Author: Michael Kopp.
Looking at statistics provided by monitoring activities can be helpful in identifying how and where improvements are needed. Michael discusses how simple averages are not particularly useful and provides alternate ways to use the data, focussing on the use of percentiles.
Author: Todd Hoff.
Live video is very demanding. At Justin.tv, with less than 250 ms delay required of a system that sees 30 hours per minute of video uploaded and daily peak bandwidth of 110 Gbps, the system has to be robust, scalable, and high-performance. This interview with Kyle Vogt, Justin.tv Founder and VP of Engineering, outlines how they do it.
Author: Javier Andrés Alonso.
Javier Alonso explores a few basic options for configuring InnoDB logging to improve performance.
Author: Dimitri Kravtchuk.
With the new changes of InnoDB code supporting OLTP read-only performance and full text search (FTS), Kravatchuk wonders if it can now replace MyISAM. This article covers his extensive tests to discover that InnoDB performs better in most, but not all, areas. FTS is notably now much better in InnoDB than MyISAM.
Author: Tim Callaghan.
In this comparison of fractal tree versus MongoDB indexes, inserts are quicker with fractal tree indexes on tables over about 250,000 inserts in this test environment. In fact the test to 5 million inserts had to be abandoned as Mr. Callaghan didn’t have the 2 weeks MongoDB would have needed to do it!
Author: Jonathan Klein.
Etsy’s latest quarterly performance report shows they are keeping up well with performance needs. Highlighted is the appropriate use of both real user monitoring (RUM) and synthetic test metrics as well as importance of hardware upgrades.
Author: Joshua Bixby.
Real user monitoring (RUM) is starting to come into play at a high rate. RUM data shows what’s happening in the real world with all its variables and is useful to anyone outside the internal devops team to know how users are experiencing the site. (DevOps have more use for synthetic data with limited variables)
Author: Sebastiaan Meijerink.
Author: Andreas Grabner
Does your application architecture include Message Queues to feed work items to backend batch processing such as “Update Product Inventory” or “Send out Notification Emails”? Message Queues work well in these use cases as they decouple your system components and allow your backend processing to asynchronously process requests. Is your XML implentation working as well as it can? This case study shows a 1000% performance increase simply by changing the XML parser.
Author: Luis Abreu
5 quick and simple tips to help improve websites that have images on them.