May 8 2018

Cloud Computing Costs & Reducing Them

Posted by Mike Brunt at 12:40 PM Web Servers | DataBase | CloudComputing | Tomcat | Languages | Performance | JavaEE | Caching | Greening The Web | Java-JVM | ColdFusion | JRun-J2EE

I have worked in capacity planning and tuning servers since 1999, in many cases I used load-testing tools, starting with Segue Silk Performer and eventually I ended up using JMeter, which I still do to this day.  These tools typically work on the principal of directing a browser to a proxy mechanism and then browsing a web site where the proxy mechanism captures all browser interactions and files-media returned, which is turned into a runnable script.  The makeup of these scripts has changed so much over these past 19 years and not for the better.  There are so many call-outs to other web sites in them now, that the longest part of preparing them for use, is removing call-outs that might look like DDOS attack when I ramp up the load. Each of these call-outs include data=$$$ in cloud-based environments where we pay for overall computer power like we pay for electricity.

Because of the high demand for developers (particularly as Blockchain demand grows) there is often hastily created code which causes performance and security issues, hence calling out to me for help.  To make matters worse H1B visa applications through 2019 are already used up, so for those in a traditional cloud environment like AWS, costs are likely to rise.  I am looking into all of this to see if I can offer a service to target reducing costs in cloud-based environments, by tuning memory, threads, SQL and heavy-duty code.