Here are the new rules for data centre consolidation
The mass-scale pandemic disruption to enterprise IT equipment refresh may not be on your radar, unless, of course, you are operating a data centre. First there were the lockdowns to contend with and, depending on whether your industry emerged as a Covid ’winner’ or ’loser’, straitened budgets to wrangle.
And now? Supply constraints, including delays to the latest generation of Intel’s data centre chips, mean that companies are experiencing component and hardware shortage, even where budgets and staff are not a problem. The cumulative effect is that customers can be quoted lead times of anything from three to nine months, Robin Gardner, Strategic Services Director of Xtravirt, the next-gen managed services and cloud consultancy firm, tells us.
With Intel CEO Pat Gelsinger recently predicting shortages will continue until 2023, there is little prospect of lead times contracting any time soon. The upshot is that many companies will be forced to sweat their existing equipment for five or six years rather than the three to four they may have previously anticipated.
In most scenarios, delaying IT equipment refresh is a false economy, which has a knock on effect on efficiency and performance. Old kit also requires more patching – and this in turn could compromise the organisation’s security posture, compliance, risk management and governance.
There’s no question that the tech infrastructure of many organisations’ is long overdue for a refresh, Gardner argues. Couple this “trigger” – as he calls it – with the business imperative to propel digital transformation and accommodate working from home and you have the perfect opportunity to revisit data centre consolidation.
“People and organisations are looking at how they consolidate their offices and their real estate in general,” Gardner explains. But where people can work from home, or transfer to new digs, servers and storage arrays are far less moveable.
This means the traditional data centre is effectively a millstone around the facility team’s neck. “It locks them to that space, and that investment. Moving to another internal location or internal facility just moves the millstone.”
Here are the new rules for data centre consolidation
Almost as long as we’ve had data centres, tech managers have been on a mission to consolidate them. Each new wave of investment in cutting edge servers, storage and comms infrastructure quickly gives way to a wave of angst over efficiency, performance, and resiliency.
It’s an ongoing cycle, but for the last decade or so, the received wisdom has been the best way to break it is to bolt to the cloud. However, this could be a tall order, considering that, according to McKinsey estimates, some 80 per cent of’ enterprise virtualized workloads reside in the data centre. Cloud migration is complicated enough, and it is important to to avoid conflating this with appmigration to cloud native, Gardner advises.
App migration to cloud native is a difficult process, requiring planning and investment, and this means that priority is given to the apps with a stronger business case, or more easily modernised. However this leaves a long tale of traditional workloads in a legacy data centre, he explains. Due to the need to remove dependencies from real estate, and difficulties and cost of hardware lifecycle management, “getting these workloads to the cloud has a new urgency, and the route cannot be via modernisation as it will take too long.”
“The triggers for data centre consolidation are actually nothing to do with changing an organisation’s software infrastructure,” Gardner says. It’s looking at the cloud migration strategy not being driven by a cloud native objective and vision. It’s being driven by the opportunities presented and the risk mitigation presented by getting out of a physical data centre.”
Also bubbling away in the background is an increasing pressure to become carbon neutral, according to Gardner. “Among the many enterprises we’ve worked with over the last 12 months, being carbon neutral has not been one of their top three concerns, yet based on our experiences, sustainability is one of the fastest growing areas of priority.”
This will become increasingly important from a regulatory point of view, and as a consumer and employee concern, he says, and there are obvious implications for anyone managing racks of servers which have to be powered and cooled and more broadly on how organisations manage their infrastructure operations.
Real estate, real problems
“One of the key things about becoming carbon neutral is knowing what your carbon footprint looks like, understanding it, tracking it, maintaining it, and introducing flexibility to be able to proactively manage it.” This is likely to be harder to achieve on-premises, where the range of resources, both in tooling and personnel, is significantly less than the public cloud giants can offer.
Consolidate all these triggers and you have a compelling case for companies to beat a retreat from some or all of their existing data centres.
Some organisations might look at consolidating their infrastructure to a single data centre. But as Gardner says, “when you’ve got a distributed data centre model, you’ve got inherent resilience, because you haven’t got all your eggs in one basket.”
A single, in-house data centre means investing in additional resilience, business continuity and bandwidth for the facility, as well as backup as a service. That is of course, if the company can secure the requisite kit.
The same problems apply if you’re looking at the halfway house of using a data centre or colocation provider. “You’re still constrained by server lead time.” And with both options, Gardner points out. “You’ve got an opportunity cost of taking resources from other projects.”
It’s also about recognising that enterprise workloads are still overwhelmingly VMware-based. In which case, VMware Cloud on a public cloud such as AWS provides “an easy answer that can be implemented quickly to address these trigger challenges.”
This means that organisations can continue to leverage the same skills and management tools they’ve been using on-prem, he explains: “One of the benefits of VMware Cloud on AWS is that from an IT operating model, and from an application change or modernization perspective, it’s relatively low risk and low overhead.”
One way of looking at this is that the longer you delay the decision on what to do with your data centres, the greater the accumulated technical debt.
Less space, more room to breathe
Although the ultimate shape of infrastructure might be up for debate over the longer term, the VMware Cloud approach offers the chance to address the hardware-based portion of that technical debt quickly. Leaders then have breathing space and bandwidth to address the software infrastructure side of the issue.
“VMware Cloud can be in place with migrated workloads within months,” Gardner says, with the initial enablement taking a week to ten days.
And organisations can free up more bandwidth by choosing the right partner. Working with a managed service provider (MSP), both on planning the migration and operations longer term, gives IT teams the time to get on with dealing with that array of other problems.
After all, says Gardner, just consider “the breadth of demand on IT, and on business right now. New business models, exploring and seeking out new sources, or replacements, of revenue. IT is under pressure from shedloads of directions”.
Then, on top of that, “there is new demand from facilities that you haven’t been accounting for before. And there’s a new risk associated with access to hardware that you hadn’t accounted for before.”
Embracing both a cloud provider and an MSP might be seen as introducing further complexity and costs. But this should be set against the much greater costs and complications of simply doing nothing.
“The business case has changed. The risk of not doing it is increasing, particularly if you’ve deferred lifecycle renewal through Covid,” argues Gardner, who says the right partner will enable its customers to remove risk, more quickly, more efficiently and deliver capacity and services more quickly.
As for the growing issue of carbon neutrality, a public cloud provider can provide scale and tooling that is unavailable to most in-house installations, and also the capability to flex in line with your business profile and business demand.
“It’s important to grasp this is not an all in model, rather it’s a ‘cloud right’ strategy. If you’ve got a distributed data centre model today, consolidating a third of them or two thirds of them to cloud and retaining the rest, that’s still part of the solution. It’s about identifying the new triggers, the new risk factors, and having an easy to- apply solution.”
Those new triggers account for the vast bulk of the problems and challenges facing organisations today, says Gardner, and “the bleeding edge solutions aren’t going to come quick enough to solve those.”
