The right infrastructure will eventually enable providers to suggest the right services for each customer based on real needs.
When I was in high school I had a job at a fast food restaurant, as many teenagers often do. One of the first things I was taught was “suggestive selling”. That’s the annoying habit of asking every customer if they’d like an additional item with their meal. Like fries, or a hot apple pie. The reason behind the requirement that employees “suggest” additional items is that studies showed a significant number of customers would, in fact, like fries with their meal if it was suggested to them.
Hence the practice of suggestive selling.
The trick is, of course, that it makes no sense to suggest fries with that meal when the customer ordered fries already. You have to actually suggest something the customer did not order. That means you have to be aware of what the customer already has and, I suppose, what might benefit them. Like a hot apple pie for desert.
See, it won’t be enough for a cloud provider to simply offer infrastructure services to its customers; they’re going to have to suggest services to customers based on (1) what they already have and (2) what might benefit them.
Unlike the real-time suggestive selling practices used for fast-food restaurants, cloud providers will have to be satisfied with management-side suggestive selling. They’ll have to compare the services they offer with the services customers have subscribed to and determine whether there is a good fit there. Sure, they could just blanket offer services but it’s likely they’ll have better success if the services they suggest would actually benefit the customer.
Let’s say a customer has deployed a fairly typical web-based application. It’s HTTP-based, of course, and the provider offers a number of application and protocol specific infrastructure services that may benefit the performance and security of that application. While the provider could simply offer the services based on the existence of that application, it would be more likely a successful “sell” if they shared the visibility into performance and capacity that would provide a “proof point” that such services were needed. Rather than just offer a compression-service, the provider could – because it has visibility into the data streams – actually provide the customer with some data pertinent to the customer’s application. For example, telling the customer they could benefit from compression is likely true, but showing the customer that 75% of their data is text and could ostensibly reduce bandwidth by X% (and thus their monthly bandwidth transfer costs) would certainly be better. Similarly, providers could recognize standard-applications and pair them with application-specific “templates” that are tailored to improve the performance and/or efficiency of that application and suggest customers might benefit from the deployment of such capabilities in a cloud computing environment.
This requires context, however, as you don’t want to be offering services to which the customer is already subscribed or which may not be of value to the organization. If the application isn’t consistently accessed via mobile devices, for example, attempting to sell the customer a “mobile device acceleration service” is probably going to be annoying as well as unsuccessful. The provider must not only manage its own infrastructure service offerings but it must also be able to provide visibility into traffic and usage patterns and intelligently offer services that make sense based on actual usage and events that are occurring in the environment. The provider needs to leverage visibility to provide visibility into the application’s daily availability, performance, and security such that they understand not only why but what additional infrastructure services may be of value to their organization.
Of course in order to offer such services they must exist, first. Cloud computing providers must continue to evolve their environments and offerings to include infrastructure services in addition to their existing compute resource services lest more and more enterprise turn their eyes inward toward their own data centers. The “private cloud” that does/doesn’t exist – based on to whom you may be speaking at the moment – is an attractive concept for most organizations precisely because such services can continue to be employed to make the most of a cloud computing environment. While cloud computing in general is more efficient, such environments are made even more efficient by the ability to leverage infrastructure services designed to optimize capacity and improve performance of applications deployed in such environments.
Cloud computing providers will continue to see their potential market share eaten by private cloud computing implementations if they do not continue to advance the notion of infrastructure as a service and continue to offer only compute as a service. It’s not just about security even though every survey seems to focus on that aspect; it’s about the control and flexibility and visibility required by organizations to manage their applications and meet their agreed upon performance and security objectives. It’s about being able to integrate infrastructure and applications to form a collaborative environment in which applications are delivered securely, made highly available, and perform up to expected standards.
Similarly, vendors must ensure that providers are able to leverage APIs and functionality as services within the increasingly dynamic environments of both providers and enterprises. An API is table stakes, make no mistake, but it’s not the be-all-and-end-all of Infrastructure 2.0. Infrastructure itself must also be dynamic, able to apply policies automatically based on a variety of conditions – including user environment and device. It must be context-aware and capable of changing its behavior to assist in balancing the demands of users, providers, and application owners simultaneously. Remember infrastructure 2.0 isn’t just about rapid provisioning and management, it’s about dynamically adapting to the changing conditions in the client, network, and server infrastructure in as automated a fashion as possible.