Part 1 – What’s So Unique About the Cloud?
When I talked to customers, I got often asked why the operations management in Cloud is different in their existing data center. This seems to be an innocent but actually is a very important question. To answer that question, we need to examine what is so unique about the Cloud compared to the traditional data center, which could have both physical and virtualized infrastructure. Many things could be the same between Cloud and traditional data center. But there are three attributes which define what the Cloud computing is about.
Elasticity – Think about pools instead of boxes
Many business buyers look into Cloud to solve their immediate business demands that their traditional data center can not solve. Some of those demands could be seasonal. Others could be because of exponential growth that outpaces data center’s expansion plan. The Cloud offers elastic resources that can be scaled up and down to meet those dynamic demands. The economical way to do that from a Cloud provider point of view is to make sure the cloud infrastructure can support larger number of resources. And these resources have to be pooled to provide high utilization and high availability. While a typical enterprise data center manges thousands servers with 20-35% of utilization, a single Cloud could easily to have hundred thousands of virtual machines and up to 70-80% of utilization. In the enterprise, you can deal with the box (the physical server or virtual server). In the Cloud, you have to live with pools of compute units, storages, and network devices. This is a big paradigm shift between the Cloud and the traditional data center.
Responsiveness – Think about services instead of applications or machines
When business buyers look for Cloud to solve their dynamic demands, they not only want the service but also want the service to be available immediately and to run reliably. With thousands or ten thousands of those service request, the Cloud has to make them provisioned, ready to be used by cloud users, and ready to be monitored and operated by the administrators within minutes. In order to do that, you can’t have a manual process for each of those requests. The whole on-boarding (and off-boarding) process needs to be standardized, automated, and self-serviced. Service reliability is not a new requirement. It is also a very important measurement in the traditional data center. However, in the Cloud, due to the dynamics of the demand and large variety of the workloads, to maintain a high level reliability becomes even more critical and challenging. This largely impacts how the operations should be. We will talk about this in more details in part 2 of this topic. But the bottom line is that most of the operations tasks and processes also need to be automated.
Efficiency – Think about sharing instead of owning
While many business buyers look into Cloud to give them business agility, IT buyers mainly seek a way to reduce the costs. More specifically, the fixed cost. Cloud provides the economy of scale for reducing the unit costs. That, combining with the on-demand model, translates into low variable costs to Cloud users. It is worth noting that the economy of scale not only applies to the CapEx through better utilization of commodity resources, but also the OpEx through automated management process.
Shifting the paradigm from boxes, applications, ownership in traditional data center to pools, services, and sharing in the Cloud requires you to evolve your current operations management into a new stage – Cloud operations management. In part 2 of this topic, I will get into more details about what pieces you should consider when you build your operations management in the Cloud.