A server farm in Chicago. (Photo: Rich Miller)
In this edition of Voices of the Industry, Danny Gonzalez, Director of Business Development at Anritsu, shares information on how to test the limits of your data center’s edge network to ensure it delivers the services that applications your customers need.
It had to happen. What started out as applications running from our centralized data centers (DCs) have grown and moved. Data centers (generally speaking) mimic our own evolution if you think about it. For many of us, much of our youth and activities revolved around the core of our home or family. We were able to test the limits and learn new things under the protection of a good support system. Then came that day when we took the next step and stepped out of this controlled and centralized environment – building a new life on the foundations of our youth. Just like us, while data center applications still have those connections at the core that they will always carry, they now function as a decentralized data center and live on the edge – with different rules and processes, and where their needs are better met. .
Applications like cloud services, gaming, ADAS, virtual reality, and social media require more than a centralized DC can provide: lower latency, faster connectivity, higher bandwidth , better performance. To deliver on their promise, they must migrate to a distributed DC where they can benefit from cutting-edge computing and realize their true potential. Yet, just like our own evolution, life outside the comforts of home tends to be a lot more work. Deployments should be designed to maintain existing 10G / 25G infrastructure and equipment while supporting newer 100G / 400G broadband networks. With increasing demand, building and testing these new 100G and 400G networks must be done quickly. While the new 400G optical modules are smaller, consume less power, and have a higher density, network administrators and field technicians need to optimize them at all levels (equipment, transceiver, and network) to be successful. ensure they meet error tolerance and forward error correction (FEC) parameters. The move to the edge gave domain controllers more capacity and also more to manage. So how do you make sure that everything goes together? Once again (as in life) you are testing the limits of your edge network to make sure it is providing the services your end users need.
- Check the performance of each application as its own service with a specific QoS – each service has a given SLA that dictates how the network will respond to it. Some services are low latency while others are bandwidth dependent. Although they all travel on the same physical medium, they are treated very differently depending on the location of the edge computing resource and the virtualized network created to support it. Once the applications are classified by service levels, you have the possibility to validate the performance of each one.
- Take advantage of benchmark tests deployed in the field to ensure interoperability and that SLA performance and QoS benchmarks are met – benchmark testing is like using your favorite family recipes – they provide a step-by-step process for ensure that each service meets its minimum accepted industry criteria as it crosses multiple network equipment vendors and helps you resolve any issues that may arise.
- For 5G applications, take latency sensitive measurements – timing is everything, especially when talking about 5G. After all, 5G latency sensitive applications are one of the determinants of peak computing resources. Characterizing this latency performance is a key metric for a successful deployment. Regardless of how you perform these measurements, whether end-to-end, stand-alone, and / or benchmark, they are essential to ensure that performance.
Yes, change can be difficult, however, it doesn’t have to be a seemingly overwhelming challenge. The good news is that just as most of us have made it out of the secure space of our youth, so will our data centers. Just as we have discovered the tools we need to help us successfully navigate the evolution of life, there are tools available that can help you navigate the transition of your applications to advanced computing.
Danny Gonzalez is Director of Business Development at Anritsu with over 19 years of experience in the testing, development, training and execution of digital and optical transport. Anritsu’s MT1040A Network Master ™ Pro 400G is a versatile and portable solution that will help you seamlessly implement a variety of network deployment test techniques as well as perform hardware equipment verification testing. . This network testing solution is designed to evaluate transport operating at speeds from 10Mbps to 400Gbps – making sure you can support your legacy base while implementing new, state-of-the-art computing resources.