Has the ‘Cult of Public Cloud’ put us on shaky ground?
Written by Barry Alston, Claritas Solutions
Cloud services are increasingly viewed as commodities, akin to utilities like electricity or water. They are essential for modern business operations, and their availability and performance directly impact organisational success. The recent change from the government re-classing them in the critical national infrastructure showcasing just how important they now are alongside the emergency services and utilities. Perhaps only the scale of this shift coming to light following the recent fallout out from Microsoft outages and the Crowdstrike update deployment. Investigating the cost of the disruption caused by the incident, cloud outage analytics specialist Parametrix Insurance estimated the outage caused direct losses of $5.4 billion to US Fortune 500 companies. These days cyber insurance policies will likely cover no more than a sliver of the losses – perhaps 10% or 20%, if anything – “due to many companies’ large risk retentions, and to low policy limits relative to the potential outage loss.”
Despite this critical role, cloud services lack a regulatory framework similar to utilities, raising concerns about potential monopolistic control and the need for future regulatory oversight. A very good guide to whether you are a commodity or not is that you are answerable to a free market. This means that the production from each of the different suppliers is similar enough that we don’t have to make distinctions between them on quality or service grounds. This must be so or we can’t really create that futures contract. And if we don’t have to distinguish between them on those grounds then the only one left is to distinguish on price grounds, bringing in the ‘cloud’ benefit argument of capital expenses versus operational expenses.
The pay-per-use model of public cloud have been instrumental in adjusting business purchasing processes with large up-front capital expenditure costs for infrastructure being a thing of the past and a scale up use models having been around for many years now from all the major vendors. Using known minimal commit levels from customers ensure that the cost is known from the start – unlike cloud, where unknowns or wrong purchased service types can end up with customers having crippling bills. Plus, the financial arguments about moving cost from capital expenses versus operational expenses is also misnomer as it’s still a business cost no matter which side of the books the payment is made.
For some time, our own Government’s ‘Cloud First’ decision, has seen over the past decade departments adopt Public HyperCloud solutions over alternative delivery platforms, with the two big hyperscalers servicing around 70% of the UK’s cloud business today. Take the headlines recently seen in the news where the UK has significantly expanded its data infrastructure, with four major US firms, CyrusOne, ServiceNow, CloudHQ, and CoreWeave, committing a combined £6.3 billion ($8.2 billion) in investments. Adding to the £8 billion AWS announced back in September to building, operating and maintaining data centres in the UK.
A single vendor strategy or ‘having all your eggs in one basket’ which involves relying on one provider for all your technological needs, from software and hardware to cloud services. While this approach can streamline operations and simplify vendor management, it comes with notable drawbacks like vendor lock-in, reduced negotiation power, and potential single points of failure with no liability on the provider for outage of service. Many vendor contracts have limited liability clauses, which means the vendor isn’t legally responsible for certain types of service failures, including outages. You, as the customer, bear the risk of the vendor’s performance issues, and there’s little recourse if something goes wrong, so the replication and backup are on you. In most vendor agreements, you have to ensure that your data is backed up or replicated across your multiple locations or environments. The vendor may provide tools, but the configuration, execution, and verification of backups are your responsibility.
I work for Claritas Solutions, a Wetherby-based IT company and we see businesses sleep-walking into cloud-based services they assume are cost effective, safe and impenetrable as well as having managed backup and resiliency built-in which is not the case. Many business leaders probably don’t realise the significance of where their data is stored and processed and this is and will continue to be the lifeblood of Gen AI. At Claritas Solutions, we have UK-based data centres and none of our data or traffic is routed overseas with security cleared (SC) staff with decades of industry experience with secure managed backup solutions to ensure business resiliency.
The UK developed a national cloud strategy in 2009/10 – parked to ‘allow the market to decide’, coincidentally just as Microsoft launched their own Global Cloud platform; but the UK’s work formed much of the thinking for the US FedRAMP model the following year. It’s definitely now time to dust that off, look at it again in the light of today’s more mature technology landscape and to create something suitable for the UK.
Most customers still rely on older systems with workloads that remain stable but gradually grow over time, and these systems aren’t designed for the cloud. Only projects that require flexible scaling tend to expand and shrink as needed. Nowadays, you can reserve 20-30% extra capacity, allowing for workload adjustments as needed, or for future growth, similar to how cloud services offer pay-as-you-go flexibility without needing a large upfront investment. While some customers have unique needs that are suited for public cloud, most businesses have a mix of requirements and should choose the right solution for each task.
If the UK’s vision about becoming a world tech leader is still on the table – you can’t be a world leader if you allow your suppliers to make all your key technology decisions for you, it will remain just that.