Redundancy is a key concept in the data center industry. It means having spare components and backup paths in place to ensure business continuity in the event of a failure. Every colocation provider claims to offer redundant infrastructure. In practice, however, the difference lies in the details, and the term redundancy is not always interpreted in the same way.

In a professional data center, the standard is to provide the customer with two independent power paths, conventionally referred to as A and B. Modern servers and network devices are equipped with two power supplies, each connected to a different path. In theory, if one path fails, the device continues operating without interruption by drawing power from the other.

The hidden compromise – “cost-saving” redundancy

There is a practice on the market that represents a dangerous compromise between security and cost cutting. It involves installing UPS systems on only one of the two power paths.

Path A – protected – utility power passes through the UPS system, which filters it and guarantees uninterrupted power in the event of a grid failure.
Path B – unprotected – utility power is fed directly to the power distribution units in the rack, bypassing the UPS altogether.

A provider using this setup can technically claim to deliver two power paths. The problem is that this approach creates a number of serious risks: no protection against disturbances on path B, increased exposure for devices with a single power supply, unpredictable consequences in the event of a failure, and a false sense of security for the customer.

Talex’s uncompromising approach

At Talex Data Center, we have followed a philosophy of uncompromising quality from the very beginning. Both of our power paths, A and B, are fully independent and fully protected. This means that each path has its own utility connection, separate main switchboard, dedicated redundant UPS system, independent generator backup, and a separate cable route leading to the rack cabinets.

A failure of any of these elements on one path has absolutely no impact on the operation of the other. This is real, full redundancy – one that provides our customers with genuine, not illusory, peace of mind. It is also one of the fundamental requirements of the EN 50600 standard, which we meet at the highest, fourth availability level.

Ask for the details

When choosing a colocation partner, do not hesitate to ask specific questions. Ask directly: are both power paths protected by UPS systems? What does the power architecture look like from the utility connection all the way to the rack? Do you hold a certificate confirming the availability class of your data center?

The answers to these questions will tell you far more about a provider’s approach to quality and risk than any marketing brochure ever could.