ANTONY SAVVAS TEST & MEASUREMENT
TEST AND MEASUREMENT:
How are test and measurement companies rising to the challenge of the impact of AI on optical networks and components, and how is AI changing their product set? Antony Savvas considers this evolution in the market, and how the test and management ecosystem will change going forward. COMPANIES ARE RISING TO THE CHALLENGE OF AI
O ptical test equipment is used to evaluate and ensure the performance of optical networks and systems, using tools like optical power meters, spectrum analysers, and network testing kit. Key drivers for growth for the market include advancements in fibre optics, increasing demand for high-speed internet, regulatory changes driving wider deployment of optical networks, and increased investments by telcos in their infrastructure. OBSTACLES But the high cost of advanced test equipment, as well as the fast pace of technological change in 5G and IoT, for instance, can be obstacles to addressing optical demand. However, increasing automation and the use of artificial intelligence in testing processes lay the possibility to streamline testing and enhance accuracy. And developing smart optical testing equipment in response to the demands of AI among service providers, can reduce human error and improve efficiency in the midst of a test and measurement skills crisis. xzProviders of hand-held and bench top AI testing kit and services include the likes of EXFO, Corning, Anritsu, Yokogawa Electric, Keysight Technologies, Fluke, AFL, Aplab, Spirent, Tektronix, Optikos, VIAVI, Gold Lite, and a host of others.
REQUIREMENTS Hamdy Farid, senior vice president of Product at connectivity intelligence specialist Ookla, says of testing requirements: “Gathering and providing data alone is no longer enough for test and measurement companies. The next wave of advancement involves transforming data as a raw commodity into actionable insights and predictive analytics, that can help providers understand how to deliver better real-world connectivity experiences to consumers.” Farid says Ookla is “actively leveraging” AI to help improve connectivity across a variety of use cases, helping it to solve common, yet costly, network planning and optimisation use cases, like automatically identifying and providing recommendations to solve overshooting cells and using cluster analysis for anomaly detection. “From an end-user perspective, it helps us to identify where to prioritise network optimisation or expansion to benefit the most users,” Farid says. AI PARADOX VIAVI’s chief technology officer Sameh Yamany outlines the paradox facing the industry. He says: “Artificial intelligence is having a profound impact on the network ecosystem. Increased data volume and complexity, maintaining high reliability and performance for intensive AI functions, and security
against increasingly complex threats are just some of the challenges posed by AI, when it comes to optical networks and components. But the solution to many of the challenges posed by AI actually lies within AI itself. Network test and measurement companies are now taking an ‘AI testing AI’ approach, leveraging the enormous analysis and automation potential of AI to provide lab-based testing, real-time insights, predictive maintenance, and more.” Yamany says AI has “revolutionised” network test and monitoring capabilities, allowing providers to facilitate things like automated anomaly detection, root cause analysis, network stress testing, and performance optimisation. This is leading to more precise and faster testing capabilities, allowing for “better handling” of high- speed network requirements, and more efficient troubleshooting, reducing the burden on field deployment teams, says Yamany. As for the kinds of things players are able to better test using AI, enhanced performance metrics such as latency, jitter, and packet loss can now be monitored in real-time. “There is also a focus on thoroughly validating the performance and integrity of next- generation transceivers like 800G and 1.6Tb, and high-density fibre optic components, which are particularly crucial for supporting AI and ML applications in data centres and cloud environments,” he adds.
www.opticalconnectionsnews.com
21
ISSUE 38 | Q3 2024
Made with FlippingBook Digital Proposal Creator