By: Kurt Milne
In a classic August 20, 2011 Wall Street Journal editorial, Marc Andreessen pointed out that software is eating the world. He is right. It is an exciting time to be a software developer.
What that means to you, is that somewhere out there right now, someone is furiously coding the next killer app with the intent to turn your industry on its ear.
The key question: how can you and your company make software that eats the world, faster, better and cheaper?
One way is to write a different kind of app. Not the legacy application that fills your datacenter with code written in the developer’s favorite language, that uses middleware or web server of choice, and a database that is optimized for use. Both the code and the infrastructure is tailored IT and custom fit for purpose.
The traditional enterprise application typically ends up as a monolithic blob that is:
- Brittle – any change to application, middleware or infrastructure has a very real probability of causing service failure.
- Hard to support – extensive documentation and training required for new developer to make changes, or for ops support team to maintain over time and then recover from outage.
- Hard to scale – does not sense and respond as business needs and usage levels ebb and flow. Even with virtual servers, adding or removing resources, and moving work from one cluster to another is based largely on manual processes.
Your datacenter is full of them.
By contrast, the killer app that will disrupt your industry is likely to be:
- A mashup of a loosely coupled set of components that each perform a simple task very well.
- That call on-premise, or hosted, or public services (e.g. hybrid service oriented architecture)
- That are designed for highly variable load conditions (e.g. rapid prototype, then fail or scale)
- That leverage virtualized resources (compute, storage, network, security) that can be added, configured, and removed via API call.
Net result, is that your killer app will be different. It will be architected to leverage services that rely on virtual resources (on premise or somewhere out there in the cloud) that join and leave the application as conditions change, and that cause the application topology to constantly shift.
Ponder that for a moment. The app that is going to delight your customers, and make IT a strategic contributor to your business, and drive your stock PE multiple far above your competitors — is going to be a shape shifter.
For an IT operations professional, the shape shifting killer app requires profound changes that needs to be addressed head on. Right now.
As a result, VMware is investing in CloudOps based on four key premises:
- Process and procedure is more important than ever before. How we do things matters. Ad hoc operations won’t cut it when managing a shape shifting killer app.
- Many of the best practices that implement a “change control” based resiliency strategy, won’t carry forward to shape shifting apps. It may be time to let go of some things near and dear that have worked well for us in the past, but that may be holding us back.
- We need a new IT operating model. This may be a controversial statement. But a service lifecycle perspective becomes an important part of a revised model that recognizes and optimizes a fundamentally new set of practices at the apps management layer, the service management layer, and the infrastructure management layer. Something like this may be a good starting point for a conceptual framework.
- And we need a set of management principles and working assumptions that optimize the separation of concerns and white space between those who are focused on apps, services and infrastructure. Not focused on white space between dev and test, or between functional silos.
How do we operate in this new world? Lets work together and figure it out!
Follow us on Twitter at @VMwareCloudOps for future updates, and join the discussion by using the #cloudops and #SDDC hashtags.
- Solomon Hykes, CEO DotCloud, Upcon 2012 keynote – World of Service: the software architecture that is eating the world.
- James Urquart, CTO EnStratus – CNET blog September 2010 IT Operations in a cloudy world.
- Bernard Golden, CIO.com blog Feb 2012 Ensure cloud application resiliency the NetFlix way