At the beginning of the cloud computing revolution the idea of just paying for what you use promised to usher in a more efficient age of IT spending. Although that has certainly been the case for some, and has allowed web scale companies to grow extremely quickly, there is always the risk of cloud bloat, where a company loses track of its cloud spending across an increasingly disparate range of suppliers and vendors.
Shifting to the public cloud might not necessarily cost your organisation more or lead to serious cost savings, but it will change the way your organisation procures and pays for technology, making costs less predictable than the old long-term licensing models.
For example, music streaming company Spotify recently completed a full migration to the Google Cloud Platform (GCP) from on-premise data centres.
When asked during the Google Cloud Next conference in July 2018 what the cost implications of this move would be, director of engineering Ramon van Alteren said: “This is a key thing to keep an eye on as we move from a centralised buying position to a distributed buying position where everyone is capable of spending money for your company. So it depends. Currently we have grown in size so it is really hard to compare and I can’t give you figures.”
The RightScale 2018 State of the Cloud report found that 81 percent of enterprises have a multi-cloud strategy, with respondents estimating that they are wasting 30 percent of spend a year. RightScale itself puts this figure nearer to 35 percent.
As a result optimising cloud costs is the top initiative for respondents to the 2018 survey, with 58 percent listing it as their top cloud priority. Despite this, the research found that only a minority of respondents had already implemented automated policies to optimise cloud costs, such as shutting down unused workloads or selecting lower-cost clouds or regions.
This changing paradigm has had a direct effect on the legacy vendors too, which are having to wrestle with their pricing policies in a world where they need to adapt, while keeping their core licensed customers happy too.
German legacy vendor SAP has been publicly dealing with this issue for years now, and revealed its modernised pricing policy in May 2017.
Hala Zeine, SAP corporate development officer wrote at the time: “In an agile world where digital reigns supreme, licensing complexity is getting in the way of innovation… Our objective was to make pricing predictable, linked to unit of value, transparent, and consistently applied.
“Does this address every indirect access scenario in the age of devices, IoT, and collaborative networks? Not yet. There is much more to do and we are eager to keep updating pricing scenarios to bring you greater value. It is, however, a step in the right direct toward pricing modernisation.”
So, how can companies ensure they aren’t overspending on cloud?
The first step on the road to avoiding cloud overspending is to be prepared and diligent when procuring new services.
As cloud vendor Nutanix advises via its in-house Next magazine: “Before you move forward with a cloud vendor, make sure you understand their pricing model. It’s critical that you understand all the “hidden” costs for various API calls and other transactions so you can make a fair comparison and choose the best vendor for your use cases.
“Once you start using the cloud, identify those services that are the highest contributors to your monthly spend. Rationalise usage to either justify costs or identify opportunities for optimisation. Over-utilisation of a service could be due to a coding error in an application.
“Also, since cloud services are available at the click of a mouse, some services may have been provisioned and forgotten about. These orphaned or unused resources can waste thousands of dollars. Identifying them can deliver a quick cost-savings win.”
Nutanix also recommends as best practice: “Assign workloads to the business unit or other functional area that is responsible for them. This will not only enable you to chargeback for services, but can help with more accurate forecasting. Budgets can be created based on the business knowledge and requirements of the teams that are using the services. Owners can be made accountable for ensuring that resources remain optimised and within budget.”
Colin Rowland, SVP of EMEA at Apptio advises: “Before adopting any new IaaS and embarking on a public cloud migration, IT leaders first need to conduct a robust analysis of the total cost of cloud migration, trading off any expected savings against the costs of doing the migration itself.”
Be familiar with your estate
Most effective cloud management comes back at some point to effective monitoring. You can’t effectively manage your spend if you don’t know what your organisation is spending on.
“Take a look at all the different services you’re consuming in the cloud – instances, load balancing, SQL and NoSQL services, etc. – and become familiar with the advantages and costs of these services,” advises Mor Cohen, Cloud CTO at Turbonomic.
“If your organisation has predictable usage patterns and application demand stays fairly consistent, you can calculate these costs pretty easily. Then you can comparison-shop from one provider to another.”
The issue here is actually keeping track of all of these services. If you can’t see your whole cloud environment it is impossible to rationalise.
Fortunately there are a number of third party and open source tools for monitoring your cloud environments, which will help you keep track of spending.
Read next: Best cloud management tools 2018
Apptio is one of the bigger names in the technology business management (TBM) space – a nascent technology area coined by the vendor itself. Apptio also helped found the TBM Council back in 2014 to support its methodology and its growing community of IT leaders.
The idea behind tools like Apptio and the techniques of TBM hinge on understanding increasingly complex cost bases, be that cloud infrastructure and services or internal labour and support. Once that visibility has been achieved, IT leaders can, in theory, start to spot pinch points and inefficiencies to reallocate funds.
As Colin Rowland says: “It’s never been easier for a business to sign up to a new software, making it increasingly difficult for the IT department to keep tabs on the spend of other business units.
“In this shadow IT environment, vendor financial data, contracts and vendor performance details are owned by disparate business units, limiting IT’s ability to manage the vendor portfolio end-to-end. This lack of a centralised view leads to greater risk and inefficiencies: redundant or overlapping vendors, unexpected overruns in spend, surprise contract renewals and vendor decisions that are not data driven.
“In order to avoid vendor sprawl and keep SaaS costs to a minimum, the CIO needs a single view of all vendor spend and contract details.”
It is proving an attractive approach in the financial services sector. UK banks Nationwide and RBS have already used Apptio to rationalise their IT spend, with Nationwide expecting to save £6-8 million in IT costs after its first year of implementation, and now Lloyds Banking Group is looking to do the same.
Keith Pearson, IT CIO at Lloyds Banking Group turned to Apptio in an attempt to get visibility of all IT assets, be that a business-critical software application, or a piece of hardware like an on-premise data centre, and be able to see: “Who owns it, what attributes it has, is it secure, is it resilient, how much does it cost, will we be fined?”These are fundamental questions we couldn’t answer about an asset,” he told Computerworld UK.
For example, Lloyds uses the Falcon fraud detection platform sold by FICO. Pearson says that the company thought it knew what it was spending on Falcon, but after implementing Apptio, saw that it was spending five times more than the bank suspected.
Read next: Best cloud management tools 2018
Cloud identity management tools like Okta can also prove useful when it comes to visibility of SaaS applications.
Rob Sansom, the global technology advisor at Capita spoke at Okta Forum in London in July 2018, saying that the tool had helped give the outsourcing giant new visibility into how these services are being used across the business.
“With acquiring so many businesses it is difficult to have a holistic view of all of the SaaS applications being consumed,” he said. “We know for example that we have a high number of duplicate tenants of particular services, Salesforce being a particular example.”
“We had no real picture of who was using a service,” he added. “Now we can evidence that and have more opportunities for efficient and effective management of software licences.”
By using Okta, the IT team can start to see how the applications it has integrated are being used.
“That allows us to identify opportunities for consolidation of those tenants and also save money on licences,” Sansom said.
Avoiding cloud sprawl
Cloud adoption brings with it an inherent risk of sprawl. As hinted at by Spotify earlier, this is where a lack of central coordination over cloud purchasing and the over purchasing of resources can lead to sprawling, and spiralling, costs.
The answer to this issue is the somewhat dry solution of effective cloud resource management. As cloud computing is effectively an operating expenditure (OpEx) instead of a capital expenditure (CapEx), IT teams can often purchase new cloud resources without requiring central sign off. This is great for agility and innovation but poses a risk to budgets.
A clear and simple cloud policy, which sets baselines for resource allocation to support an application, should help avoid any abuse in this area however.
If your organisation is running at serious scale a central cloud monitoring team should also be put in place to ensure policies aren’t being violated and that spend it being optimally allocated. This specialist team could also ensure that teams aren’t over-provisioning resources for their workloads, purchasing a dedicated cloud server when that isn’t required, for example.
Mor Cohen from Turbonomic says that “to truly realise the potential cost savings of the cloud, you have to continuously optimise, delete idle resource and right size across all resources.
“If you do this, it can generate incredible savings. It all sounds simple until you realise that there are 1.7 million configuration options for EC2 instances alone, 90 additional services available in AWS, and similar numbers in Azure. On top of that, in an ever-changing cost landscape where 70% of the these options change within the course of a year, it is no surprise that most organisations consume public cloud almost statically.
“Simply put, the complexities of the cloud are beyond human scale and the only reliable way to balance price, performance and compliance is to automate.”
Another potential route here is to shift towards a modern serverless computing model, where code execution is fully managed by a cloud provider, instead of the traditional method of managing, provisioning and maintaining servers when deploying code.
The main cost saving here comes in the form of developer time as there is no need to provision, configure and manage infrastructure, and in increased utilisation as users are only charged for the time they are actively using the platform.
Avoiding cloud jail
Cloud management specialist Turbonomic also talks about avoiding ‘cloud jail’, which is a fancy way of saying vendor lock-in when talking specifically about infrastructure-as-a-service (IaaS) providers.
As outlined in a blog post from product lead Ben Yemeni: “Cloud jail is simply waking up to discover you’re spending way too much money on infrastructure and are completely beholden to one provider. Once this happens it becomes very hard to switch. Inmates have often architected for one cloud. It becomes very difficult to migrate to another (the network tunnels cost money) and re-architecting to operate across multiple providers is a significant undertaking.”
To stay out of cloud jail Turbonomic recommends three actions: cost optimisation, having a great app experience and being multi-cloud from day one. We’ve already covered cost optimisation but the other two are worth unpacking.
As Yemini writes: “If your cloud journey is all about that next great app and user experience is paramount to delivering the best app you can. Make sure you are set up to tracking the end-user response time and performance of your app from day one and be particularly leery of network performance issues your provider can’t or won’t fix (e.g. packet loss for certain geographies or internet providers). If you can’t solve issues by leveraging CDNs or SD-WAN acceleration you may want to think twice about your chosen provider.”
Then when it comes to being multi-cloud from the start, Yemini advises to “set up a test environment that has cross-cloud connectivity. This could be as simple as vCenter and AWS or more complex with multiple private and public connections,” he writes.
Don’t bet the farm
It has become something of a given that moving to the cloud will make everything better, but some workloads are simply better left alone.
“In some cases, it doesn’t make sense for an organisation to move everything to the cloud and a hybrid approach is best. Build a data-driven cloud strategy based on your applications, your data and your requirements,” says Mor Cohen from Turbonomic.
“When you decide which workloads are best-suited for the cloud, don’t simply lift and shift everything without understanding how to optimise it. That’s like de-cluttering your garage, but instead of throwing out the old and unused items, writing a check every month to keep them in a storage unit across town.
“One of our customers did this and migrated 400 VMs to the cloud. As a result they ended up paying about $4 million per year for those applications – 6X what the cost of the equivalent applications was on prem,” she said.
Shift to archival storage where possible
One quick win when it comes to reducing cloud storage costs is to shift older data to cheaper storage tiers. AWS offers a service called Glacier for just this at just $0.0045 per GB per month. Google Cloud’s Coldline is a viable alternative to AWS too, at $0.7 per GB per month for data at rest.
All in all the answer to effective cloud management is not a simple one. It requires astute management, solid policies, the right tooling and perhaps even a dedicated team, but as the world moves more and more towards the cloud best practices will emerge.