No Jitter is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

Will 802.11ac Kill Wired Ethernet?

The new 802.11ac standard provides bandwidth equal to or surpassing that which is available to most wired users, but abandoning wired networks isn't for everyone just yet.

In both consumer and enterprise networks, wireless is king. According to Nemertes' benchmark research, spending on mobile devices and services will take up about 8% of overall IT budgets in 2015 -- a number that we expect to rapidly climb. More than half of companies are increasing spend on mobile devices, with a mean increase of around 25% a year. The shift to mobile goes beyond phones and tablets: On average about 29% of the typical user population will rely on Wi-Fi as their primary means of corporate connectivity -- a number that has almost doubled in the last 24 months.

Wireless standards continue to rapidly evolve to keep up with this growing demand for Wi-Fi access. The latest iteration of the venerable IEEE standard, 802.11ac Wave 1, offers upwards of 1.6 Gigabits of bandwidth, with Wave 2 upping potential per-user bandwidth north of 6 Gigs later this year. With support for QoS, and spectrum management enhancements that increase WLAN reach and density, the time may finally be right for companies to abandon wired networks all-together, but there's a catch; actually, three primary catches that IT leaders should consider before cutting the cord.

The Uplink: Let's step back for a minute and think about how most wireless networks are designed today. Access points are distributed around offices and other facilities, connecting back over Cat 5e or 6 cabling to an Ethernet switch located in a closet via 1 GigE switch port. Typically, that Ethernet switch powers the AP.

Obviously upping per-user access bandwidth to anywhere from 1 Gbps to 6 Gbps means that the 1 Gbps uplink is likely to quickly become saturated. While it's unlikely that end-user devices will use the full available bandwidth, even a dozen or two devices demanding 200-300 Mbps will strain the 1 Gbps uplink. So the answer is simple -- just upgrade the 1 GigE uplink to 10 GigE, right? Not so fast.

Going from 1 GigE to 10 GigE introduces some additional complexity and cost. For starters, 10 GigE switch ports are typically more expensive than 1 GigE and upgrading large numbers of ports will require significant investment in not only Ethernet switch upgrades in the closet, but upgrades from distributed switches to LAN core.

10 GigE also requires Category 6e or 7 cabling for runs of up to 100 meters, which means that APs may not be close enough to closet switches to support 10 GigE and potentially need to switch to fiber (which supports runs of up to 260 meters using Multimode). Finally, 10 GigE doesn't support Power over Ethernet (PoE), so if you are currently powering your APs via Ethernet, you'll need another approach (it's worth noting that the IEEE is now looking at developing a standard for PoE over 10 GigE, but it's not available yet).

But wait, all hope is not lost.

Recognizing the impracticality of upgrading AP uplinks to 10 GigE, WLAN and LAN vendors are offering another approach; enabling customers to upgrade their uplinks to 2.5 Gbps or 5 Gbps while maintaining support for PoE. But as there's a catch here as well: Absent a ratified IEEE standard for 2.5 or 5 GigE, switch, AP, and component vendors have created their own consortiums to partner to develop competing, non-interoperable approach. One side there's Cisco and the NBASE-T Alliance, while on the other side there's Alcatel-Lucent, Aruba, Avaya, Brocade, and Extreme Networks with the MGBASE-T Alliance. At some point each of these efforts will roll up into the IEEE's own standards development group, but that likely won't happen for a year or more.

The Network Backbone: So if you've solved the uplink problem by going to 2.5, 5, or 10 GigE over copper or fiber, you've still got the cascading effect of adding more access bandwidth. When switch uplinks are typically no more than 10 Gbps, or sometimes bonded 1 Gbps links, those uplinks will require bandwidth upgrades, likely to 40 Gbps, with a corresponding need to invest in switch and/or blade upgrades, wiring upgrades, etc.

PoE For Phones: Even in a fully wireless office it's likely that desktop phones will continue to exist (not everyone is moving to softphones and headsets). The standard way of powering Ethernet phones is via PoE. In a wireless-only environment, those phones will require another source of power -- typically a power adapter attached to each phone that plugs into the nearest outlet. Going this route means that IT architects will need to think about power reliability for phones, especially in locations with poor cellular coverage. This means the potential cost of not only a power pack for each phone, but also a UPS, or a building-wide UPS to keep phones running in the event of a loss of building power.

So back to our question: Will 802.11ac kill wired Ethernet? In many small offices and open workspaces the answer is clearly yes. Eventually, the IEEE will standardize a 2.5 and 5 GigE approach that will offer vendor interoperability. IP phones are rapidly becoming software running on a laptop, PC, or mobile device. Network backbones will scale as they always do to meet increasing bandwidth demand at the edge. But the road to the wireless-only office will have some bumps along the way. Make sure you understand your vendor's approaches to scaling not just your APs, but your backbone Ethernet network to support 802.11ac.

Follow Irwin Lazar on Twitter and Google+!
@imlazar
Irwin Lazar on Google+