| How to Get Unbelievable Load Test Results |
Misc. Articles |
  |
Like the fantastic characters in a Ripley's museum, this article examines some fantastic results produced by performance experts. Come inside and see how to keep your response time constant under any load! After all, seeing is believing ... or is it?
|
| Of Buses and Bunching: Strangeness in the Queue |
Misc. Articles |
|
This article explains how correlated or bunched requests can impact capacity planning results. An especially important example is the impact of highly correlated Internet traffic on sizing buffers in IP routers and HTTP servers. |
| Capacity Calculations: Handle with Care |
Misc. Articles |
|
This paper discusses avoiding calculation results that are more precise than is justified by precision of corresponding measurement input data.
|
| "The LA Triplets" Quiz |
UNIX® Load Average Series
|
|
This is a little quiz to test your understanding of the triplet of numbers that appear in the UNIX® load average (LA) performance metric.
|
| UNIX Load Average Part 1: How It Works |
UNIX® Load Average Series
|
|
In this online article Dr. Gunther digs down into the UNIX kernel to find out how load averages (the "LA Triplets") are calculated and how appropriate they are as capacity planning metrics. If you haven't taken the quiz, you should consider doing so before reading this article. With apologies to Jules Verne and his novel "20,000 Leagues Under the Sea" I might equally have entitled this piece, "20,000 Lines Under the Shell." |
| UNIX Load Average: An Addendum |
UNIX® Load Average Series
|
|
Several readers have queried my statement in Section 4: Kernel Magic of my online article that the sampling rate of the calc_load () function in the Linux kernel is once every 5 seconds rather than once every 5-th second. This addendum tries to address that confusion. |
| UNIX Load Average Part 2: Not Your Average Average |
UNIX® Load Average Series
|
|
This is the second in a two part-series where we explore the use of averages in performance analysis and capacity planning. Part 1 described some simple experiments that revealed how the load averages (the LA Triplets) are calculated in the UNIX kernel. In Part 2, I'll compare the UNIX load averaging approach with other averaging methods as they apply to capacity planning and performance analysis. |
| UNIX Load Average Reweighed |
UNIX® Load Average Series
|
|
This is an unexpected Part 3 to the discussion about the UNIX load average metric. Part 1 describe what the kernel code does, while Part 2 explains how the load average actually computes the equivalent of something called an Exponential Moving Average (EMA), such as you find at BigCharts.com for doing stock market analysis. I say "unexpected" because I thought I'd said everything there was to say about that topic. But recently someone asked me about where the weight factor Exp(5/60) comes from, and I realized I had somehow skipped over the details of that point. Here, I rectify that omission. As you'll see, it's a rather deep topic in itself. I hope you find it interesting and helpful.
|
| How to Write Application Performance Agents in TeamQuest Performance Software 7.2 or 8 |
TeamQuest Tutorial
|
|
TeamQuest Performance Software provides unintrusive mechanisms for instrumenting applications and analyzing application performance. This paper describes how to use those mechanisms.
|
| Commercial Clusters and Scalability |
Server Scalability Series
|
|
In this paper we present an introductory analysis of throughput scalability for update-intensive workloads (such as measured by the TPC-C or TPC-W benchmarks) and how that scaling is limited by serialization effects in the software-hardware combination that comprises any platform.
|
| How to Measure an Elephant |
Server Scalability Series
|
|
This paper shows how to apply some of the scalability concepts introduced in the previous paper to the performance analysis of real computer systems. |
| Evaluating Scalability Parameters: A Fitting End |
Server Scalability Series
|
|
This is the final online article concerning the concept of application scalability. Here, you will learn how to determine value of the parameters that control scalability.
|
