Massive growth in network traffic is urgently driving the need for bandwidth upgrades from 100 gigabit Ethernet (GE) to 400GE in the data center. Large hyper-scale data centers have more than 50,000 optical fibers in them. With a transceiver at each end of the fiber, these data centers house upwards of 100,000 transceivers. Keeping the cost of optical transceivers low is a high priority for data center operators. To be competitive, transceiver manufacturers must find ways to drive down production costs.
Like most new technologies, the price of next-generation optical transceivers tends to drop sharply after introduction to the market, and development costs amortize as volume ramps. Next-generation transceiver technology, such as 400GE, will reach mature pricing within a year of introduction. At maturity, the cost of transceivers is directly proportional to the complexity of the design and the number of optical components. Test time contributes significantly to the overall transceiver cost. More efficient testing of the broad range of transceiver data rates accelerates innovation and lowers cost. Here are five tips that can help you significantly reduce test time and subsequently the cost of next-generation transceivers.
Tip 1. Design and Simulation
In the research and development (R&D) phase, starting with powerful design simulation software is the first step to ensure test efficiency and lower test cost. Today’s design and simulation software enable transceiver designers to optimize their designs, ensure performance and robustness, and avoid costly additional board design cycles. They can identify the most sensitive design components early on and decide how to set specifications to improve manufacturing yield. Once optimized, designers can test design performance using post-processing data analysis functionality without rerunning simulations.
Tip 2. Characterization and Compliance Test
Once deployed in data centers, marginally performing transceivers can bring down the network link, lowering the overall efficiency of the data center as the switches and routers need to re-route the faulty link. The cost associated with failed transceivers once deployed in the data center is enormous. In a hyperscale data center with more than 100,000 transceivers, even a small one-tenth of one percent failure rate would equate to 100 faulty links.
Industry standards organizations, such as the Institute of Electrical and Electronics Engineers (IEEE), International Committee for Information Technology Standards (INCITS) and the Optical Internet working Forum (OIF), generate and maintain optical transceiver specifications and define test procedures to ensure interoperability of modules from different vendors. The operating margins in 400GE optical links are the tightest of any generation yet, creating additional test challenges as small measurement errors can quickly consume the entire operating margin. Fortunately, the methods that describe how to characterize 400GE designs are becoming more stable, and engineers can review and follow the guidelines outlined in the standards when developing their transceiver characterization test plans.
Tip 3. Test Automation
Test automation software can reduce test time down from hours to minutes. Choosing automated compliance test software verified to test to the exact specifications of each technology standard is essential. Test automation software provides insights to the test engineer about any detected issues and can quickly pinpoint failures, saving hours of debug time. Test automation software guides the test engineer through setup, calibration and compliance measurements, and allows them to quickly run through test cases without being an expert on test procedures. More importantly, ensuring transceivers are compliant to standards will minimize the risk of interoperability issues with network switches and routers once installed in data centers around the world.
Tip 4. Data Analytics
The results of automated tests are available to test engineers in real-time reports. These reports typically include details about the test setup and configuration, the measurements made, the pass/fail status, margin analysis, and output waveforms. With this information, it is easy for test engineers to replicate the test scenario later. However, sometimes the sheer amount of data collected is overwhelming for a test engineer to analyze and understand. Therefore, test automation software with integrated data analytics capabilities is ideal.
Data analytics tools provide insightful analysis of test results. Visualization methods, such as line and histogram charts, show pass/fail limits and statistical information, so test engineers can see at a glance the performance of the device under test. Data analytic tools often store results in a cloud repository, making it fast and easy to share the results among global and distributed teams. Test engineers can quickly make design decisions with confidence that would otherwise take days or weeks for them to analyze and resolve.
Tip 5. Manufacturing Test
Research and development of 400GE transceivers is well underway. Engineers are still struggling with how to test PAM4 modules, and 400GE standards are continuing to evolve. 400GE transceivers have a tight time-to-market window to meet the demands of emerging technologies such as 5G and IoT, as well as the massive growth of the public cloud infrastructure. Once 400GE transceivers reach the manufacturing phase, any issues found will mean a costly rework of designs. The ideal is to design for manufacturing. Several tools can help transceiver manufacturers create manufacturing-friendly designs. Costs increase significantly as issues are uncovered later in the development process.
Get to Market Faster
The data center transceiver market is extremely cost sensitive and competitive. Test time is a significant factor that contributes to the overall cost of transceivers. By reducing test time, transceiver manufacturers can reduce costs and be first-to-market with next-generation transceivers. At each phase of the product development life cycle, there are test solutions that can be used to maximize test efficiency. These tools can shorten design cycles, dramatically improve productivity, ensure quality and significantly reduce costs.