Data Centers with Fewer Outages

Although data center failures have dropped in recent years, this should not decrease the vigilance of operators and their customers. The Uptime Institute, which specializes in certification of  data centers, has studied the global market from this angle. Numbers of failures are indeed dropping, but when they occur, they are more costly. There are also facilities which have not yet experienced a single failure.

Fewer problems in the ‘serious’ category

According to the Uptime Institute report, in 2020, 78% of surveyed data centers faced outages. In 2021, it was 69%, and in 2022 only 60%. The conclusion? A large number of facilities are experiencing outages all of the time, but a relatively small proportion of disruptions are treated as  ‘major’. Critical disruptions have historically been estimated at around 20% of all failures; however, this percentage dropped significantly to 14% last year.

Is there anything to worry about?

The results of recent surveys mean that problems in an off-premise, professionally managed data center are less likely to occur. Still, if they do happen, they can be costly and difficult to control. This is bad news because a failure of data center infrastructure can cause severe consequences for organizations that rely on such services, e.g., hosting companies, e-commerce, manufacturing, logistics, finance or SaaS sectors. Failures often lead to measurable financial and reputational losses.

The failure of AWS, one of the world’s largest public cloud providers, which took place in December 2021 and lasted several hours, affected the operations of many global companies such as Netflix, Disney, and Spotify. A fire at OVH’s headquarters in France two years ago caused equally serious problems. Polish customers could not use the services of major e-shops or email providers. Critical services such as those provided by  TOPR  informing mountain hikers about the weather and risk of avalanches stopped working.

Similar problems also affected companies in Poland. For example, in June this year, 2 commercial services of the country’s largest hosting provider were unavailable due to a failure in the data center maintaining the company’s physical infrastructure.

What if there are problems? 

Representatives of the data center market profess to have many safeguards and practices in place to increase security and ensure access to data even in a crisis. Still, even they point out that the devil is in the details.

– It is in the interest of the data center operator to ensure the highest possible level of security and business continuity for customers. However, firstly, the standards offered by data centers vary widely, and secondly, it is not always possible to assess them objectively as only some facilities undergo independent certification. It is worth noting, however, that a failure does not always have to mean permanent unavailability of data – it can be costly to remedy. Still, if procedures and technologies allow it, the customer can regain access to data and services. Another thing is that a lot depends on what level of service and SLA the customer decides on; not everyone makes a conscious decision to ensure continuity of their services, especially if the decision driven by cost, notes Wojciech Stramski, CEO of Beyond.pl, a provider of data center, cloud, and Managed Services

Security is key!

Technology experts primarily draw attention to the redundancy of the physical infrastructure in a data center, which means duplication of a number of solutions – from independent sources of primary power, backup power (their redundancy) or implemented fire protection technologies.

This is why contractual provisions defining guaranteed service availability, or SLA (Service Level Agreement), are so important.

– Service level provisions are the key element of any contract with a technology provider. They should contain not only a precise definition of the level of services provided, together with a precise definition of their possible unavailability but also the formulation of penalties in the event of a failure to achieve the agreed targets, emphasizes Łukasz Węgrzyn, Partner and Head of IT and Data at Osborne Clarke law firm. How does an SLA work? With an SLA of 99.5%, the data center service provider undertakes that downtime in the facility’s operation during a month cannot be longer than 3 hours and 36 minutes. The lower the SLA level, the higher the potential risk of service unavailability to the customer.

Certified security

Companies should verify the level of security of the data center by verifying its certificates and standards under which it operates, the granting of which is preceded by a thorough verification of the data center. Ideally, these certificates would be awarded by independent institutions and require periodic recertification. The most advanced of audits can contain up to thousands of requirements that require verification. For example, for a data center to boast an ANSI/TIA-942 Rated 4 standard, it must meet more than 2,600 requirements in architecture, mechanics, power supply, and telecommunications. As a result, such a facility can be described as fault-tolerant. However, there are only a few such facilities in Europe and only one in Poland.

– Security in a data center facility should be considered as early as the investment planning stage. Many issues, such as the construction, the separation of zones in the server rooms, the implemented energy or cooling solutions, communication redundancy, fire protection or the physical location itself as well as the operational procedures associated with the day-to-day running of the business, are dictated by specific and very high requirements. A certificate confirming that a facility is among the safest is the culmination of this process. Representatives of many industries such as banking, logistics, medical, manufacturing, retail, and e-commerce, need uninterrupted business continuity. We are noticing an increasing awareness of this fact from customers. Anyone who has experienced a failure and its consequences – loss of profits or customers and their trust – does not question the importance of ensuring business continuity, concludes Wojciech Stramski.

Many years of industry research show that professional data center facilities are doing increasingly well when it comes to ensuring fault tolerance and business continuity. In addition, there is increasing knowledge and customer awareness in this space. When choosing a data center operator, companies are no longer guided by the price alone but focus on more important elements such as the selection of a reliable and proven partner that will provide them with peace of mind, allowing them to focus on developing their core business.