Is it Time to Look at a Multi-Cloud Network?

The numbers from the 2018 Rightscale State of the Cloud Survey are in – and the data show that multi-cloud strategies are in as well. 81% of enterprises are now adopting a multi-cloud strategy. As far as the NOC is concerned, however, things are a little less clear. Running a multi-cloud strategy is by no means the same as running a multi-cloud network.

Given the popularity of multi-cloud, you may currently be experiencing pressure to adopt your NOC for multi-cloud management. Don’t feel as though you have to hurry! If you go about this the wrong way, you’ll find that your multiple clouds have become unmanageably complex. This article will discuss the right way to build a multi-cloud management system to complement your multi-cloud strategy.

What Does a NOC have to Do with Cloud Strategy?

Traditionally, a NOC looks like a control room with a bunch of people in large chairs looking at a bunch of large computer monitors. The monitors display key information about a single data center and its associated physical and virtual machines. Traditionally, the data center would belong to a single enterprise, and comprise what became known as “private cloud.”

This traditional view is no longer valid, and there are multiple scenarios in which a single NOC might be expected to manage multiple clouds. For example:

  • An organization might retain a private cloud with a dedicated NOC, but also purchase resources from one or more public clouds to scale its operations. The NOC is now handling a hybrid cloud.
  • A company might consolidate its NOC teams, leaving one NOC team to monitor a physical data center in person plus one or more additional data centers remotely.
  • In either of the above example, the NOC’s parent organization might take the form of an MSP providing public cloud services to multiple clients.

Monitoring multiple data centers or multiple clouds goes well beyond the scope of an ordinary NOC. Traditional monitoring only covers one data center at a time, along with its associated storage, compute, applications, and internal networking. Typically, each data center subsystem gets its own screen within the NOC. Adding multiple data centers either means adding more screens or more personnel – or taking an entirely different approach.

Overcoming the challenges of multi-cloud networking

It’s going to fall on network operations centers to create a network architecture that lets them control and maintain multiple data centers from a single pane of glass. This is hard to do with private clouds because many of them rely on incompatible vendor ecosystems – one might be supported by HPE, one by CISCO, one by Juniper, and so on.

Where the public cloud is concerned, administrators will have the same problem. For example, a company may wish to purchase extra long-term storage from AWS, while relying on Microsoft Azure for compute power. It’s difficult to find a solution that will integrate both capabilities into a single pane, because each cloud vendor wants you to use only their product.

Hitachi has begun to solve this problem by placing proxy servers in each subsidiary data center that it wishes to control. The server, known as a T-Box, is connected to the NOC via the web service interface. Each connected T-Box node pushes monitoring and ticketing information to the central server, which notifies the NOC operator in turn. In addition, the central node can push configuration changes to each connected box, which lets a single operator effectively control multiple data centers.

Adapting the NOC for Multiple Public Clouds

In the very near future, NOC operatives are going to need a solution that includes multiple public clouds as well multiple private clouds. Most enterprises are currently running up to six clouds at once. Supporting these clouds is going to require more than proxy servers – it’s going to require the rise of extremely user-friendly automation.

Again, there’s not yet a cohesive tool that lets users tie multiple clouds together with ease. That means making manual connections between clouds, an operation that doesn’t scale when there’s thousands of connections to be made.

Fortunately, there’s an interim technology that may let administrators scale the manual connection process, as well as other rote administration tasks. Known as Robotic Process Automation (RPA), this technology lets users rapidly and easily automate tasks. In a nutshell, the user performs their task once, with the RPA program running. The RPA automatically captures and records their clicks and keystrokes, and then repeats the task as many times as the administrator needs.

It may take a long time for cloud administrators to find tools that let users automate tasks between multiple clouds. In the meantime, tools like RPA will provide customizable, user-friendly solutions which take care of these tasks on their own. Automation will allow data center operators to tie multiple public and private clouds into a seamless whole – and automation as a service represents yet another convenient data center technology provided by Hitachi.

Cloud is here to stay. Harness the power of Cloud before it’s too late!

Add A Comment

Your email address will not be published. Required fields are marked *