Infrastructure as a Service (IaaS) makes developing cloud-based infrastructure simple. But that simplicity masks highly complex pricing practices that can lead to underutilized resources. Iaas optimization is the key to delivering maximum business value.
It’s easy to lump the monthly spend into a single expense category, but CIOs can do better and reduce overall costs by identifying and monitoring individual areas of use and evaluating their efficiency, then adjusting contracts and usage to optimize on a granular level.
IaaS reduces the pain of traditional infrastructure management because the necessary computing platforms have already been set up and made available for immediate use. What’s more, they offer instant and unlimited scaling on demand meaning IT never needs to install another server, networking equipment or storage system. While this reduces the management tasks within the organization, it doesn’t come cheap.
IT leadership must understand the services and costs associated with their IaaS to balance performance against costs. Put another way, plan to optimize IaaS to take the best advantage of it continuously.
Five factors to deliver IaaS optimization
#1 Start and never stop
One significant difference between managing internally hosted infrastructure and IaaS is the upgrade/expansion process. IaaS handles expansion automatically as services and storage capacity is needed to accommodate increased loads. This is a seamless process and one reason for the popularity and expansion of cloud deployments.
Internally hosted systems require distinct processes including purchasing, plugging in, configuring, and optimizing the resource.
Most IaaS platforms aren’t monitored to a deeper level than is convenient because billing is challenging to decipher, and charges fluctuate minute by minute. That means finding the right balance of utilization vs. price requires continuous vigilance and evaluation.
#2 Contract specifics
Initial IaaS contracts can be simple enough and provide pricing based on utilization, and even provide caps to limit run-away processes. But the particulars of the contract can produce invoices with thousands of lines of detail that are difficult to understand and evaluate. CIOs should be aware of the details included in their IaaS contracts and develop methods to assess actual billing against utilization with an eye toward optimization.
#3 Everything changes
As IT becomes more comfortable with mixed cloud presences, there’s a tendency to add new processes to existing IaaS providers. CIOs need to evaluate new instances with the same eye to detail they used when they first started. Even though a new app instance may be the same as another added earlier, it may be used differently or with different frequency. Also, evaluate IaaS provider charges change over time - especially when adding new services.
#4 Know that over-allocation is the norm
The nature of IaaS systems is that they are flexible and are designed to automatically increase capacity when applications call for more. But baseline capacities can be overestimated when services are first set up in anticipation of specific loads. IaaS systems don’t automatically adjust baseline allocations, and when those base allocations prove to be more than required, IT gets billed for more than they use. Even automatic increases need to be reviewed regularly to assure they return to appropriate levels for each application.
#5 Network matters too
Network connections can enable or cripple a cloud-based infrastructure. This is particularly true in mixed-cloud architectures where multiple cloud providers support various applications that interact. It’s also essential in hybrid-cloud environments where locally-installed computing couple with cloud-based services. Network efficiency needs to be reviewed and monitored continuously to detect performance issues that cause bottlenecks and data loss that interferes with application performance.
Each of these components affects the performance of the enterprise application environment and the costs of running them. It’s easy to let IaaS systems automatically tune themselves to deliver high performance, but they may allocate more than is necessary in pursuit of better performance. IT leadership must step in to mediate the trade-off between excellent performance and delivering services at the lowest cost.
»Read more: Eliminate Cloud Sprawl
Optimize by focusing on business outcomes
Organizations optimize infrastructure spend when they focus on business outcomes first, and technology second.
IaaS is just a delivery mechanism for technology—by itself, it’s no more likely to be optimized than on-premise infrastructure or hybrid solutions.
To drive optimization, infrastructure and operations (I&O), teams need to limit IaaS to the business outcomes it’s most suited to support. Those limitations define the size of the IaaS footprint.
To achieve optimization goals, focus on the business outcomes that you are looking to deliver.
Here are four business outcomes that IaaS can support:
#1: Fund innovation
- Tie IaaS to RTB and GTB spend categories
Through rationalization and consolidation, organizations reduce their run-the-business (RTB) spend to fund innovation. But some organizations build a technology-centric IaaS strategy (for example, pushing 30% of infrastructure spend to the public cloud by end-of-year) that blurs the distinction between RTB and grow-the-business (GTB) spend. Without these categories, IaaS RTB cannot be identified and reprioritized to fuel innovation.
- Maximize infrastructure you are already paying for before expanding to IaaS
A poorly thought out push to IaaS leaves I&O with excess legacy infrastructure. On-premise infrastructure will always have excess capacity to cover peak-utilization needs—it’s an operational imperative. But that capacity will spike after a move to the public cloud, and I&O needs to either use or decommission it.
Decommissioning fixed assets carries a financial liability that throttles innovation. I&O teams want to turn off unused hardware, but a server array decommissioned 12 months into a 48-month depreciation schedule immediately places all the remaining depreciation as an operating expense on the balance sheet. The benefit of the asset is realized in a year; the cost must be recognized in that time too. There are scenarios where this is valid (a failing project with sunk costs), but this is bad news for organizations looking to fund innovation using IaaS. How can I justify IaaS when I’m still paying for decommissioned on-premise capacity? The better approach is to optimize on-premise assets you are already paying for before leveraging IaaS.
- Leverage IaaS to for cyclical innovation patterns
IaaS is not a one-size-fits-all solution to deliver innovation. Application development work is a natural fit for IaaS: building up and tearing down fixed-cost, depreciating assets that you own is more expensive than using as-you-need-it IaaS. But innovation and experimentation may be so ingrained in an organization’s development culture that the concept of tearing anything down may seem ridiculous. If innovation is permanent, and the infrastructure is available, development work should stay on-premise; an innovation agenda with peaks and troughs (application development project portfolios) is most suited to IaaS.
- Identify RTB IaaS spend and clarify fixed cost commitments
Organizations looking to IaaS to fund innovation must also assess their fixed-cost obligations. On-premises infrastructure, with a locked depreciation schedule, needs to be used effectively before considering IaaS. Otherwise, IaaS will add costs without any reduction in legacy on-premises costs.
#2: Increase employee productivity
All infrastructure platforms can deliver business value, but the winner in a bake-off between IaaS, on-premises, or hybrid infrastructure is the solution with the best chance of adoption. For example, an application development team wants the agility of the cloud to spin up and down IaaS as they need it. They don’t wish to corporate IT to provision an on-premises solution. The team’s culture is biased to IaaS, and I&O understands that any other solution faces a harder path to adoption. That’s not to say there aren’t other considerations (the application development team may have a workload better suited to an on-premise solution that is cheaper to run and support than IaaS), but team culture influences adoption and productivity.
#3: Mitigate change management risk
Hybrid infrastructure, with a blend of on-premises and public-cloud offerings, can be a transitional step to IaaS. A large infrastructure change (for example, a reduction in infrastructure demand due to an Office 365 migration) doesn’t usually happen in just one switch—an incremental approach minimizes change management. Hybrid infrastructure supports this incrementalism. An existing legacy email service runs in parallel with the roll-out of Office 365. Eventually, the roll-out is complete, and the freed-up infrastructure is repurposed or decommissioned.
#4: Strengthen vendor relationships
An I&O team needs to have the right skill set to optimize IaaS. The person with in-depth technical knowledge to run a data center may not have overlapping expertise in brokering and managing a cloud vendor relationship. They aren’t at fault for this blind spot; it just hasn’t been part of their job remit. Organizations need to commit to providing I&O with the skills to manage IaaS vendors. If they don’t, they are unable to optimize their IaaS spend.
»Download these eBooks to optimize hybrid and cloud by right-sizing legacy spend, shifting savings to innovation and efficiently embracing cloud:
- Eliminate cloud sprawl
- The essential guide to building a cloud center of excellence
- Forrester report: Build a pragmatic cloud strategy that delivers real value to your organization
»Read next, on Emerge: