Efficient management of resources is in particular a considerable challenge for network and communications testing, especially in large or distributed testing labs. Eitan Lavie, director of technologies at QualiSystems, explains the benefits of an automated resource management system.

For network equipment manufacturers and service providers, network testing is a mission-critical, multifaceted task that is essential if products and services are to be high quality and priced competitively. Testing efficiency and efficacy directly impacts time-to-market, product quality and overall profitability including capital and operational expenditures.

Testing teams find themselves needing to manage thousands of local or networked resources. These resources can be used by a single station or shared between multiple users, in a single location or across different labs. If these expensive resources are not managed properly, they are under-utilised or end up in usage conflicts within and between test teams. This translates directly into expensive bottlenecks or delays.

Let us consider in more depth some of the issues that can arise from a lack of effective resource management. They add up to a pretty long list of problems for the test engineers themselves and for the organisation as a whole. Some of the bigger issues include:

Utilisation and cost: Equipping and running test laboratories is a costly operation. Physical resources – equipment like traffic generators or switches can run into millions of dollars, and are expensive to replace if they are damaged or become outdated. Logical resources (managed functionalities like switch node configurations or software licenses) are often costly as well. So any level of utilisation under 100 percent is an instant money-loss for the organisation.

Setup: As network environments and equipment become more complex, it gets more and more time consuming to configure and set up equipment for testing. Required configuration adjustments for different test batches or Devices Under Test (DUT) often mean our engineers have to manually wire up and re-cable the equipment, and sometimes this can lead to time consuming configuration errors.

Sharing: In most organisations, test resources must serve multiple test teams. Sharing these resources can be quite frustrating. Making sure the test equipment you need is available when you need it, and furthermore – guaranteeing your test is not being interrupted just a second before it ends, is sometimes an impossible task. Resource use conflicts between individual testers and testing teams, plus inefficient use of expensive equipment result in longer, unpredictable testing times and efforts.

Control: When sharing resources across large or distributed test labs, resources are often stored in test-equipment farms. Control over the equipment becomes increasingly difficult as more and more Layer 1 Switches are required to automatically manage resources cabling and connections. Control gets even trickier as large clusters of switches with complex topologies are required.

Meeting with network equipment manufacturers and telecom service providers’ testing teams reveals these challenges are faced on a daily basis. While trying to figure out potential solutions to these pains, one can consider that the continuous evolution of automated testing solutions can be deployed to the rescue here as well. Testing teams can improve the way they manage multiple test resources, while save time and avoid frustrating efforts by using a centralised, automated resource management system.

Some of the crucial issues I mentioned above can be solved by ‘good friends’ like a scheduling and reservations mechanism, an application that automatically calculates switch topology paths and a well-managed cluster of Layer 1 Switches.

Automatic topology configuration: As mentioned earlier, to save re-configuration and manual re-cabling efforts, Layer 1 switches can come in very handy. The Layer 1 switch provides a wire-once technology, which users can use to connect any port to any other port within the system using software commands. Once the lab network and equipment is connected, test topology changes are accomplished by a simple mouse click, and there is no further need for manual cabling.

As multiple labs share the same resources, multiple Layer 1 switches are required and topologies get complicated once again. One solution would be to wrap all the switches into one cluster or cloud and then control it using a background service to automatically calculate the optimal topology, according to the test requirements and definitions. Ideally, the test themselves remain unaware and unaffected by the solution in the background, and doesn’t require any modification.

Automatic reservation and scheduling: An automated service that can lock and release test resources according to predefined rules can help avoid those situations where different test engineers are stepping on each others’ toes, and interrupting each others’ tests. With all resources under a centralised repository, managing resource reservations between test executions, test development teams and other resource customers is highly effective and helps eliminate resource usage conflicts.

On top of a locking service, a scheduling capability can upgrade lab control. Being able to automatically schedule and allocate specific resources at specific times, can help guarantee optimal resource usage, and minimal test interruption.

However, some flexibility is required. For example if a test overruns its allocated time and requires an additional few minutes to finish a long and complex test process, it would be a real shame not to allow it to run to completion by reallocating a resource to another test. In order to avoid cases like this and to assure that the test with the highest priority gets all the resources it needs, a rules engine that defines a hierarchy for tests and resource users, and that also provides an automated conflicts management system is a great plus.

Coordinated Execution: When dealing with growing multi-site and large scale testing environments, achieving timely and cost effective test execution requires not only well managed resource sharing but also control over test stations, test requirements and an ability to match and coordinate all these components together. In order to automatically allocate tests to the appropriate available station, we need a centralised application that can dynamically abstract resources by defining their structure (chassis, card, port and so on) and capabilities like protocols and rules etc; and to remotely control all resources and stations. With these tools, a well-coordinated execution can be achieved.

To return to our initial set of challenges, we have seen that in network testing labs where multiple users design and execute large numbers of tests for different product components, poor test resource management directly impacts on test efficiency and from there affects product quality and overall bottom line profitability. To maximize testing, test labs need to share their resources efficiently and avoid lengthy test setup times.

The centralised automated resource management system that I describe, combining dynamic resource definition, resource reservation and locking, and automatic test scheduling and topology configuration, with remote control over test resources and test stations for coordinated execution, is not just some futuristic concept. It is a concrete solution that leverages the value of physical, logical and human test resources and increases overall profitability by maximizing test lab capacity.

In the final analysis, an automated resource management solution can help manufacturers and service providers achieve consistent high quality products and services that will ensure their ability to succeed in the highly competitive world of networking and telecommunications.

www.qualisystems.com