Disruptive by Design: Robert Frost and the Form Factor Future

March 1, 2019
By Lindsay Clarke

In his famous poem, “The Road Not Taken,” Robert Frost writes, “Two roads diverged in a yellow wood.” If you have read the poem or analyzed it, as many a high school English teacher has required, you know that Frost suggests taking the road less traveled is the better choice. And while this may be true for adventure seekers and wanderers out there, here in the world of IT I recognize the benefits of not wandering off on my own. The life cycle of network equipment can be five to seven years, or even longer, so on this cusp of 400G it is important to choose optics that offer interoperability for the long term.

In my last column three years ago, I pointed out the need for adherence to open standards. This was primarily focused on nonproprietary protocols and other soft concerns. The benefits of interoperability and the undeniable market forces that drive innovation and reduce costs depend on it. Currently, as we find ourselves at the edge of 400G adoption, we’ve reached a different sort of fork in the road, pondering optical interface types.

Down one path we find QSFP-DD. For those of you who may one day participate in an IT-based trivia game, that stands for quad small form-factor pluggable—double density. You will be pleased to find that it looks much like the QSFP+ and QSFP28 you have come to know and love for 40G and 100G, respectively. In fact, you can plug either in to a QSFP-DD port without need for a mechanical adaptor.

Down the other path is the OSFP, also known as the octal small form-factor pluggable. OSFP looks similar to the QSFP family, although slightly larger in size, and it requires a passive mechanical adaptor if you want to use your QSFP+ or QSFP28.

So which path is right for you? Both are standard and both allow for backward compatibility with QSFP technology. Many vendors are developing to both standards. Pricing is similar, and both will ultimately help you achieve your 400G dreams. But there is a reason that the IT powers that be have given us OSFP. The OSFP has been designed with 800G support in mind. It has thermal capacity and meets the signal integrity requirements for 800G, while QSFP-DD is unlikely to do so.

According to Cisco’s Visual Networking Index, global Internet traffic is expected to triple by 2022 from 2017. The demand for capacity and speed is not diminishing, and it isn’t going away. By planning for 800G now, you will future proof your network without incurring great cost or sacrificing the backward compatibility that is necessary during transitional periods.

Even dear Mr. Frost acknowledges that the path you choose now frequently leads to the paths you choose in the future: “Yet knowing how way leads on to way, I doubted if I should ever come back.” Plan for the next path. With more than 80 companies having joined the OSFP multisource agreement group, you won’t be alone. Industry support is strong, and the OSFP ecosystem is robust enough to carry the transition from 100G to 400G and beyond.

And for those of you still having a tough time making this decision, fear not. The OSFP and QSFP-DD describe the physical form factor of the module. As long as the Ethernet media types are the same, OSFP and QSFP-DD modules will peacefully coexist.

Lindsay Clarke is director of special programs at Arista Networks. The views expressed here are hers alone.

Enjoyed this article? SUBSCRIBE NOW to keep the content flowing.


Share Your Thoughts: