I'm going to say something now and you may laugh at me, but here it is: This is going to be the year of Ethernet.
Yes, Ethernet -- the one that was developed by those two smart people at Xerox more than 40 years ago and that gets 15 minutes of fame in the press every few years when a faster standard comes out. What could possibly be so earth-shattering about Ethernet, of all things, that I'd put it up against its more glamorous cousins, like software-defined networking and mobility, as one of the big networking trends to pay attention to in 2015?
For the first time, Ethernet wonks aren't trying to make networks faster.
Think about it. Five years ago, the IEEE ratified standards for 40 and 100 Gigabit Ethernet (GbE). Three years later, the standards body formed a study group to address 400 GbE. But last summer, things changed. A vendor-led consortium popped up to say it was taking a step backward, in a sense, and working on specifications for 25 and 50 GbE. The IEEE launched its own effort to work alongside vendors on an open standard. Why? Because cloud providers and large-scale data center operators had outgrown 10 GbE in their top-of-rack switches, but many 40 GbE-based products were neither cost-effective nor power-efficient.
A few months later, another consortium -- two, in fact -- put out press releases saying they were working on new Ethernet specifications for 2.5 and 5 GbE, a move aimed at enabling enterprises to get the most of out 802.11ac without requiring them to make a costly move to 10 GbE in the access layer.
The goal of these initiatives is not necessarily to develop faster Ethernet standards but more strategic ones, according to John D'Ambrosia, chairman of the Ethernet Alliance, speaking during an interview for our cover story on 2.5 and 5 GbE in this edition of Network Evolution ("Could vendor divide on new Ethernet speeds hurt enterprise networks?").
"Look, I have spent 15 years chasing this speed dragon -- always trying to go faster," said D'Ambrosia, who has also chaired several 802.3 task forces in the IEEE. "But the simple reality is that not everybody needs to go fast, and not every application needs to go fast."
These changes have breathed new life into Ethernet, according to D'Ambrosia, who said the number of Ethernet-related projects in the IEEE is nearly double the level it's been for the past several years.
"I've been in the IEEE since the early days of 10 Gigabit, and I've never seen us as busy as we are now," he said. "You just see people saying, ‘OK, now what other problems can we solve?'"
And as the demands on networks continue to grow, it seems safe to say we can expect to see more innovation in Ethernet that focuses on specific use cases instead of simply adding raw speed.
Also in this issue, contributing writer Steve Zurier takes a look at the network's role in identifying, stopping and, hopefully, preventing large-scale cyberattacks like those executed recently on Sony and others ("Latest network security threats prompt one question: ‘Are we next?''"). And be sure to check out contributing writer David Geer's piece on a group of researchers hoping to eliminate the need for IP addresses with a project called Named Data Networking ("Named Data Networking project wants to retire TCP/IP").
As always, don't miss this edition of "The Subnet," in which we catch up with one wireless networking engineer who shares her tips for building your own inexpensive wireless site survey kit ("Why you might not need an expensive wireless site survey kit").
10 Gigabit Ethernet tutorial: What you need to know
802.11ac Wave 2 likely to prompt wired network upgrades
Should you upgrade to 802.11ac now or wait for Wave 2?
- Containers: Where Linux Host Networking Meets Network Infrastructure –Cumulus
- David's Bridal Modernizes Network Infrastructure –CradlePoint
- Scale Out Networking for Hyper-Converged Infrastructures –Big Switch Networks
- The Maturity Model for Network & Infrastructure Management –SevOne