One thing I have always tried to remember when people ask “Why?”, is to try to think of the end result of the Cable Plant you wish to have. Is this something you can be proud of, and something your CIO would like to show off to his colleagues? More importantly, is this something that is useable when you get that call at 2:00AM saying “‘X’ piece of equipment is down, please come in to take a look.” In a matter of minutes you know where everything is and can easily eliminate or target a physical layer problem. Those are the times you will thank yourself that you took the extra effort to ensure you have a sound professional cable management design.
As far as cable management goes there are many schools of thought around this subject. So while there is no one right answer, I will provide you with the ammunition you need to deploy an efficient cable plant which will go a long way to supporting your environment well into the future.
The thing I find interesting is that almost none of the storage vendors (Server, Disk, or Switch) provide any real detail around this subject, potentially because everyone’s data centers are different. There are many facets of cable management depending on how detailed you want to get. These include:
•Inside the storage device rack
•Inside the server device rack
•Inside the physical network layer
All of them are equally important to the bottom-line of why we want to make sure all our cables are dressed and labeled properly. The most important line being “trouble-shooting”. Those of you who have had the luxury of visiting environments with no cable management will appreciate the true need of ensuring this is accomplished properly.
Rather than digging thru a spaghetti mess of cables, wrapped around power-cords, twisted in knots, fiber cables being zipped tied and etc., you will have the luxury of knowing where everything is from your documentation and physical inspection. Other aspects of cable management also assist in hardware maintenance of servers, switches and storage, as well as new server, switch and storage physical rack installation (e.g. not having to unplug 12 production switches in order to install one new one because of poorly designed cable plant). Finally when it comes to data-center cooling, having a properly strung cable plant will allow for all your systems to be cooled efficiently versus having the cooling environment blast thru a mess of cabling.
As important as it is to have proper physical cable management, it’s just as important, if not more so, to properly document that information. This can be done in any number of ways. Choose whichever suites you best, but the key here is constant maintenance of the cable plant and then documenting the changes as they occur in your physical layer. As with any data-center process, cable-management needs to be repeatable over and over again so ensure the cable management guidelines you deploy in your facility are adhered to.
Only you and your team can track down to the physical layer. This may seem like a lot of busy work but when you have issues or a new storage / switch rollout you will thank yourself that this detail is written down somewhere and easily referenced. Not only does it allow you to easily locate things, but it will allow you to logically add new connections into the cable plant, prior to physically doing it, thus giving your accurate counts of switch/storage ports available in your environment.
If you haven’t looked into implementing proper procedures for cable management, I encourage you to do so. It’s painful and time consuming but it will greatly assist your organization during times of install and maintenance. At the end of the day, having properly strung glass and copper cables is going to make the life of the technical system admin much easier to bear.
Part II in this series on data center management will look at labeling of cables, servers and storage systems.