Key Technologies Driving the Stampede to the Cloud

Mark Bagley, VP, Innovation Scouting, BT
127
235
42
Mark Bagley, VP, Innovation Scouting, BT

Mark Bagley, VP, Innovation Scouting, BT

The stampede to cloud is the number one trend in today’s IT landscape. Why finance, buy, install, operate, maintain and protect your own IT infrastructure when you can rent everything? Security concerns about the cloud are passé. If it’s good enough for the banks and the US government, it’s good enough for the rest of us. When you add in the knowledge that a lot of companies never wanted to run so much technology anyway, it’s clear that the shift to cloud is unstoppable.

Several fast-developing technologies are accelerating this transition. In particular, ‘serverless’ computing, containerization and more dynamic network services are coming together to replace yesterday’s static, inflexible IT infrastructure with something much more responsive, agile and intelligent.

No more servers?

Serverless computing means just that you don’t have servers anymore. A cloud provider runs all the infrastructure, servers and operating systems that you need. And whether it is software engineers piloting a new digital service, or a retail business gearing up for Cyber Monday, the cloud infrastructure will automatically provide the resources you need and shut them down when you’re done. In the past, a retailer had to guess at how much capacity they would need to cope with peak sales periods and formally request it in advance. In the serverless model, an event triggers the request for resources, which are intelligently allocated, scaled up or down, machine to machine with no human intervention.

It has always been tricky to predict the computing resources you might need and easy to end up with too much or too little. The costs of unmet customer demands are high: the US IRS website crashed on the day millions of people had to pay their taxes. And the on-demand rental model is very pure commercially. You never end up with more or less than you need.

The idea of serverless computing also turns big data into something anyone can use. When the CEO asks for sales figures you haven’t got, you can turn to a big data service running on serverless and be ready at a moment’s notice, turning those requests around almost ‘on demand’. If the data is coming from multiple sources, perhaps partner systems, perhaps an Internet of Things (IoT) sensor or a drone, the service can coordinate it all via public APIs whether in the same cloud or an adjacent one. With serverless computing, as soon as you ask a question (even via a virtual assistant, Siri or Alexa-type interface), the underlying infrastructure will work together to collect and analyse the data and give you the answer ’automagically’. We don’t need legions of data scientists after all. You could do all the above with regular cloud IT but with serverless you don’t have to set it up and you don’t need to remember to switch the lights off as you leave the room.

Contain yourself

For any enterprise already running 100 per cent on cloud, going serverless will be a natural move over the next year. However, a lot of businesses still have to deal with legacy infrastructure (and some may want to keep resources on premise for commercial reasons). This is where containerization can help. Like real life shipping containers, software containers are a way of packaging something so it can share space with others without revealing its contents. For example, you could use containers to move applications temporarily to the cloud during a big load and then bring them back in-house.

  For any enterprise already running 100 percent on cloud, going serverless will be a natural move over the next year  ‚Äč

The primacy of the network

The network is the ‘longest pole in the tent’; it holds everything else up. And the more you depend on cloud, the greater the importance of your network.

Inflexible, legacy networks are a poor fit for the cloud-based world. Whether using MS Office 365 on your laptop or connecting your data centres with Amazon’s in Northern California and Microsoft Azure’s in Singapore, the network needs to be available, scalable, resilient and secure. Enterprises need network services that mimic the same event-driven, on-demand environment offered by cloud vendors. Programmable networks based on software defined networking (SDN) and network function virtualization (NFV) are the enabling technologies.

SDN allows enterprises to create dynamic networks that match the flexibility of cloud. It offers greater visibility and control of traffic flow, so critical applications and data can be prioritised to suit business demands. For fast-growing businesses or enterprises expanding internationally, it provides the ability to add new sites to the network quickly and simply. NFV simplifies and speeds up the deployment of new functions in the network, for example a firewall or a retailer’s electronic point of sale application. Using virtualization, enterprises can deploy new functions as software rather than devices, enabling them to be rolled out across offices and branches nationally or internationally in hours rather than days or weeks.

Programmable networks being built today using SDN and NFV will give enterprises unparalleled agility. They will allow them to make the most of their infrastructure, using secure gateways to offload non-critical traffic onto the Internet and using highly secure and resilient MPLS services for core business applications and services.

For established businesses, these technologies promise agility and simplicity. But in a situation mirroring cloud adoption, applying SDN across layers of existing network services demands skills, technical capability and security experience. The solution, like cloud, is to buy a managed software-defined wide area network (SD-WAN) and let a service provider manage the hybrid network while you get on with your business.

Getting from here to there

It’s easy to talk generally about technology trends. It’s a lot harder to work out how to use them in your particular situation. Here’s what CIOs should consider:

1. Stop running third party applications in house. Put everything in the cloud. Do an inventory. There’s probably 20 or 30 items you can offload, along with their associated costs.

2. If your IT infrastructure is intrinsic to your business, then take advantage of the cloud with containers. The cloud is especially good for bursts of work.

3. Expect your network partner to provide a secure, managed SD-WAN. Your cloud experience will only be as good as your network.

4. Think and act more like a software house. Focus on the development, not the operation, of new services.

Finally, when the dust of the stampede has settled, how will all this change the IT department? The job of the CIO? Immeasurably for the better. Moving infrastructure to the cloud and developing an agile, software-defined network will free up talented IT people to work with their business colleagues to find new ideas and bring them to life. And CIOs will at last find the time to be business leaders for the digital age, to contribute creatively and commercially to the success and reputation of the enterprise.