Written by Kaushik Mittra, Kai Ng, Jose Raygoza, and Pat Chou, Product Managers, Transceiver Module Group
In Part 1 of this 3-part blog series, we talked about the breadth of Cisco’s transceiver and platform portfolio. Here in Part 2, we describe how Cisco is dedicated to and uniquely positioned to ensure peace of mind for end users of Cisco Optics, thanks to the high quality and reliability of our products.
Network downtime and the accompanied service outages are prohibitively expensive and disruptive. To prevent service disruptions, network engineers rely on extremely low failure rates of equipment deployed in their networks. Some network failures occur at the component level, such as the optical transceiver. Cisco avoids such failures with a focus on stringent qualification and testing, along with high quality manufacturing of transceivers.
Transceiver Testing at the Platform Level to Guarantee Link Operation
Any notion of transceiver quality and functionality necessarily involves system level interaction between host and transceiver. What distinguishes Cisco from other transceiver vendors is that Cisco can test these interactions across the broadest range of platforms and the largest portfolio of optics. For example, the Cisco transceiver portfolio covers speeds from 100Mbps up to (soon) 400Gbps and across a variety of form factors (from SFP to QSFP-DD and others). Testing across such breadth allows Cisco to detect the widest variety of failure modes related to transceiver operation. Additionally, Cisco applies the most rigorous standards in transceiver qualification, beyond those required by standards such as Telcordia GR-468.
Standards-based specifications such as those from the IEEE, optical MSAs, and Form Factor MSAs do not always ensure interoperability between host-and-transceiver or between transceiver-and-transceiver. The most common failure modes include signal integrity challenges on high-speed electrical data paths, logical issues on the low speed electrical interface, mechanical fit issues, inconsistency in EEPROM settings, and incompatibility with the host software. Cisco completes all transceiver xDVT (Design Verification Tests) using a variety of host platforms and stands firmly behind the qualifications.
Optical Design Validation and Testing (ODVT): ODVT ensures proper optical performance of the link as voltage and temperature are varied across pre-specified ranges. ODVT tests include physical layer parametric testing of key parameters such as optical wavelength, transmit signal power and integrity, and receiver sensitivity. Additionally, to guarantee optical interoperability, each transceiver is tested against multiple other Cisco suppliers and other transceiver types if applicable.
Electrical Design Verification Testing (EDVT): EDVT ensures proper traffic and low-speed electrical interface performance of the transceiver on Cisco host platforms across voltage and temperature conditions. Successful completion of EDVT ensures overall robust link stability and recovery.
Mechanical Design Verification Testing (MDVT): MDVT ensures good mechanical fit of transceivers into Cisco ports, robustness during non-operational activities such as transportation and storage, and tolerance to minor seismic vibrations during link operation. These tests guarantee that transceivers will not get stuck during insertion and removal, and will not get easily dislodged during normal operation of the link.
Electromagnetic Compatbility and Interference (EMC/EMI): These tests guarantee the transceiver’s ability to operate in a shared operating environment alongside other networking equipment. The stringent EMC/EMI test procedures at Cisco ensure that the amount of radiation interference from transceiver operation into neighboring equipment lies within the limits defined by FCC (Part 15). Simultaneously, the transceiver design guarantees immunity from external sources of radiation that are also within the FCC limits.
Software Development Testing (SW Dev Test): SW Dev Tests guarantee software compatibility of transceivers with a wide variety of routers and switches. In addition to functional traffic testing, these tests check for compliance to MSA register specs and alarm settings. Additionally, regression testing is completed on a per release basis to ensure backward software compatibility with each release of system software.
The successful completion of these xDVT tests guarantees that Cisco optics are the most thoroughly vetted for stable link operation. The compatiblity of Cisco transceivers with this wide variety of platforms is continuously updated in the Cisco Transceiver Compatiblity Matrix.
Reliability and Quality
All credible transceiver manufacturers comply to Telcordia testing standards such as GR-468. However, the Cisco qualification process goes beyond them with additional testing such as extended power cycling and accelerated aging requirements. And before transceivers are qualified, Cisco audits factories and manufacturing lines. During these audits, Cisco reviews best practices to identify gaps in transceiver manufacturing processes. And Cisco implements On-Going Reliability Testing (ORT), which is the process by which randomly selected transceivers from the manufacturing line are stress tested. This catches unintended variations in process over time. Both factory audits and ORT allow Cisco to preemptively correct quality issues before transceivers are deployed in the field.
In the rare case that a quality issue is detected during transceiver deployment, Cisco contains it immediately, identifying the root cause and rectifying the source of the problem. For customers who have purchased service support, Cisco’s TAC (Technical Assistance Center) provides assistance while the specific root cause analysis is being completed.
For nearly two decades, Cisco has been developing expertise in optics and best practices around transceiver design and testing methodologies. These investments have resulted in minimal downtime and service disruptions for Cisco customers. Cisco ships several million units each year with a field return rate of less than 100 ppm. That is less than 100 failures for every million units shipped and spells peace of mind for network engineers!