Reining In And Optimizing Cloud Spending

The fundamental advantage of using cloud services to deliver IT resources needed to support daily business operations is the flexibility they allow: on demand applications and instant infrastructure that can be ordered, provisioned, and delivered in minutes without the delays often involved in submitting equivalent requests to internal IT departments or waiting for suitable on-premise architecture to be implemented and configured.

But that convenience comes at a cost. By making it so easy for various stakeholders – from IT personnel down to line of business managers, partner systems integrators and even individual employees – to commission their own services with little or no central management, it can be equally difficult to keep track of who is spending what and where.

Spending Is Difficult To Control

“If account control is dispersed across a company, cloud spend may be recorded in cost centers unrelated to IT,” wrote analyst David Wright in a research report by Gartner earlier this year. “It can take time to uncover all spend points and implement a cooperative governance model. Without coordination, these dispersed and often underutilized environments can easily rack up excessive monthly charges.”

Gartner estimates that “shadow IT” – hardware, software, and services procured without the knowledge, approval, or control of an organization’s central IT department – could represent as much as 30 percent to 40 percent of total IT spending. And much of that stems from the proliferation of infrastructure, software, and platform services – IaaS, SaaS, and PaaS, as they are often called, are the biggies – and other cloud hosted “as a service” solutions provisioned ad hoc using personal credit cards or departmental expense accounts that bypass standard IT purchasing policies and procedures. The analyst firm estimates that 80 percent of organizations will exceed their IaaS budgets due to lack of optimization approaches by 2020.

Managing that spending is even trickier for organizations with a multi-cloud strategy that mixes and matches specific workloads to a diverse range of public, private, and hybrid cloud infrastructure, often delivered by multiple external providers as well as in-house IT departments. Public cloud spend is usually the hardest to track – unfortunate, given that pay-as-you-go capacity doesn’t always come cheap. Yet none of that appears to be putting anyone off. Global spending on public cloud services and infrastructure is set to double between over the next four years according to analyst firm IDC, growing at a compound annual growth rate (CAGR) of over 22 percent between 2019 and 2023, rising from $229 billion to almost $500 billion.

Beware Of Variable Cloud Costs

Warning signs are apparent if cloud spend cannot be traced to business needs, exceeds what is actually required, and/or regularly balloons beyond prescribed or available budgets. Unexpected monthly bills may also stem from lack of visibility into the detail of cloud service billing schemes and tariff frameworks, with the fine print often overlooked in the rush to provision additional capacity at short notice. Buyers can often end up paying for virtual machines (VMs) that come with pre-previsioned software licenses by default that are either unnecessary or unsuitable for the primary purpose intended for example.

Other cloud costs can vary considerably according to where information is stored and workloads are hosted depending on a variety of factors – the cost of labor, property, facilities, electricity, and water for instance – regardless of local data protection laws that may incur additional expense derived from compliance initiatives. While cloud service provider fees are invariably based on time used, the costs of data volumes can also vary for storage (data at rest) and network (data transferred). Inbound traffic may be free but outbound traffic may be chargeable with different rates according to where it is going, while hundreds of thousands or even millions of tiny charges levied every time hosted data is accessed can quickly mount up. Extra fees for IP addresses, resilience, security, backups and patching will also serve to amplify the bill.

The Cost Of Poor Spending Management

Gartner estimates that organizations lacking any coherent plan for cloud cost management may be overspending by as much as 70 percent or more. That is unsustainable for most if not all public and private sector organizations and improvements clearly have to be made. While keeping tabs on that spend is not easy, it is worth the effort if it leads to identifiable cost savings in a short space of time.

“Recommendations for remediation showing unused capacity or configuration corrections quickly pay off,” wrote Forrester analyst Lauren Nelson in the company’s Cloud Cost Monitoring and Optimization Report, with return on investment quoted at between two and six months.

The only question is which of the numerous software tools and services available to IT and accounts departments work best? While cloud service providers themselves often provide basic cloud expense and budget reporting tools to analyze monthly billing for example, few of these give subscribers the depth of detail or the timely insight they need to apply and maintain tighter controls on their spending.

Automated Cloud Cost Management Tools

To bridge that gap a new generation of cloud cost management and optimization (CCMO) tools have emerged. Ideally, they should help IT departments manage cloud utilization according to individual workload, application, project and business cost center for example, and recognize multiple technology environments and workloads stretching from containers and GPU-powered servers to high performance compute (HPC) and machine learning applications. Identifying and sharing available software licenses between different workloads and business departments too can help minimize application costs.

Few organizations have the available man hours to track the details of multi-cloud spend manually, so automated tools that control which applications are migrated and map them to the most cost-efficient hosting infrastructure will boost IT staff productivity and free up scarce resources for other tasks. Auto scaling capacity to match the needs of workloads also takes a lot of the guess work out of provisioning decisions, though the organization must be careful to set accurate thresholds to avoid over supply and under utilization of available on-premises and cloud resources.

Ease of use in configuring those automated tools is another key requirement – simple dashboard-based views listing details of spend, available budget and resource consumption via a single pane of glass. By providing staff with direct visibility of their cloud costs, they are more likely to stay within the allocated budget.

Granular Cost Visibility

Staff need to see exactly where the money has been spent, on what, by who and for how long. Looking carefully at spikes and anomalies in usage and traffic can help to identify underutilization at other periods or where network misconfiguration is creating unnecessary volumes of Internet traffic which could also indicate a security or distributed denial of service (DDoS) attack.

Cloud cost management tools should also be able to pinpoint the relationship between specific applications and workloads and the storage volumes hosting related data in order to avoid continued payments to third party providers for that storage when those projects are terminated. Unnecessary replication and multiple back-ups of either non- or no- longer mission critical data sets too can waste money, irrespective of whether they are hosted on public, private or hybrid architecture.

Most IT departments could also use additional insight into how often data sets are accessed to help them to gauge which grade and cost of storage is best suited to hosting them. That’s the type of actionable intelligence which provides multiple benefits, including a boost to application performance when in demand information is migrated to faster storage media such as flash-based solid state drives (SSDs) rather than spinning disk.

Forrester also recommends CCMO systems that augment cost insights with features that track and visualize their success, help developers access and launch cloud resources, and deliver comprehensive analytics which hone more accurate decisions and instill confidence in IT departments and CIOs. Solutions that send alerts to stakeholders warning them that they are about to exceed their allocated cloud budget should provide an additional boost to that confidence, and go some way to instilling a culture of tighter organizational control across the whole business.

Combine Cost Management Policies With Spend Control Tools

For all their worth, it’s important not to see cloud-cost management tools as a panacea that will instantly solve all overspending problems in one fell swoop. You also need basic cost control practices that all organizations should implement in tandem while also shopping around cloud providers to get the best deal and exploit available volume discounts before subscribing.

Knowing what cloud applications and services are already in play and exactly how much they cost is the key. The alternative could be messy, overly expensive exercise in IT application and infrastructure provisioning that negates every advantage of cloud migration in the first place and leaves the business in a much poorer state of repair than before it started.

“Without a formal action plan, organizations will increasingly waste money, sign inappropriate contracts and obtain substandard service,” according to Gartner’s Wright.

Sign up to our Newsletter

Featuring highlights, analysis, and stories from the week directly from us to your inbox with nothing in between.
Subscribe now

Be the first to comment

Leave a Reply

Your email address will not be published.


This site uses Akismet to reduce spam. Learn how your comment data is processed.