You know the drill: If you keep doing what you’ve always done, you keep getting what you’ve always gotten.
If you’re not careful with your cloud deployment, you’ll give yourself the same headaches you have with your current infrastructure. That spider web of servers, storage, networks and firewalls in your data center will be just as spidery in a cloud environment. You won’t know what’s out there, what functions the different services are performing or how to track them.
The bottom line is that deploying to a public cloud requires some fundamental changes in the way you do things.
Draft New Teams
If cost control and agility are top objectives, you’ll need IT and apps teams organized to fully leverage the cloud environment. Otherwise, you’ll end up with just another environment that is providing no incremental value.
Here, you might consider drafting a combined cloud and application team. At its core, public cloud is a combined set of services that a single team should be able to provision, manage and monitor effectively.
Compare that to the traditional model, where you have individual network and server teams. Under them, there are Linux-specific teams, database teams, and storage and backup teams. Everyone has their own specific area of technical expertise, and they are all trying to monitor, provision and manage these discrete technologies.
There are different standards. There are different queues and processes. It’s totally inefficient, and it creates this whole level of access and role complexity. You end up with too many teams, and you have to submit too many tickets.
The reality is that if you try to follow that old-school model in the cloud, you’ll blow your cloud optimization opportunities right out of the water.
A Better Way
A better approach is to have a specific team dedicated to leveraging the cloud. If they need support and depth on a networking or routing question, they can then reach out to specific areas of expertise. Ditto for a load balancing or operating system issue.
Your cloud-leverage team can set up golden images of the machines, along with applications that can be readily deployed when the app team wants to deploy those workloads. Those standardized images become the underpinnings for auto scaling functionality, and the ability to automatically dial-up and dial-down capacity.
Change Your Management Style
Managing a cloud environment requires a skilled technical staff, either in-house or through a managed services provider. The scope of management will need to include the cloud provider’s services, as well as the systems and resources that are deployed.
In terms of day-to-day management, an optimized public cloud team partnered with the application folks is the model that seems to work best. The challenge is that it’s all software driven. Your team doesn’t have the ability to walk up and touch a physical server or storage array. Scripting and programming become key capabilities. And your team will need to avail itself of the new tools available for monitoring, management and provisioning.
Put It All Together
Just realize that you might not make everyone happy with this team approach. Your applications team might be worried that you’re going to slow them down. There could be concerns from the C suite about compliance and security.
To be sure, there is a balancing act involved. You obviously want the public cloud environment to be as fast and agile as possible. But you need to balance that with the ability to have appropriate request-fulfillment processes. You want to have information travel downstream about why something was deployed, what the application is supporting and what the cost is.
With that information in hand, you can then make better decisions technically, while also optimizing the different cloud purchasing options that are available for the best TCO. And you want to do all that while still governing the environment and addressing any compliance concerns related to what you deploy.
We Are That Team
Speaking of spiders, one fear to overcome is the fear of working with a managed service provider.. Clients are rightfully concerned about whether their providers are going to deliver what they promised Are they going to be easy to work with? Are they going to be an extension of their team?
At Ensono, we are that cloud-leveraging team. It’s how we’re organized. I manage a cloud-delivery team that provisions, monitors and manages all of our cloud services. Our team in the UK is organized the same way, with a single group managing AWS environments for our clients.
It’s as simple as this: Do you want to be afraid of spiders … or do want to avoid a tangled web of infrastructure in the cloud?