Thorsten Punke, global market manager at Huber+Suhner AG, discusses the evolution of data centre switching technology in an increasingly automated world…
The Internet of Things (IoT) is one of the most common buzzwords in today’s rapidly-evolving technological world. As a consequence, overall data centre capacity requirements are increasing and the type of facility needed to cope with this rise in data will become more and more segmented as business models and specific requirements diverge from one another.
The enterprise and the colocation segments within the data centre market are striking examples of this trend. Not only do they have a different purpose, they also require a specific technical approach to meet different needs. For instance, social media companies and CRM suppliers have specific requirements and usually manage their own data centres. Some larger companies have even developed their own hardware that supports their special requirements, while others are in the process of changing standard applications into fully-based IP data centres.
As a general rule, one requirement is important for all of them – uptime. Uptime has significantly increased over the years as IT has become more important thanks to many business processes being made IT-based. To support this, the data centre has many areas with redundant systems and organisational procedures.
Any human wrongdoing is critical and can lead to possible downtime
IT cabling infrastructure has had to undergo a similar evolution and optimisation process, which was vital in reducing link loss and improving overall link performance and efficiency with the rise in data rates. Developing cabling solutions that are user-friendly, fast to install and future-proofed is thus a pre-requisite.
Another important requirement that has evolved over the years is automation, rising from a market pull on the one hand and technical push on the other. At the top, efficiency and low power usage effectiveness (PUE) are key in maintaining control and guaranteeing the continuous functioning of all critical parts.
With this automation trend however, the only area that is still manual or ‘100% human’ is the patching itself. With the growth of bandwidth and data rates, the human factor and its impact on the overall wellbeing of the data centre is becoming more and more apparent. Any human wrongdoing is critical and can lead to possible downtime as a direct result of wrong patching.
A technology to overcome this issue is all-optical switching, where ports are connected via a light beam. The device itself is SDN-enabled and allows a new way of patching on the physical layer in no time without human intervention. The main function of all-optical switching is the ability to patch any incoming port to any outgoing one in the same switch, something that 20 years ago was deemed optimistic at best.
Combining this simultaneous switching functionality on the physical layer and making it remote-controlled with control features such as protection switching and network measurement, means that all ingredients to use this model in the data centre space are present without disrupting the active service. On top of this, because all-optical switching works on the physical layer, it is independent of the type of application and data required in the strongly segmented data centre market.
New technologies are vital and offering future-proofed connectivity solutions that can not only cope with today’s challenges, but also of those to come, is critical. All-optical switching could potentially enable automated cross connections by replacing human patching on Layer 1 while significantly improving the reliability and uptime of mission-critical connections.
This post originated at Data Centre Management magazine, from the same publisher as The Stack. Click here to find out more about the UK’s most important industry publication for the data centre space.