Course Enrollment System – KAIST

KAIST replaced Sybase with Altibase to resolve the strains of massive concurrent logon attempts while providing a solid performance.

Use Case: Course Enrollment System

Current Customers: Korea Advanced Institute of Science and Technology (KAIST)

Industry: Public

Who

Korea Advanced Institute of Science and Technology (KAIST) is one of South Korea’s top universities specializing in science and engineering.

Problem

KAIST suffered from its legacy DBMS’s inadequate performance and reliability with its online enrollment system.

  • Due to massive logon attempts during the enrollment period, KAIST’s legacy DBMS could not handle the load. As a result, logon attempts were unsuccessful, and in-use disconnections were widespread. This led to inevitable system downtime.
  • Increased number of online lectures led to an increase in online users. As the number of connecting sessions increased, KAIST‘s legacy DBMS could not keep up with the workload. As a result, KAIST’s efficiency and reputation were suffering.
  • KAIST was in dire need of cutting edge technologies to remedy those problems.

Solution

KAIST replaced Sybase with Altibase in 2003. Altibase in hybrid mode with replication features allowed KAIST to effortlessly handle the increased capacity while providing reliable HA.[1]

Results

  • KAIST provided normal services during enrollment periods.[2]
  • The spike in system demands during enrollment periods was no longer a problem, and concurrent logons proceeded glitch-free.
  • Altibase’s in-memory feature drastically boosted performance speed.[3] With Altibase’s hybrid architecture, all historical and current account information is now accessed and processed seamlessly.
  • Readily available personal, proprietary and data-driven information permitted KAIST to process a vast number of course enrollments and online courses.

 

 


[1] KAIST installed Altibase in hybrid mode on an SUN FIRE 15k system with 1.2 GHz 8-core; 64 GB of memory with an active-active replication configuration.

[2] KAIST’s previous DBMS system could handle at most 300-400 connecting sessions at any given instance. Now, thanks to Altibase, KAIST can handle more than 1,000 connecting sessions without impacting overall system performance or stability.

[3] KAIST processed 800,000 transactions periodically in the legacy system. KAIST now processes 1.2 million transactions in real time.

 

Downloading is believing.