Choose your cloud with integration in mind
September 6, 2011
Those already working with public clouds understand that integration links should go into existing enterprises systems. For example, customer data entered into a cloud service should be synchronized with customer data in your enterprise systems. Common sense, right?…
Most organizations that use cloud computing understand the need to use interfaces to access both information and services from the cloud, whether direct links from application to application via APIs or through some sort of integration engine. But you may not know that some cloud providers limit the number of interfaces they can provide in a single instance.
The limitations are understandable. If a provider creates hundreds of interfaces into cloud-based applications or server instances, eventually customer usage could saturate network services and other resources. Thus, cloud providers often allow only 10 to 20 maximum interface connections at a time.
That can be a real problem for a large, enterprise-critical system that must support both core interfaces and integration links to the enterprise. That number of integration points can be quite large to support the business processing between the on-premise applications and the cloud-based systems. That required number is often not available due to the performance concerns.
I suspect the amount of interfaces allowed by most cloud providers will go up over time, but some enterprises that adopt cloud computing today will find themselves hitting their heads on this limit in the short term. The reasons for the limits are legitimate, but you need to know about them before you depend on such a service, so you can factor the limits into your usage. Include these limits in your list of questions you ask before you buy.