Whether we are here as readers, writers, or both, there is something that all of us on the WANspeak blog have in common — we either have, or have had, problems with application performance.
If you are already using WAN optimization then I hope the problem for you is less than it was, but it will still be around somewhere. It’s there every time users drum their fingers on the desk while waiting to log into a portal or start an application, and it’s there every time a screen takes so long to update that they hit Enter again, just in case.
Indeed, when research firm Vanson Bourne earlier this year surveyed 650 IT directors and other professionals in large organizations across Europe, almost 80% said their key business applications — that’s things like CRM, ERP, and unified communications — had suffered performance problems in the last 12 months.
To make matters worse, 18% said the problems were already frequent and 54% of respondents said that they were becoming more frequent.
Some of the claims made for the effects of application performance problems are really rather scary. Adrian Thirkill, COO & UK managing director for network and hosting company Easynet, which co-sponsored the Vanson Bourne survey downloadable here (requires registration), says that losing a mere five minutes a day — which is easily built up, once you think about it — is equivalent to losing one percent of your productivity. It’s like a train that loses a few moments at each station: the individual delays are small, but the overall effect is that it arrives late so you miss your connection.
And with the average enterprise application portfolio costing $360 per employee per month to supply and run, according to this report, those five minutes represent a double-hit on the organization’s bottom line.
How can you tell if your applications are suffering? More importantly, how can you tell where and why the problems are arising, so you can fix them? Well, there’s the rub — most organizations can’t, not easily anyhow.
To start with, the problems will have built up over time. Networks grow organically, and the task of keeping track has been made far tougher in recent years with the rapid changes that have seen the arrival of VoIP and other elements of unified comms, plus cloud computing, SaaS, and ever more video and social traffic. Plus those new web-based workloads probably have completely different profiles to the client-server traffic that your network was set up to carry.
The next issue is that application performance is hard to pin down. For a start, most enterprises don’t have enough information about how their network is being used, which applications are consuming bandwidth, and so on. Even when they do have metrics, they may be too abstract to be of real use here — metrics such as server response, packet loss, and jitter will not tell you what the end-user is really experiencing.
Lastly, the issues are also human and psychological. The fact is that people adjust and adapt, so while they gripe at first, over time they stop reporting application performance issues. Instead, they go and make another coffee while the application loads — and more importantly, they lose the train of thought that lead them to open it in the first place. The result is a general yet unreported decline in business performance and productivity.
The key thing then is monitoring. There are specialist application performance management (APM) tools, but you probably already have some to get started with. For example, Silver Peak WAN optimization devices can do application classification and prioritization, with deep packet inspection to track even complex applications. Also, ask your network provider if it offers application monitoring and management services. If it does — and if it doesn’t, it should — you may be surprised, and perhaps even shocked, by the results.