Transforming your infrastructure with cloud services
Promo Infrastructure as a Service can make your computing operations more scalable and agile. It may also be the first step to something larger.
If you’re considering a move to cloud computing, then you’re not alone. An increasing number of organizations are taking advantage of its scalability and agility. The first step for many is to begin using Infrastructure as a Service (IaaS). Here are some things to consider when testing it out, and some next steps that will help you to mature your cloud computing environment.
Cloud computing adoption is moving too quickly to ignore. The Canadian federal government is leading by example, having already invested heavily in cloud through its Shared Services model, and other organizations are following suit. Intel Security’s April 2016 report on cloud computing said that in Canada and the US, organizations would take on average 14 months to devote 80 per cent of their IT budget to cloud computing services overall.
Infrastructure as a Service (IaaS) is the first step for many cloud adopters. It goes beyond simply virtualizing infrastructure including computing resource, storage and networking. IaaS adds a software management layer to allocate these resources as necessary, either automatically or at the behest of a human operator, to maximize hardware efficiency.
IaaS can happen in a remote data centre, removing the need for hardware at the customer site. Alternatively, for more sensitive customers that want to retain control of the hardware, it can happen on their premises.
Citing a SANS report, Intel Security found that four in every five respondents planned to invest in infrastructure as a service. Clearly, there’s a rising wave of IaaS adoption. But how can you ride it?
IaaS in the private cloud
IT teams may be able to use existing hardware infrastructure to deploy a private IaaS environment, but older equipment will only go so far. In planning a private cloud-based IaaS environment, designers must consider performance and reliability. Servers must be sized for the appropriate virtual machine density and storage IOPS, allowing them to scale as part of the new IT infrastructure.
Organizations have several options when sourcing this new hardware. One approach involves experimenting with a small pilot project using hyperconverged equipment. Hyperconverged solutions are ‘cloud in a box’ systems that handle storage, compute, and networking in a single integrated environment often built on commodity hardware, with a value-added software layer that handles cloud orchestration behind the scenes.
The danger with hyperconverged systems is that when IT teams see them working, they will want to expand them. These solutions are often designed to support a certain number of nodes. They must eventually be integrated with other hyperconverged boxes, or with a broader cloud infrastructure, as the cloud project grows beyond a certain size. Another option involves purchasing cloud-ready servers and storage equipment over time, although customers adopting this route may run into budgetary constraints early on, making it more important to implement private cloud IaaS incrementally.
Such incremental deployment is a probable strategy for many companies anyway, as they feel their way around this new approach to computing.
A structured but speedy approach
Researching cloud’s possibilities is key here, says IBM Canada's cloud business unit leader Mark Noppe. “Understand the use cases for IaaS, and rank and prioritize them. Choose a specific use case and implement that, then move along the continuum to others,” he advises.
IT teams can then map these use cases against their own business needs. They may find specific areas in which a discrete IaaS deployment makes sense, enabling them to test their assumptions, enjoy some quick wins, and build a business case to justify further deployments. It is sensible to begin with lower-risk tier two or three workloads. Good targets here include backup and disaster recovery.
But a sensible approach to IaaS implementation doesn’t necessarily mean a slow one. “Ensure that you have leadership in place, and a strategy, so you’re not taking an ad hoc kind of approach,” Noppe says. “But then you double down on your choices and focus on these projects with a sense of urgency and speed.”
Companies moving too slowly run two main risks: the first is to miss out on savings. The second, and perhaps the most dangerous, is to risk being disrupted by competitors moving more quickly than you and then taking advantage of cloud computing’s agility to outmanoeuvre you.
IT teams eager to move quickly on cloud projects needn’t be constrained by capital budgets, though. An alternative is to pay for hardware under a subscription arrangement. Procurement teams will want to balance the overall TCO against available funds when considering such contracts and include an estimate of how long they hope to use their computing assets for before refreshing them.
As companies discover the practical advantages of IaaS, they may decide to complement it in a remote cloud data centre. This can serve multiple purposes. IT teams can use remote locations for disaster recovery data, and there may also be some applications that they feel comfortable offloading to an external infrastructure. These may include not just line of business applications, but development and testing systems, too.
Although it can be more technically complex, hybrid IaaS architectures can also be used for cloud bursting some applications, particularly those that can be easily load balanced, such as web-facing front ends. IT teams that feel comfortable with IaaS in hosted cloud environments often migrate to the next level, according to Noppe: Platform as a Service (PaaS). This provides software development tools and libraries for developers who want to develop applications designed for cloud operation from the beginning. Development and operations teams can then run applications tailored to meet the business’s exact need in a native hosted cloud environment.
“We do see a trend toward starting with IaaS and then graduating to PaaS,” Noppe says, adding that PaaS can provide more cost savings and service improvements. “IBM customers that started with IaaS a couple of years ago are now maturing and using our Bluemix PaaS capabilities to write their applications and run them on the IBM cloud platform. The need for mobile applications and the need to connect to multiple APIs is driving PaaS adoption.
One advantage of using these cloud-based development platforms is that they provide instant access to cloud-based services such as cognitive computing. This concept builds an understanding of a subject area (such as healthcare or web analytics) using empirical data. Cognitive systems then learn from that data, offering new insights and recommendations in a format easily understandable by human operators. It is closely tied to analytics, which is another application area that benefits from cloud computing.
The software behind those services is complex and built upon years of research. PaaS providers can expose them using easily-consumable APIs, and then provide all the compute power necessary to scale them at will. The journey to the cloud may start with IaaS, but it can extend far further. As IT teams gain confidence in their IaaS strategy, they may find that it’s the first step to a more functional and scalable computing future.