When you hear the phrase, "Don't call us, we'll call you," it's usually a bad thing. It means the end of the line for a sales pitch or a job search lead. But in the world of application integration, it's actually a very powerful design strategy.
For events that occur at unpredictable intervals, the act of polling a server to see if something has happened can become a waste of CPU resources and bandwidth. To see why this might be the case, imagine an analogous and more familiar situation -- a child repeatedly asking a parent, "Are we there yet?"
Granted, application servers don't run out of patience like humans do, but they do tend to slow down as their resources get overused. Having a way to tell an application, "Notify me when event X happens over on your side," becomes a much more efficient approach because communication happens only when needed.
Last summer we introduced a feature in our corporate blogging platform, called a callback, that allows a client to specify that a notification message be sent when events like post submissions, comment approvals, etc. occur within our application.
The message is delivered as an HTTP POST operation. It contains a digital signature to verify authenticity. Moreover, it can be delivered over SSL to ensure that only the recipient sees the data payload.
Callbacks make it possible for Compendium to be integrated with other applications and online services. All the client needs to do is supply an adapter service, which is usually just a server-side script that processes the data and then forwards it on to a third party service that expects it in a different form. In the figure below, a client has developed an adapter service that submits a request to a service that uses SOAP when an event of interest has occurred.
Additional information on implementing callbacks can be found in an online guide at the Engineering team's documentation website.
For events that occur at unpredictable intervals, the act of polling a server to see if something has happened can become a waste of CPU resources and bandwidth. To see why this might be the case, imagine an analogous and more familiar situation -- a child repeatedly asking a parent, "Are we there yet?"
Granted, application servers don't run out of patience like humans do, but they do tend to slow down as their resources get overused. Having a way to tell an application, "Notify me when event X happens over on your side," becomes a much more efficient approach because communication happens only when needed.
Last summer we introduced a feature in our corporate blogging platform, called a callback, that allows a client to specify that a notification message be sent when events like post submissions, comment approvals, etc. occur within our application.
The message is delivered as an HTTP POST operation. It contains a digital signature to verify authenticity. Moreover, it can be delivered over SSL to ensure that only the recipient sees the data payload.
Callbacks make it possible for Compendium to be integrated with other applications and online services. All the client needs to do is supply an adapter service, which is usually just a server-side script that processes the data and then forwards it on to a third party service that expects it in a different form. In the figure below, a client has developed an adapter service that submits a request to a service that uses SOAP when an event of interest has occurred.
Additional information on implementing callbacks can be found in an online guide at the Engineering team's documentation website.
Comments for Compendium Callbacks: An Easy Path to Event-Driven Integration
Leave a comment