- 3 minute to 2 second Performance Challenge

ServiceDesk Plus is widely deployed by businesses across countries, time zones and sites. The challenge was to make ServiceDesk Plus scalable without complicating it. It was important to make sure that scalability to be implemented with the default database, MySQL. ServiceDesk Plus enables you to deliver world class performance with MySQL and MS SQL Server. ServiceDesk Plus is optimized for MySQL. However, we recommend MS SQL Server for high volume and large scale installations.

The Real Simulated Environment

We created a simulated environment with a help of a few customer's sample data. The idea was to create a scenario exactly as customers with large scale deployments so that we have requests with text and attachments and assets with data. The simulation was configured on a system with the following configuration with more than 5000 requests.

SetUp Details of the Simulated Environment are below

	Server Information
Operating System	Windows Server 2003
Hard Disk	80 GB
RAM	4 GB
DataBase Used	MySQL

Request Count

	Numbers
Total Number of Requests	5000
Total Number of Requests having attachments	2000
Average replies per request	3

Rows Present in DB

Tables	Rows of Data
WorkOrder Table	20000
Notfication Table	78000
WorkOrder_Threaded Table	5000

Roles and Operations:

Technicians and Requestors	Numbers
Requesters accessing ServiceDesk Plus	400
Number of Requests created per day	100
Technicians accessing ServiceDesk Plus	100
To simulate technican's operations we had also simlated random technician frequent actions such as Accessing List View Page Editing Requests Accessing Group View Request Detailed View Page

The simulated production environment is tuned before executing the operations,

1. JVM heap size is configured to 1300M

2. Mysql DataBase configuration are such that it uses 2GB for buffering the data

3. URL for /custom is excluded - to avoid context switch in tomcat framework

4. Database Connection is set to 150 with corresponding max-pool-size as 200

5. Tomcat is configured to support 180 parallel threads/requests

6. Expiry is set of .html files so that browser will not send request for html files.

7. Database cleanup - deleted unnecessary  entries in aaaaccsession, auditrecord tables etc and deleted outofoffice, undeliverable mails.

The following results were Observed in the Simulation

Tables	Previous Versions	Build 7019	Build 7019 after performance pack	Build 7500
Open Requests	170.4 secs	24.2 secs	1.84 secs	2.19 secs
OverDue Requests	160 secs	23.6 secs	2.08 secs	2.19 secs
Custom View	159.6 secs	25.8 secs	1.96 secs	2.21 secs
All Requests	172.2 secs	25.8 secs	1.96 secs	2.36 secs
Request Details Page	108 secs	13 secs	4.76 secs	4.38 secs
New Request Creation	24.8 secs	16.8 secs	5.28 secs	7.29 secs
New Req with Attachment (150 kb)	24.2 secs	17.2 secs	5.4 secs	7.40 secs
Edit Request	16.4	16 secs	5.14 secs	5.05 secs