Skip to main content

Tuning and optimization J2EE web application for HighLoad

Last few months we are developing a portal for 3rd largest bank in Europe. Unique visitor of the bank grows more than 1 million visitor in a day. The main non functional requirements of the project is the high availability of the portal and giving high through output. One of the main feature of the portal is to giving user to customize their pages with widgets and provide different services for targeted auditory. After a long discussion and analysis, bank decided to use java based engine to build up the portal and we have got the following stack:
1) Java 1.7_47
2) IBM WEBSphere 8.5 as Application server
3) Nginx as web server
4) Alteon as load balancer
5) Oracle 11gR2 as DataBase
6) SOLR for content search
Main challenge for us to supported legacy browser such as IE8, opera 12 e.t.c and one portal for all device (desktop, smart phone and tablet pc). Java based portal engine generated a lot of java script which didn't give us very good performance. For these above reasons we decided to use hybrid method of page rendering (server side (jsp) + client side (java script)) and rest service for business functionality. We eliminate of implementing any business logic in DataBase because, RDBMS is not suitable for scaling and minimize the network roundtrip. Here is our main design decision:
1) Implementing business logic through Rest service in application server
2) Serving all the static content from web server
3) Cashing is as much as possible in every layer
4) Hybrid method of page rendering (server side (jsp) + client side (java script))
Now it's the time for describe briefly what we have done in every layer, most of all steps are well known and i would like to summarize it in one place:

1) Web server optimization (nginx):
- Gzip compression level 6 for xml, json, css, html e.t.c
- Cache control Http header for 3 days
- Cache control for java script
- Etag
2) Client side optimization:
- Minify java script and CSS
- Minimize http request from browser to server. At the beginning we have more than 150 http request from browser to server. It should be remember that modern browser can make 7-8 request at a time to one domain
- Optimize every images (lossless)
- Using CSS sprite
- Aggregate CSS and JS in few files
- Minify Html
3) Server side (backend) optimization:
- Caching every rest response
- using distributive EhCache
- Hibernate + MyBatis second level Cache
- Optimize Connection Pool for database in IBM WAS
- Optimize heap size and GC policy for JVM in IBM WAS
- Optimize thread pool size in IBM WAS
- Optimize session management for IBM WAS
- Scheduler to drop long running and hanged SQL connection from IBM WAS (There is a bug in IBM WAS with connection pool)
4) Database optimization
- Using result cache for Data dictionary
- Move Data dictionary to Oracle KEEP POOL

If you like this article, you would also like the book
Labels:

Comments

Post a Comment

Popular posts from this blog

Tip: SQL client for Apache Ignite cache

A new SQL client configuration described in  The Apache Ignite book . If it got you interested, check out the rest of the book for more helpful information. Apache Ignite provides SQL queries execution on the caches, SQL syntax is an ANSI-99 compliant. Therefore, you can execute SQL queries against any caches from any SQL client which supports JDBC thin client. This section is for those, who feels comfortable with SQL rather than execute a bunch of code to retrieve data from the cache. Apache Ignite out of the box shipped with JDBC driver that allows you to connect to Ignite caches and retrieve distributed data from the cache using standard SQL queries. Rest of the section of this chapter will describe how to connect SQL IDE (Integrated Development Environment) to Ignite cache and executes some SQL queries to play with the data. SQL IDE or SQL editor can simplify the development process and allow you to get productive much quicker. Most database vendors have their own front-en
Post a Comment
Read more

8 things every developer should know about the Apache Ignite caching

Any technology, no matter how advanced it is, will not be able to solve your problems if you implement it improperly. Caching, precisely when it comes to the use of a distributed caching, can only accelerate your application with the proper use and configurations of it. From this point of view, Apache Ignite is no different, and there are a few steps to consider before using it in the production environment. In this article, we describe various technics that can help you to plan and adequately use of Apache Ignite as cutting-edge caching technology. Do proper capacity planning before using Ignite cluster. Do paperwork for understanding the size of the cache, number of CPUs or how many JVMs will be required. Let’s assume that you are using Hibernate as an ORM in 10 application servers and wish to use Ignite as an L2 cache. Calculate the total memory usages and the number of Ignite nodes you have to need for maintaining your SLA. An incorrect number of the Ignite nodes can become a b
Post a Comment
Read more

Load balancing and fail over with scheduler

Every programmer at least develop one Scheduler or Job in their life time of programming. Nowadays writing or developing scheduler to get you job done is very simple, but when you are thinking about high availability or load balancing your scheduler or job it getting some tricky. Even more when you have a few instance of your scheduler but only one can be run at a time also need some tricks to done. A long time ago i used some data base table lock to achieved such a functionality as leader election. Around 2010 when Zookeeper comes into play, i always preferred to use Zookeeper to bring high availability and scalability. For using Zookeeper you have to need Zookeeper cluster with minimum 3 nodes and maintain the cluster. Our new customer denied to use such a open source product in their environment and i was definitely need to find something alternative. Definitely Quartz was the next choose. Quartz makes developing scheduler easy and simple. Quartz clustering feature brings the HA and
Post a Comment
Read more