Businesses migrating to the cloud are understandably enthusiastic about realising all the benefits of scalability, flexibility and economy that the best cloud offerings can deliver. However, many encounter a stumbling block when confronted by the complex array of pricing models offered by different providers. Making sense of all the different options can be a major headache along the path to cloud adoption.
Back in 2013, a report from analyst firm 451 found that cloud pricing was excessively complicated, and blamed the immaturity of the market for the lack of clarity and standardisation of offerings and pricing. This was making it very difficult for customers to compare different vendors and indeed sparked the creation of several cloud comparison indices in a bid to resolve the problem. Four years later, are we doing any better?
Businesses migrating to, or starting out in, the cloud will find themselves courted by a beauty parade of providers, all claiming to provide just the right solution for their needs. The challenge, however, is that all organisations are different and have varying requirements at different times. It’s the drive to deliver this flexibility – which, after all, is what cloud computing is all about – that has resulted in the complex pricing environment that we continue to see.
Attempts to make cloud purchasing more of a transparent off-the-peg activity have resulted in a massive range of alternatives and options aimed at persuading purchasers to buy a package. Take for example, the hyper scale public cloud providers such as AWS and Azure. These companies use the idea of instance or t-shirt sizes for their virtual machines, with the customer selecting the “size” that they feel best fits their requirements. There is a vast range of alternatives depending on the CPU type, number of CPUs, RAM, the number of virtual disks that can be attached, whether they have GPU acceleration and myriad other elements.
Typically, these are priced on a pay as you go basis, where you are billed by the hour (or minute or more recently by the second) for the particular t-shirt/instance size. Persistent storage is normally billed separately, unless you are using ephemeral VMs, where the storage associated with the VM is not persistent, and effectively disappears when the VM is turned off. In some cases, even with persistent storage, a virtual disk is often shown in a VM which is ephemeral and is used for caching or temporary space while the machine is running.
As well as the instance-based pricing models, other models have recently launched:
Reserved pricing: under this model you agree a contract that allows you to use the VMs 24/7 for a one or three-year period. Large discounts are available but, rather like that special offer gym membership that you take out in January, you are paying for the VM regardless of whether you use it or not.
Convertible reserved pricing: this is effectively the same as above, but allows you to recoup outlay by selling time slots for your reserved instance that you don’t need on a marketplace.
Spot pricing: in this model you bid for the maximum price you are prepared to pay for a VM. Depending on the provider’s spare capacity, your VM will be powered on when your spot price is reached, and powered off again when your spot price is not high enough. This might be useful for batch processing requirements when scheduled short term bursts of capacity are required.
Away from these complicated pricing models offered by hyper scale providers, other cloud providers are working hard to make life easier for customers by tailoring solutions based on the characteristics and maturity of the customer’s business and with more straightforward pricing. These basically come down to three alternatives that take a lot of the pain out of identifying the right approach:
The Pay As You Go model is based on the actual usage of CPU, RAM and storage by the minute and charged monthly. The benefit here is its bespoke nature: rather than having to estimate which “size” they need ahead of time, customers can provision any size VM (within reasonable limits) and change the specifications whenever they need to. This model suits highly dynamic, growing and evolving businesses who value the ability to scale and shrink at a moment’s notice.
The second option, referred to as ‘Reservation Pool, is generally lower cost, though slightly less flexible, and offers customers a reserved “bucket of resources” which have guaranteed availability. This option works well for mature businesses that have steady demand requirements and really want to see the cost-efficiencies that the cloud can deliver. Even greater discounts are available if businesses commit to multi-year deals.
Organisations that sit in the middle of that spectrum can opt for the third way – a mixture of the two approaches – reserving a bucket of resources (that can be easily increased if required), but taking an option to burst into PAYG resources when the occasion demands it. These three options offer a very cost-effective alternative to the off-the-shelf approach and are arguably more transparent for customers weighing up the various alternatives.
Of course, it’s worth remembering that, despite the apparent “race to the bottom” taking place between the hyper cloud providers, it’s not all about the money. While some services are becoming standardised and commoditised, if you are looking for a specific offering – DRaaS, for example – you will want to know far more about it than simply that it’s the cheapest on the market. That’s why I firmly believe that consultation has a critical role to play when businesses are choosing cloud solutions.
The pricing environment remains complex, despite our best efforts, and there is more to vendor selection than price alone. In any case, enterprises moving to the cloud need to be viewing it as a strategic, not just operational, decision and as such they need to take advantage of the advice that cloud providers offer. This will ensure that they don’t just get the best price, but they also get the best service that meets the real and changing needs of their organisation.