FEATURE FIBRE OPTIC SYSTEMS
Change Management – Confronting the Erosion of Order On the Rocks By Keith Sullivan, Global Datacentre Director, Corning Cable Systems
Change management is a strange expression when you think about it. It implies that there must be a need elsewhere for managing things in permanent stasis, where the onset of time doesn’t exist.
Network cabling infrastructures are extremely susceptible to change, and that shouldn’t be a radical concept for anyone to understand. First, there are all the upgrades. 1Gbps Ethernet used to be a pretty whizz-bang core infrastructure standard back in its day, and only a few years have passed to put that well and truly into shadow. Then, what about all the growth in scale of network infrastructures to cope with our appetite for data in every facet of life, or the impact of greater compliance, virtualisation or energy efficiency goals? However, the biggest change agent
of them all is the cumulative effect of all the small, incremental adds, moves and changes that happen over time. Like tiny water drops landing on bare rock, each one barely registers. After a while though, the rock has been changed forever. In exactly the same way, a few years
down the line, the infrastructures that built advanced structured cabling solutions don’t look the same in reality as they did on the original blueprint. Bundles are more taut or slack, cable trays a little untidier. Look carefully and the odd dangerous twist or kink can be found among the miles of cable distance you took so long to ensure were perfect on day one of the network’s life. To put it another way, at a scale too small to resolve itself in plans and blueprints, your network evolves inevitably from an ordered state to a less ordered one. A laterally thinking physics student might refer to this phenomenon as a kind of ‘entropy’.
No one likes it
A little sensitivity to change is something that we all suffer from. But sensitivity to change is not something you want to impact your key IT services in the middle of the night. You expect each length of your complex cabling infrastructure to be a pathway between two important places, not time-bombs waiting to go off when you least expect it. Bend-insensitive fibre (BIF) is an
innovation that overcomes the age- old physical challenge of exploiting the light-carrying and manipulating quality of polished glass without smashing it into a million pieces (the
14 NETCOMMS europe Volume II, Issue 1 2011 Anticipating change starts from day one.
www.netcommseurope.com
glass, not the light). With BIF, you can be less delicate and more utilitarian with your cabling plans and your cabling experience. It means tight radii rather than wide arcs. You get more room for manoeuvre, literally. There is a guy employed by Japanese FTTH operator NTT DoCoMo who has spent most of his weeks during the course of the last 6 or 7 years travelling the world’s trade fairs and industry seminars with a lab coat and attenuation reader to demonstrate how he can tie a reef-knot in such a cable without the reader even showing a blip. Not many people go to the demos anymore; it’s so old hat. Yet in the datacentre, it’s a much
newer hat. Datacentre professionals are increasingly keen to test the virtues of BIF; not because they want double- bows in their cabling trays, but because they don’t want a phone call in the middle of the night to tell them some tiny incremental change has ‘broken a glass’ and brought their billion dollar datacentre to its knees. Figures aren’t available for the total amount of BIF deployed today in datacentres worldwide, though Corning itself has shifted over 500,000km and counting.
Playing with fire
So how do you begin to measure the impact of datacentre entropy? Specifically, when do niggles start turning into nightmares? Where is
the threshold breached and entropy tips over from simply bringing about ‘imperfection’ into becoming a dangerous risk? We all perform regular maintenance on our cars - brake pads, oil changes and so on. But at a fundamental level, all the steel wants to do is rust. Rusting is driven by entropy, and we all know that there comes a moment in a car’s life where no amount of scheduled maintenance can outweigh the effects of that entropy. Leave that car in the wrong environment and one day it will inevitably move beyond economic repair and no amount of change management theory is going to give you the opportunity to turn back time and do things differently. Anticipating change starts from day one. Relying upon calculated entropy ‘tipping points’ in the datacentre is playing with fire. Another way of approaching the
question is to extend your use and understanding of the Mean Time to Repair (MTTR), averaged over a good sample of server-switch and switch- switch links. Each link consists of the active elements at each end of a length of cable. The manufacturers will have calculated MTTR figures for the active components. When the MTTR deviates from the predicted curve for the active elements, then you’ve potentially got entropy effects in the passive element, the cable. It would be really interesting for fibre cabling manufacturers to standardise
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60