. Most businesses won't consider cloud adoption unless internal IT resources can be harmonized with cloud hosting, which is critical for overflow and backup applications. These businesses look to a public cloud provider for advice on how to create hybrid applications with in-house virtualization, multiprogramming/SOA hosts, etc. It's often a requirement for the provider to identify a private cloud “stack” that works well with their public cloud offering.
Availability of management tools for public and private clouds. Complex public and hybrid cloud apps have spurred the need for management tools to help connect and deploy apps to a cloud environment. Early adopters of public cloud services focused on management interfaces. However, a growing majority of businesses testing cloud find that management interfaces matter less than general provisioning and support toolkits that cross boundaries between public and private, management and interface. Some vendors offer these toolkits as “templates” or “frameworks.” Open source projects can address cloud-application provisioning in a flexible and provider-independent way.
Willingness to develop and run a pilot test. Very few companies are prepared to make a major cloud commitment without testing their applications in the cloud. The problem is the cost of a pilot test itself is difficult to justify without some assurance of success. A cloud provider should work with companies to develop a pilot test plan that reduces risks and gradually increases cloud service charges as confidence in the project grows.
A cloud provider with all these traits is a valuable partner; one that can’t promise these differentiating points poses a larger risk. These qualities point to the issues shaping cloud application deployment, so all cloud providers should be expected to have them.
Tom Nolle is president of CIMI Corporation, a strategic consulting firm specializing in telecommunications and data communications since 1982.