Rants
Questions
Soapbox
Best Practices
Apply today for a FREE subscription to CIO Magazine!
Tue, Apr 21, 2009 16:42 EDT

|
Posted by: Thomas Wailgum in Best Practices Topic: ApplicationsBlog: Enterprise Software Unplugged
Current Rating: |
Sometimes I forget that selecting, implementing and paying for core enterprise applications are just the first steps in the long and arduous journey of managing the lifecycle of enterprise applications.
There is, of course, all the fun stuff that happens after the "go live" date: Ensuring that users are actually using the new ERP, CRM, supply chain or BI apps, that uptime stats have lots of 9s in them, downtime is minimal and planned, response times are lightning fast, and business users have a gut feeling that the overall availability of these critical applications is perfectly acceptable.
According to recent survey data from Aberdeen Research, several interesting trends have taken place during the last year relating to organizations' view of application performance.
A survey of 158 organizations in February and March 2009 reveals several bad business consequences of poor application performance:
>Nearly 65 percent of respondents reported declines in employee productivity, up from 58 percent in the 2008 survey.
>Just over 60 percent stated that problems with application performance have caused decreases in customer satisfaction, a jump up from 47 percent in 2008.
>And lastly, nearly 40 percent said that application problems damaged the company's brand, compared with 32 percent in 2008.
But when Aberdeen segmented the survey data by respondents' functional area (IT versus business), the data revealed a significant disconnect between the two camps: The business side was 93 percent more likely to report a direct impact on revenues and 30 percent more likely to report an impact on operational and infrastructure costs than IT respondents were.
In other words, IT didn't figure app downtime was that costly.
"This comes as a result of the lack of visibility and resulting understanding of how applications are impacting revenue-generating business processes," notes research analyst Bojan Simic, in the survey report.
Rather than grumbling quietly in their offices, line-of-business managers "are becoming increasingly interested in how these applications are performing," notes Simic, which, to me, is a great and much-needed development. And, in fact, the business side now wants to prevent application performance issues before end users are impacted, the top challenge cited in the Aberdeen report.
Simic writes that "it is no longer good enough to be able to address performance issues in a timely manner. Organizations are increasingly looking to be more proactive when managing application performance."
IT departments, Simic adds, face a new challenge: How to identify and resolve potential performance issues before they frustrate users by impeding their ability to get their jobs done.
Shall IT try to stop global warming as well? How about fix the banking crisis? While we're at it, let's have CIOs develop a peace plan for the Middle East.
I'm not saying that IT can't play its part in remedying the problem. There are plenty of networking tools to help do this (including the vendors that sponsored the Aberdeen survey). But network monitoring tools that can detect performance trends aren't enough.
Once again we're treading on familiar ground: the classic business-IT disconnect. This time it reveals that IT doesn't know which apps are crucial to the business, and the business hasn't cared enough to clue IT in on why those apps are so vital.
Do you Tweet? Follow me on Twitter @twailgum. Follow everything from CIO.com on Twitter @CIOonline.
Hi Bojan,
Long time no word. As typical, the proglem is ALWAYS a network issue. Over the last several years I've had to prove that the application is at fault using OPNET. About 3% of the time, it id a network issue. The other 97% of the time it is due to bad programming or using Citrix for a program that had prior issues.
Although I do not support any vendor, Cisco included, OPNET points out that it boils down to the coding and hardware supporting it. SQL is chatty and is HTTP which it terrible fo file transfers. Until we prove the bottleneck a delay time, the netork catches the grief. All on our dime with every vendor.
From my experience, a problematic application is ported over Citrix which makes it worse. My only solution is the QOS levels on the router until the application is re-designed.
Although I am a Network Engineer (geek), I could care less if it's the network nor the application. My focus is making my cutomers happy. Typically I run into chatty applications, hard drives that aren't RAID, and other cheap hardware solutions on the servers. Add in international latency along with internaional services (China runs it's infrastructure), the only way that I've been able to provide a amiable solution, it boils down to the server location and hardware.
I hope that this helps and good to see you staying on top of the quirky issues. Like you, my issues have been in multiple queues for way too long. And no good deed goes unpunished so I have to be very understanding and my boyish charm (at 39 years old :)
Best of Success ! If you'd like to chat, e-mail me so we can set up a time to discuss this in further detail while sharing our similar experiences. I think that it would be mutually beneficial.
Most Sincerely,
Gary Janssen
I have to agree with Gary regarding the application development, although I'm not sure I concur with him regarding Citrix exacerbating the issue.
I have seen in the past that the application itself is poorly tested, particularly in the Performance Volume area (if it's tested at all), and therefore no one knows how it will respond under full user mode. Only using tools after the fact, such as Quest's Foglight for DotNet, will the bottlenecks be identified. Obviously at that point the Genie is out of the bottle.
Too often the infrastructure team is getting blamed, when the network is not having problems, the SQL server is fine (barely breaking a sweat) and the Application servers are paging like there's no tomorrow.
I would argue that in many/most cases, systems designers don't know how to elicit application performance specifications from users and users don't know how to articulate them. How many times have we heard "I need it to be fast" or "I need sub-second response time" or one of my all time favorites "I need real-time updates." Systems performance modeling is a lost art, IMO.
Chris Curran
CIO Dashboard Blog
Hi Chris,
I would have to agree with your statement's. Many of the issues you have highlighted are due to lack of effort in understanding once's requirements before asking somebody to develop a product.
I hold no fancy C-level title's but i dream of what i can make it a reality. My years of experience in designing solutions for various types of customers provided me the insight to let people know what is possible and impossible up front. This has cost me dearly, but keeps me happy for not giving any false promises. I have also seen people who failed by giving false promises.
Systems performance modelling is an art, it needs creativity, but without discipline it is just gardbage.