Those
of you who have been doing Capacity Management for a while will read through
this post and nod and understand. Remember those days when people would
call you and say that an application was "slow" and the first thing
you'd do is verify that there was enough hardware capacity to handle the
existing load? CPU utilization not too high? Check. I/O rates
reasonable? Check. Enough memory? Check.
When
I was just starting out as a capacity planner sometime last century, one of the
missing key pieces of information we always wished we had was the end-to-end
response time of business transactions. Sure, we could always eliminate
pieces of the infrastructure as potential capacity bottlenecks, but it was
always with the wish that we could easily see how much time is being spent
within transactions and where that time is actually being spent.
Something that told us where to focus; where to look next.
Efforts
were made over the years in making ARM (application response measurement) more
mainstream. ARM is an attempt to insert lines of code into applications
that would capture statistics about the response time. While there's
still an ARM working group within the Computer Measurement Group (CMG), ARM has
never taken off in a widespread manner. My theory has always been that
people that are writing applications have no desire to add code to write
response time data into log files for performance and capacity geeks like me.
The cynical side of me always wondered if commercial software houses
really wanted the organizations they sold to to really know how poorly the
application performs in the real world.
A
relatively new technology (relatively is one of those words that means
different things to different people -- some of the mainframe people I know
think anything done since 1980 is pretty new) allows organizations to capture
this type of data without having to alter code within applications.
Application response monitoring (APM) software allows organizations to
look at how well transactions perform, where bottlenecks are, and how well
service level agreements are being met.
Metron
sells an APM tool called SharePath -- it captures end-to-end information on
every production transaction that's completed. To us, that's important.
No synthetic transactions are generated and no sampling of the production
transactions take place.
For
years we've been users and sellers of athene® -- a leading Capacity Management
software solution. SharePath and athene® are complementary solutions --
athene® does an expert job of looking at the hardware infrastructure and
helping to monitor, manage, and predict future needs, while SharePath looks at
the transactions and how well they are (or are not) meeting required service
levels. Data from SharePath can be brought into athene®'s Capacity
Management Information Store (CMIS) for use within a client's existing
reporting structure and to help the Capacity Manager make predictions and key
recommendations to management.
On
Thursday May 9 we'll have a webinar that looks at Application Performance
Monitoring and why we feel it's an important piece of our 360-degree Capacity
Management strategy. You can register here:
http://metron-athene.com/services/training/webinars/index.html
Rich Fronheiser
Chief Marketing Officer
