Curbing cloud sprawl to keep IT costs down
In this guest post from Henrik Nilsson, vice president for Europe, Middle East and Africa (EMEA) at machine learning-based IT cost optimisation software supplier Apptio, offers enterprises some advice on what they can do to stop cloud sprawl in its tracks and keep costs down.
On the surface, it seems there are plenty of reasons for businesses to jump head-first into the cloud: agile working practices, the ability to scale resources, and boost the resiliency of their infrastructure, for example.
However, enthusiasm for the newest technology creates a tendency for business leaders to make investments without analysis to support strategic decision-making. Cloud is not immune.
Before long, cloud costs can escalate out of control. Usage can go through the roof, or business value from cloud-based applications can plummet and accountability is replaced by a “Wild West” approach to resource use, whereby whoever needs it first, gets to use it.
In this type of scenario, CIOs should take a step back and consider how to harness the power of the cloud to align with the wider objectives of the business.
Managing cloud sprawl can be the hardest part of aligning cloud usage to a business strategy. Cloud sprawl is the unintentional proliferation of spending on cloud, and is often caused by a lack of visibility into resources and communication between business units. Different departments across organisations want to take advantage of the benefits of cloud, but need to understand how this impacts budgets and the goals of the business.
To successfully master the multiple layers of cloud costs , IT and finance leaders need to see the full picture of their expenditure. They need to utilise data, drive transparent communication, and continuously optimise to stop cloud sprawl, achieve predictable spending, and build an informed cloud migration strategy.
Strategic cloud use
Using consumption, usage and cost data to make cloud purchasing decisions is the first step to stopping cloud sprawl at its root. Migration decisions should not be based on assumptions or emotion.
The cost to migrate workloads can be very high, so businesses need to understand not just the cost to move, but also how moving to the cloud will impact costs for networking, storage, risk management and security, and labour or retraining of an existing IT team. They also need to evaluate the total value achieved by this migration and align decisions with the strategic needs of the business.
A key driver of cloud sprawl is the assumption that cloud is the solution to any given business need. Not every department needs to migrate every part of its operation. In some instances, on-premises might be the right decision. A hybrid approach is considered by many to be the best balance – one survey suggested that a 70/30 split of cloud to on-prem was the ideal mix. This enables certain mission-critical applications to remain, while the majority of computing power may be moved to the public cloud.
Visibility through division
When cloud is carved up among stakeholders (from marketing to human resources to IT itself) it can be hard to get a clear picture of usage and costs. Multiple clouds are used for different needs, and over time usage creeps up on a ‘just in case’ basis, even where demand isn’t present.
To get a handle on this, the ever-expanding cloud demands of a business need to be calculated and then amalgamated into one single source of truth. A cohesive cloud approach is necessary across departments, or there is no hope of maximising the potential benefits.
Ideally, there needs to be a centralised system of record, whereby all cloud data (as well as its cost) can be viewed in a transparent format, and clearly articulated to any part of the business. Without this, cloud usage becomes siloed between departments or spread out across various applications and software, as well as compute power and storage. This makes it nearly impossible to have a clear picture of how much is being paid for and used – or equally, how much is needed.
Once a strong base has been established to visualise cloud usage, and different departments can make investment and purchasing decisions, optimisation is key. This may be something as simple as looking at whether a particular instance would be more efficiently run under a pay-as-you-use model versus a reserved spend or calculating the value that has been gained from migrating depreciated servers to the cloud. This can then inform similar future decisions.
Optimising in this way ensures that cloud usage isn’t allowed to spiral out of control. Cloud is a necessary modern technology to fuel innovation, but businesses need to reign in waste in spend and resources to eliminate cloud sprawl and ensure the right cloud investment decisions are being made to support broader business strategies.