66 reputation
4
bio website broadcast.oreilly.com/…
location
age
visits member for 1 year, 9 months
seen Oct 3 at 12:41

I'm a performance nerd, formerly with Sun, who don't exist any more. Sigh...


Oct
3
answered What percentage of its capacity should I aim to run a network at?
Sep
24
awarded  Autobiographer
May
6
answered Calculating required uplink capacity of network switch
Mar
4
comment How many IOPS do I need? My workload bottleneck is storage
Transactions are what the user sees, IOPs are the disk IOs required to do the work. For example a bank transaction is debit followed by a credit, with each requiring a read to get the two accounts, followed by a write of the updates to both accounts. In that case, 1 TP takes 4 IOPs. 1 Transaction per second (TPS) will therefor require 4 IO Operations per second (IOPS), and so on.
Feb
28
answered How many IOPS do I need? My workload bottleneck is storage
Jul
17
awarded  Supporter
Jul
17
answered Performance issue on solaris 10 zone. Multiple pollsyscalls
Jul
17
answered Maximum throughput on switch with different link speeds
Jun
23
answered Dramatic disk I/O performance difference for two disks on same machine
May
24
answered ASP.net site on IIS7 - intermittently very slow on all remote connections, but always very fast locally
May
14
answered High await in sar
May
13
answered High await in sar
Apr
20
awarded  Editor
Apr
20
revised How do you do load testing and capacity planning for web sites?
Try to format...
Apr
20
answered How do you do load testing and capacity planning for web sites?
Apr
20
answered Read IO Wait Increase
Apr
20
answered How to measure I/O performance of ramdisk (/dev/ram0)?
Apr
20
comment strace vs. log4cxx performance hit?
Ltrace sticks an "interposer" function between the application and the library that implements the syscall, so the cost is one extra call and return, plus one printf. Plus linking the library at start-up, of course!
Apr
19
awarded  Teacher
Apr
19
comment strace vs. log4cxx performance hit?
Strace (or truss) uses ptrace, which to oversimplify sets breakpoints in the code and traps to the tracing process when they're struck. This has start-up costs in finding and setting them, and two process-switches when they're used. As a rule of thumb, think of them as being like a system call. Ltrace/apptrace is lighter-weight, and logging more so again. I was one of the apptrace authors, which we wrote to be able to trace production program. Customers found truss too heavy to allow us to use on production machines.