Java™ Scaling Feature Story

Unshackle Your Growth Engines

CIOs hear it every day: Align IT with the business needs, support growth, be agile, and, of course, the now famous line: do more with less. Successfully meeting this challenge can be the difference between a great company and the rest of the pack. Effective IT strategies can drive competitiveness and both the revenue and cost variables of the profit equation.

The pressure to grow while lowering costs is a constant. Providing compute capacity as a shared network service and removing the resource constraints of traditional systems, network attached processing from Azul Systems® makes it possible to scale while minimizing total cost of ownership through systemic savings.

Fixing “problem applications,” rolling out new services quickly, and moving towards the data center of the future are often contradictory ideals. Network attached processing manages to accomplish all three! And Azul has the real world results to prove it in the following common scenarios.

Scenario 1:

Large HR Application On A Scale-Up Architecture

Problem: Needing to scale from 10,000 users to over 200,000 users, this major project was experiencing delays and uncertainty about ever reaching the business goals. Large scale servers were believed to be the only solution, but even these powerful systems were experiencing stability issues and were unable to meet service level requirements. The organization faced the prospect of procuring multiple scale-up servers with no guarantee of success.

Solution: Augmenting the deployment with network attached processing provided the solution. The company can cut the number of required systems in half while adding headroom for future growth and to absorb spikes in workload. Projected cost savings are dramatic – over 60% reduction in infrastructure costs over three years.

Scenario 2:

Large eCommerce Application On A Scale-Out Architecture

Problem: The largest application for a logistics company needed to meet exponential growth requirements over the next 2 years. A recently implemented eCommerce solution, the number one revenue generating application in the company, was struggling to meet customer requirements at the current user level. Continuously adding more and more hardware the data center was reaching a hard barrier: the data center was nearly out of floor space. The cost and complexity of the scale-out approach was hindering the business and inability to scale to keep up with demand was causing service stability issues.

Solution: Network attached processing simultaneously addressed the scaling and customer service issues. The large capacity of Azul compute appliances can easily handle the current workload and maintain response time Quality of Service even at high throughput levels. Adding capacity also becomes extremely simple. Utilization levels in the compute pools are monitored and when a given threshold is reached, an additional compute appliance can be added. Because they are a shared resource, Azul compute appliances operate at much higher utilization rates: in the range of 50-60% compared to the 10-20% utilization common with scale out systems. This method of planning and expanding capacity is analogous to the way storage is managed in a network attached storage deployment. And don’t forget about the floor space issue. By reducing the number of CPUs required to run the application and the high CPU-density of Azul compute appliances, the space issues in the data center can be a thing of the past. Read more »