CHAPTeR 9 DXP PeRfORMANCe OPTIMIzATION Following are the key findings from performance analysis and profiling of the components: • At the presentation layer, the page size or supplier landing page and distributor landing page was more than 2 MB each, leading to higher latency. Each of the landing pages loaded more than 20 JavaScripts and 6 CSS files. There were five images used on the pages. • Initial login process was very slow due to multiple LDAP and database calls. • On the server side, the ORM layer was performing complex table joins that further degraded the overall performance. Further to these, the following specific issues were noticed: • The database-based modules were querying all records (select * call) from the database. This is done for both “Pending requests” and “Supplier Profiles” tabs. • This will have the following impact on page performance: • Even though we are displaying 20 records (10 each in two tabs), we are querying all records (which would potentially run into thousands). • This will increase the data transferred between portal server and database server. • This will also increase the initial page load times of the page and the HTML page size. • As these queries were done via the ORM tool, it created Java objects for each database record, which could potentially increase the consumed memory. Recommendations and Improvements Based on the findings from the performance analysis, the performance engineer recommended and implemented the following suggestions. After applying all the performance optimizations, the page load time was reduced from 120 seconds to 10 seconds. 256