AlexOakenman - Fotolia

Tips on reducing unnecessary cloud service costs

As IT departments prepare their 2019 budgets, we assess how to avoid paying too much for cloud services

This article can also be found in the Premium Editorial Download: Computer Weekly: Mapping the future at Ordnance Survey

Infrastructure as a service (IaaS), platform as a service (PaaS) and software as a service (SaaS) free business users from the constraints of IT, but without suitable oversight, costs can quickly get out of control.

As initial costs are relatively low, business managers can sometimes end up being in control of their own cloud budgets, which is simultaneously a good thing and a bad thing.

While as-a-service offerings might enable organisations to work unhindered by IT, cloud contracts need to be managed effectively – yet business managers may not have the expertise to do so. Some might opt for a cheaper, lower-level service or use a free subscription, only to end up paying a premium when they ramp up usage, rather than moving up to a more economical service that meets their scalability requirements at a better price.

For example, agile methodologies used by developers to speed up software development through DevOps can empower them to enable their applications to request IT resources programmatically. But although such an approach offers huge benefits to the business, enabling pilots and projects to consume cloud resources and run unhampered by IT bureaucracy, agile development practices can leave a gaping black hole in the IT budget, with IT managers unable to gain an accurate insight into how much is being spent in the public cloud by their in-house DevOps teams.

Earlier this year, a survey by Densify reported that a quarter of IT departments have little understanding of how much is being spent on public clouds such as Amazon Web Services (AWS) or Microsoft Azure.

For central IT, this lack of visibility on overall costs means companies can end up paying far more than they need for cloud subscriptions and cloud-based services.

AJ Witt, an industry analyst at The Itam Review, says it is generally acknowledged across the industry that there is about 30% wastage in cloud subscriptions, much of which comes down to buying too much. “For instance, you buy a bundle and don’t use all the products, or you’ve got 100 seats and only use 70 of them,” he says.

Just as when a software asset manager assesses on-premise licences against usage, Witt advises business users to evaluate how they use their cloud subscriptions.

“A single Adobe product [such as Photoshop] is cheaper than buying the whole bundle,” he says. “The other pitfall is buying the wrong product, such as a full version of AutoCAD when, for most users, an AutoCAD LT subscription will do, or sometimes even a free viewer is enough.

“The issues we had with on-premise software in this regard are still present with subscription-based licensing.”

Because it is so so easy to buy SaaS, a business may find that marketing, sales and engineering all have their own subscriptions and user counts, and each will have unused subscriptions, which could be pooled if there was only one contract, says Witt.

SaaS is so easy to procure and deploy that it often bypasses procurement and IT policies, he points out.

“Most providers of SaaS optimisation tools focus on discovering this uncontrolled spend by scanning expense and accounts payable records first, rather than looking at more formal stores, such as a contracts database or asset register,” he says.

Without central control and a single record of supplier engagements, says Witt, there is no way of knowing whether the business is paying the right price or using the subscriptions it has purchased.

“You also lose any economies of scale and have less leverage with the supplier, with piecemeal department-level engagements,” he adds.

Over-provisioning

Another problem is over-buying. For instance, traditionally, IT departments would buy resources such as enterprise storage up front, and this investment would support the company’s data growth for a number of years, by which time the initial investment would have been written off. Clearly, the elasticity of IaaS means it makes no sense to buy more than is required at the time.

David Chamberlain, general manager of License Dashboard, says IT buyers can sometimes be over-cautious. “IT people can be worried about things like an Exchange server going down, so there is a tendency to over-provision, and then people will forget to decommission the cloud service when it is no longer needed,” he says. “The cloud is very elastic and is easy to throttle up, which is a big change from on-premise servers.”

Witt says virtual machine (VM) sprawl has always been an issue on-premise, even when companies have had good processes in place. “It is always easier to spin up a VM than it is to decommission it,” he says. “Most datacentres will have a significant proportion of unused VMs – in my experience, it’s around 30-40%.”

While only a fraction of storage and compute is consumed by these unused VMs on-premise, the VM is charged per second in the cloud, Witt points out. “You’re charged for the VM size regardless of whether it is fully utilised,” he says. “This is waste with an exact, measurable and recurring cost associated with it. As such, controlling sprawl in the cloud will yield immediate cost savings.”

Minimum spend limitations

Witt says some suppliers ask customers to commit to a minimum number of users, with a minimum growth expectation. Salesforce contracts are often structured this way.  

“So you buy up-front, and maybe it takes six months to deploy them all, but then your focus changes and you don’t use the product as much as you expect to, and you’re locked into a three-year commitment,” he says.

While the major software providers have been looking to migrate their customers to cloud subscriptions, Witt says it is worth checking exactly how much cloud is included with existing enterprise deals. For smaller software providers, there may be an overlap, and customers end up paying twice.

“The key takeaway is that the cloud is the strategic priority for all suppliers,” he says. “Use that to your advantage and get your procurement teams to structure the contract so that duplication is limited and you don’t end up paying for subscriptions you haven’t yet deployed.”

Read more about cloud cost optimisation

  • Without proper checks and controls in place, it is all too easy for the cost of using cloud to spiral out of control in enterprises – but it doesn’t have to.
  • Scalability is one of the cloud’s biggest advantages, but misuse of the feature can be costly. Pay special attention to these three rules to avoid unwanted billing surprises.

“On-premise enterprise agreements often have programmes to transition you to the cloud. For instance, a Microsoft Enterprise Agreement can certainly be structured that way,” adds Witt.

“You can often get a credit for the unused software assurance maintenance portion of your on-premise licences if you commit to cloud. And certain perpetual licences come with cloud deployment rights, if you have active maintenance [Windows Server and SQL Server, for example].”

Witt says he has also encountered situations where there has been a trade-in offer, with the customer receiving a discount on the first year of a subscription. “If you’re in the EU, the other option is to resell your old perpetual licences on the secondary software market,” he says.

Challenge of visibility

The nature of cloud services and SaaS means business users can buy subscriptions and consume cloud resources directly, without IT’s control, which means IT lacks a single view of use across the IT estate, says Witt.

“You are paying on a monthly basis,” he says, so how do you know if you’re still using a particular product or service? And if so, how many users are using it, and for how long?”

Although cloud providers have their own management portals to provide usage information for their own service, even if you know which suppliers you are using it can be difficult to track all this information manually each month, says Witt.

A recent Forrester report, Cloud cost monitoring and optimisation, Q2 2018, noted that while many organisations turn to the cloud to reduce their spend on infrastructure or to avoid steep up-front costs for new investments, this is hard to achieve.

“Success depends more on the maturity of cloud management and governance practices than the nature of the workload,” wrote the report’s author, Forrester principal analyst Lauren Nelson. “With cost complexity continuing to increase alongside growing usage, users, accounts and instance types, IT professionals increasingly depend on tools to enable visibility, consistency and scalability of management practices.”

Software asset management tools

Forrester has identified a category of software tools, including Apptio, CloudHealth Technologies, Densify, Rightscale and Turbonomic, which aim to provide software asset management-like capabilities for cloud resources.

According to Nelson, these tools’ central capability is to increase the visibility of cost and user dynamics across cloud platforms. In the report, she wrote: “As management practices mature, this information will become a critical element of governance and automation programmes, as both efforts rely heavily on consistent availability of data.”

What is clear from the experts Computer Weekly has spoken to is that the best practices of software asset management need to be applied to IaaS, PaaS, SaaS and other cloud resources, so that IT leaders can understand what cloud services are being consumed and for what purpose.

Witt believes departments procuring their own SaaS will not necessarily have the skills to manage such products and services, and this is where IT has a chance to take control. “For IT departments to re-engage with the business, they need to provide a service the business requires,” he says. “This requires a step out of the traditional role of an IT department into budgeting, compliance and supplier management.”

Read more on Software licensing

CIO
Security
Networking
Data Center
Data Management
Close