Clusterscale logo
Home page spacer Company information spacer Products spacer Our customers spacer partners spacer Support spacer purchase spacer contact information

Case Studies


pdf file crocus.co.uk case study
(pdf) 345kb
word document crocus.co.uk case study
(web)
   
word document INPS case study
(web)
   
word document Mymovies case study
(web)
   
We're always interested in what our customers have to say. We respond positively to feedback whether it bee good, bad or indifferent. It's what helps us stay on top of the competition!
Email: feedback@clusterscale.com

Case study - INPS

INPS is a leading UK supplier of high quality clinical management systems with an enviable reputation for excellent customer service. The company is part of the CEGEDIM Group, which supplies health care technology and information solutions to over 30,000 GPs, pharmacies and pharmaceutical manufacturers across Europe.

Cegedim RX (formerly NDCHealth UK and Enigma Health plc) is now part of Cegedim’s UK operations. They provide IT systems to over 5,000 pharmacies throughout the UK supporting pharmacy groups as well as independent community pharmacists.

The INPS office in Dundee specialize in the development of hosted, web-based software applications for large-scale solutions to meet the needs of GP practices, primary care trusts, acute hospitals and community organizations.

INPS strive to deliver innovative clinical systems combined with the highest levels of training and technical support to our users - a philosophy which has seen the Vision user base more than quadruple since the company became part of the Cegedim Group in 1999. This organic growth has been achieved whilst still maintaining high quality service levels to our customers.

The Challenge/Problem

In order to ensure high availability of IT systems to a large number of pharmacies across the UK INPS previously used a custom round robin DNS based system to load balance their Microsoft Terminal Server environment. Due to rapid growth they needed to improve the performance, availability and reliability of this system. The key issues for INPS were:

  • Cost Effectiveness

  • Resiliency – Active/Standby pair of load balancers.

  • Maintenance – Easy maintenance of servers in the pool

  • Health Checks – Rapid removal of servers failing to respond

  • Persistence (sticky) – Keeping users on the correct server

  • Performance – Ensure that servers being hit hardest get less traffic

Cost effectiveness was a key issue so several of the low end load balancers on the market were tested on try before you buy deals. While these device specifications clearly matched the requirements when it came to real environment testing they repeatedly failed to achieve expected results.

Rectifying these problems was further hampered by having to wait for 3rd line support from California for both Barracuda and Kemp.

“I would definitely recommend you to test the Pegasys Clustered Pair for your Terminal Server environment, and ClusterScale were really helpful helping us with the system tuning.”

Kepa

The Solution

In order to ensure high availability of IT systems to a large number of pharmacies across the UK INPS chose a clustered pair of ClusterLoad Pegasys units.

"We tested several load balancer appliances for our Terminal Server Farm including Barracuda and Kemp Technologies and were disappointed they did not fully meet our requirements, Specifically replication of session data between the master and slave load balancer for high availability. We tested the ClusterLoad Pegasys units and were glad to find it did pass our tests. As an added benefit the Direct Routing mode enabled us to remove the requirement for a private load balanced subnet and therefore a layer of complexity from our architecture."

The primary benefit of the ClusterScale solution over the likes of Barracuda and Kemp Technologies was the option to replicate session data between the master and slave load balancer to ensure high availability. Furthermore, the complexity of the set-up was easily reduced as the Direct Routing mode enabled INPS to remove the requirement for a private load balanced subnet.

“The implementation of the ClusterScale load balancer was a simple and surprisingly quick procedure thanks to the easy-to-use web based interface, in fact we had both units up and running in under 10 minutes.”

“The 24*7 support offered provides an extremely prompt and intelligent answer to any queries from a UK based office and the knowledge of 4 hour on-site response thanks to the use of Dell hardware provides peace of mind.”

“Since choosing ClusterScale we at INPS have had no performance or reliability issues resulting in quiet and predictable daily operations, just the way we like them to be!”

Future:

INPS recently purchased another pair of ClusterLoad Pegasys units for a new large scale HTTP based application in development. The product requires proof testing under extreme load >2GBps. Once the load testing has been completed INPS will move the ClusterLoad Pegasys units into production for this product with confidence that the system will easily match the required performance parameters.


All contents Copyright © ClusterScale 2008 | press | contact us
web site designed by revolver design //=¬