In the quest to reduce operational costs and streamline business processes to improve company performance, organisations are increasingly Web-enabling their business critical applications and centralising operations to the data centre. This “virtual proximity” makes a lot of sense, but in making the move, many IT managers are discovering how ill equipped the WAN is to provide the backbone of application delivery. Slow or inaccessible applications mean infrastructure investments are wasted and employee productivity suffers dramatically.
My job leaves me all too familiar with the daily struggles of applications as they try to traverse the complicated labyrinthine bottlenecks that make up the WAN, but experience tells me that many struggle to understand what these poor applications have been through before they appear on a remote desktop. So to refresh your memory (and because it seemed like a fresh perspective over the Christmas break), I asked one application to complete a diary and tell us exactly what it gets up to in a typical day. So how would an application describe its journey across the WAN (if it could!)?
0:01 – Boot up. Things certainly happen fast at the server!
0:05 – Sync up briefly with the other applications in the office today. Apparently the CRM system is getting upgraded later on.
11:30 – Leave the office on my way to a branch office in Adelaide. I am travelling economy class with everyone else on the WAN. Stuck next to a particularly awkward VoIP call – it’s going to be a bumpy journey.
11:34 – Stuck in traffic just outside the data centre. This is a bottleneck every morning. Someone needs to do something about it.
11:52 – Arrive Adelaide, 21 minutes late. Well, most of me is here – just waiting for some crucial code so I can get my work done.
11:58 – The rest of me arrives in Adelaide. Successfully do my work. Leave the office with smiles all around. I think they have forgiven me my late arrival.
12:25 – Get back home to head office. Sync again – gosh, a few things were upgraded while I was gone!
12:26 – Need to go to urgent meeting in Lagos, book priority ticket.
12:32 – Arrive Lagos. I am not used to working with this sort of equipment. Would have got here earlier, but someone accidentally requested a Wweb application upgrade which took up most of the room for the journey.
21:00 – Rest for essential maintenance.
Ok, so applications do not complain about bumpy journeys, but you get the gist. This story from a typical application demonstrates that traversing the WAN is a complicated operation with latency and congestion obstacles. What we can conclude, therefore, is that the move towards centralised and consolidated IT operations in the name of business agility must be seen in the light of its impact on remote users and the network.
Poor application response time come down to two basic factors – congestion and latency. Congestion occurs when there is simply too much traffic for the network to handle, causing delays or information to be discarded altogether. As outlined in the application diary above – parts of data quite literally get left behind. Too much data loss makes it impossible for an application to work at all. Latency on the other hand has long been the cause of performance degradation and to be quite honest, there is really no way to get rid of latency. Latency is the time it takes for light to travel between points and the reason it is so tricky to address is because a transaction is not just a one-way trip from client to server – there are many, many roundtrips required to transfer the content in just one transaction if everything goes perfectly – plenty more if there are complications.
The good news is that state-of-the-art WAN optimisation can easily address congestion and while latency cannot be eliminated, the effects can be mitigated. With the majority of employees working outside of headquarters, enterprises today must have a plan in place to support these remote workers and the poor old applications trying to reach them! If productivity is a measure of work completed within a certain timeframe – then delivering better application response within that timeframe is not just a bonus but an essential contribution to the bottom line.
Steve O’Brien is the Vice President of Expand Networks APAC