In 2008, the IEEE ratified the 40GBASE-SR4 standard that uses 4 lanes at 10 Gig per lane over multimode fiber for a total of 8 fibers (i.e., 4 transmit and 4 receive). At the same time, the 100GBASE-SR10 standard was ratified that uses 10 lanes at 10 Gig per lane over multimode fiber for a total of 20 fibers (i.e., 10 transmit and 10 receive).
While very few data centers are running 100 Gig today, and most that are tend to use singlemode fiber, we all know that 100 Gig over multimode is on the horizon. Therefore, many data center managers are looking to prepare themselves for an eventual migration.
With active equipment planning to use a single 24-fiber MPO interface for 100 Gig and the channel currently requiring 20 fibers, many are considering the use of 24-fiber MPO solutions. For 40 Gig, some manufacturers are also promoting the use of 24-fiber MPO trunks that break out into three 12-fiber MPOs at patch panels, claiming that it offers an easier migration path while reducing the amount of fiber in the pathway.
But it’s important to understand why “caveat emptor” applies when it comes to 24-fiber MPO solutions.
As we look towards next generation fiber transmission and future technologies like multicore fiber, we see that fiber counts are going down—not up. As such, IEEE is currently working on the 100GBASE-SR4 standard that intends to reduce the fiber count of 20 for 100GBASE-SR10 to 8 using 4 lanes at 25 Gig per lane (i.e., 4 transmit and 4 receive).
This would not only reduce the fiber count and result in less cabling for 100 Gig, but it would also enable an easier migration from 40 Gig to 100 Gig as both applications would use just 8 fibers. It would also likely be the end of life for 100GBASE-SR10.
At this time, it’s not a matter of “if” the 100GBASE-SR4 standard will come to fruition, but rather “when.” And when we can run 100 Gig over just 8 fibers, we’re back to using 12-fiber MPO solutions—precisely why I do not recommend the use of 24-fiber MPO solutions today. But it’s more than just what’s on the horizon for standards.
Optical loss budget is a big concern among data center managers, and due to limitations in the protocol, standards now require a total connector loss budget of 1.0 dB for 40 and 100 Gigabit Ethernet channels. A 24-fiber MPO connector typically has a loss of 0.5 dB versus the 0.2 dB loss that Belden offers on its low-loss 12-fiber MPO connectors. This is mainly due to the fact that the more fibers in a connector, the higher the loss, and the 24-fiber MPO connector and its end-face geometry are more difficult to manufacture and control than a 12-fiber MPO.
The higher loss of the 24-fiber MPO limits data center managers to having just two mated pairs in a channel. As discussed in my previous blog, a two-point topology results in high-density patching at the core, longer point-to-point connections and unsecure access to critical switch ports. In other words, 24-fiber MPO connectors do not allow for the use of flexible cross-connects or zone distribution areas that improve security, scalability and manageability. Furthermore, there is currently no test equipment available for field testing 24-fiber MPOs.
While a 24-fiber MPO solution might work for a simple, point-to-point architecture, the reality is that the technology is not all it’s cracked up to be. A 12-fiber MPO solution today will easily support 10, 40 and eventually 100 Gig with better loss, more flexibility and industry-wide acceptance.
Be sure to read our follow-up post on this topic, as well!
Dwayne Crawford has more than 20 years of experience in the datacomm industry. He has served on several international standards committees to advance high-performance/low-latency protocols (such as IEEE-1394, GigE Vision and CameraLink) used in real-time image processing and utilizing high-performance computing platforms.