Roundtable: Are all colocation data centres created equal?

07 July 2023

With the colocation data centre market booming, how can enterprises distinguish between the ‘top notch’ and the ‘just ok’?

What should a user expect from a top-quality colocation data centre (DC)?

Matt Seaton, director, Netwise: A DC is all about providing a high uptime environment to critical IT services, so a natural focus should always be on the resilience of the core infrastructure like power, cooling, and connectivity. Any top-quality facility operator will also give focus to supplementary services, such as support and on-site amenities.

John Hall, managing director-colocation, Proximity Data Centres: The DC should have state-of-the-art infrastructure with redundant power systems, backup generators, multiple network carriers for diverse connectivity, advanced cooling systems, remote hands services, and physical security measures like biometric access controls and surveillance systems to protect infrastructure and data. It is important that this is a truly diverse infrastructure as regular maintenance is vital to ensure high availability: the facility should provide a highly reliable and redundant network infrastructure to ensure maximum uptime. This includes redundant power feeds, backup systems, and multiple network connections to minimise the risk of downtime.

Steve O’Brien, managing director, cloud, Redcentric: It’s important to look at the features that provide added value. Sustainability is a critical issue, so features like immersion cooling can help to provide higher levels of energy efficiency. Robust backup and resilience plans should be expected from a top-quality colocation DC; even the largest hyperscalers have faced issues.

Charlie Stace, solutions architect, iomart: A user should expect robust, reliable, and redundant infrastructure, including, power, cooling, fire suppression and physical security. It’s important that a DC has redundancy in each area to ensure that in the event of failure the DC continues to run smoothly. Features such as 24/7 physical security, monitoring, biometric access controls, and video surveillance are necessary. The DC should have a multiple carrier neutral network provider available to ensure clients have high-speed low latency connections. SLAs should cover response times for technical support, guarantees on power and network. These should also include schedules for maintenance. Compliance and certifications are also an important consideration.

What makes one colocation DC better than the next?

John Hall: As well as features such as the M&E infrastructure and connectivity, the most important factor is the team that is responsible for the day-to-day management. Do they have the experience and expertise as well as the ‘can do’ attitude to really understand client needs? Geographic location is also important as your team will probably need to visit the site.

Matt Seaton: Given the well-established tiering system for quickly identifying a DC’s redundancy level, it should be simple to categorise groups of facilities together, which should all be of a comparable technical standard. It’s everything that sits beyond this, at each level of operational standard, that sets one facility apart from the next. Do they offer 24x7 support and access? Is the facility well-maintained? Are they able to demonstrate the underlying technology? Do they have on-site amenities? Is the facility well-located? Are there dry loading capabilities? Do they have a delivery handling function?

Charlie Stace: Uptime and reliability are critical. Look for DCs with a proven history of minimal downtime, investigate how power is made resilient, and how long the facility can run with loss of power from the grid. The diversity and quality of the network connectivity also plays a significant role; look for DCs with Tier 1 carriers and ISPs. DCs with good access to major internet exchanges and content delivery networks offer improved performance and efficient data transfer. DCs should also be able to recover from disaster easily; look out for backup power systems (multiple generators), UPS, and diverse physical routes into the DC for both power and connectivity.

Steve O’Brien: This varies because of the differing needs of organisations and their specific digital transformation goals. If data sovereignty is the key priority, for example, then that will be the differentiating factor between one DC and the next.

What should potential users look to avoid?

Steve O’Brien: An absence of physical security safeguards and cybersecurity measures could indicate that a supplier doesn’t prioritise security, which is critical for protecting sovereign data.

John Hall: Look at the team and see how they understand your requirements. Have they taken the time and effort to understand your needs and provide a solution to satisfy them 100%? Have they given good advice and discussed your current and future requirements? If not, then walk away.

Charlie Stace: You want to see as much resilience, and reliability as possible in most cases, however there could be some cost advantages in finding lower grade DCs if your requirements are not as critical.

Matt Seaton: Any visual clues during a tour that a site may not be well looked after are a red flag - a top-quality DC should always be extremely well presented. Beyond this, one of the biggest things to look at will be the financial stability of the company.

How big an impact does pricing have on quality?

Charlie Stace: It’s important to remember that colocation is in a competitive market - pricing does not always have a significant impact on quality or service. That said, it’s not unusual to find the top Tier DCs charging more.

Matt Seaton: As with most things, you do generally get what you pay for, and while some operators do unfairly leverage this based on their brand equity, for the most part, the very best facility operators will be positioned further up the pricing table. Designing, building, and running world-class DCs doesn’t come cheap, and while the best DCs in the world do offer exceptional value, you can expect to pay a little more for that peace of mind. The race to the bottom on pricing is simply not conducive to operating a top-quality DC in the long-term, so if you see pricing that’s too good to be true, it probably is.

John Hall: There are four broad areas of costs associated with DC services – setup costs for the initial installation; rental of the area used (rack footprint); power consumed; and connectivity. Power costs have been much discussed in the press and are important, but the cost of the rack footprint rental can vary considerably across the UK.

Steve O’Brien: Due to the current high costs of energy, many DCs have increased prices and mitigated passing on costs to customers to stay in business. Location matters and DCs located in high areas of energy demand will have to pass on those costs, making their services less value for money. It’s worth examining how a given DC incurs costs and as a result, the impact that it has on quality.