x
Optical/IP

Cisco's Speculative Spec

Cisco Systems Inc. (Nasdaq: CSCO) is proud to let its engineers explore new ideas – but that doesn't mean the company plans to use them.

Although Cisco engineer Glenn Zorn's name is on an Internet Engineering Task Force (IETF) memo promoting "dumb" 802.11 access points, no one should expect that the company will turn these ideas into actual hardware, a spokesperson says.

In other words, it could just be an exercise in intellectual freedom. [Ed. note: a.k.a. 'nut-scratching'?] This would not be unusual. In the whiteboard jungle that makes up the world of vendor R&D labs, engineers with large foreheads (five-heads, really) frequently vie to outdo each other in publication of IETF drafts. Generally this is not a pointless exercise, as the dialectic produced by different specs often really does result in stronger standards.

Zorn (who, disappointingly, does not actually hail from the planet Tharg) is one of the authors of the IETF paper describing the methodology for a lightweight access protocol for "thin" (a.k.a. "dumb") APs – units from which most of the intelligence has been stripped, making them (theoretically) cheaper and easier to install. These boxes would be used in conjunction with a dedicated wireless switch, or "brain," which would provide most of the network smarts (see Access Point Tiff Simmers). However, the involvement of Zorn [ed. note: take me to your leader!] doesn't mean Cisco itself will necessarily follow that path, according to Ron Seide, product line manager for wireless LANs at the company.

"Cisco has long supported the idea of intellectual freedom for its engineers," says Seide. "But there's a big difference between exploring ways of solving a problem and actually developing a product."

There has been plenty of speculation over Cisco's approach to the wireless LAN switch market (see especially Cisco’s LAN Switch: Build or Buy?), but so far the company has shown little inclination to concentrate intelligence at the center of the network. Instead, it claims that the benefits "thin" APs are designed to deliver – including simpler interoperability and tighter security – can be accomplished using their current model: "We've been delivering quite a lot of those features already," says Seide.

— Dan Jones, Senior Editor, Unstrung

Page 1 / 2   >   >>
airbb 12/5/2012 | 12:06:57 AM
re: Cisco's Speculative Spec As we've seen with PC vs. NC, it really doesn't matter when AP price drops at the current rate.

We'll have the "smart" AP at $100 or two. Customers will still wanted to buy something other than IBM PC.

The wireless switch will be off the shelf component managing multi-vendor APs, smart or not.
dmitton 12/5/2012 | 12:05:58 AM
re: Cisco's Speculative Spec The link in the article is obsolete and doesn't work. The current draft is -01 and I don't see a Cisco name in it. All Airespace and DoMoCo!

- Dave.
jacksullivan66 12/5/2012 | 12:05:50 AM
re: Cisco's Speculative Spec So within a week of the publication of the "thin AP" spec by the IETF, a second draft comes out with Cisco's name no longer on it?

I guess they "really" don't like the idea of the thin AP. Maybe it's because it might mean they don't get to sell as much equipment, or get the margins they'll need on those dedicated AP boxes...

The question is, will the market force their hand? Another article by Unstrung entitled "PCTEL Goes Soft", details the trend toward converting regular PCs and laptops into Access Points.

Let's face it - the smart AP as a dedicated, IT managed box is evolving to a thin, dyanamic, possibly all software solution running on just about any computing device within the enterprise, with a WLAN switch/router sitting on the backbone providing the intelligence. That can't be good for Cisco - the current leader in enterprise WLAN / Access Point equipment...
joset01 12/5/2012 | 12:05:40 AM
re: Cisco's Speculative Spec Indeed, I'm interested in how everyone sees WLAN infrastructure evolving over time.

Here's my take on it at the moment (probably subject to change as I learn more):

In the enterprise, especially in larger deployments, we'll see an increasingly thin architecture evolve, possibly even with software radio implementations on the front-end and the "smarts" on the backend. I wonder how much of this will actually be a hardware market in the future (5-10 yrs), perhaps in the end the only hardware you'll need is tiny little radio nodes on the backend and a single RF chip at the client.

I imagine that the access point as we know it now is likely to find a fuller evolution in the public access hotspot market (if that does turn out to be successful) with software handling the management features and probably a lot of antenna technology on the front end.

Of course, this view doesn't really take into account emerging tech like UWB or other specs and standards or the fact that the technology that seems most suitable for a deployment doesn't always get picked (anyone remember the whole network computer versus PC debate?).

Anyway, be interested to know what you think. Maybe we should have a poll about this topic?

DJ Unstrung
airbb 12/5/2012 | 12:05:37 AM
re: Cisco's Speculative Spec Cisco makes much much much ... more money in switch/router than WLAN AP. Switch/router
was, is and will be their business. This is why they will jump into WLAN switch market once the market accepts the product.

Centralization of "smart" is the fundamental of Cisco's business. Cisco picks up companies like Aironet just to test the market and can drop it as easy. Aironet is not Cisco. Cisco is not about to re-invent the wheel on how to do network business because of Aironet.

Wait and see...

jacksullivan66 12/5/2012 | 12:05:36 AM
re: Cisco's Speculative Spec Dan - thanks for running the "Zorn is Shorn" story... Maybe now we can get some response from the Cisco folks on the message boards. Pulling their support of the IETF draft spec a week after it was submitted is a pretty strong statement "against" the thin AP movement. I'd love to hear somebody from Cisco answer the question, "Why?". If you've read my previous statements, I think we've already guessed at their reasons.

With regard to the evolution of the WLAN within the enterprise, I've also chimed in suggesting that the dedicated WLAN AP equipment business will be a non-existent business within 5 years (within the enterprise). The WLAN intelligence will be pushed (1) back to the network core and (2) embedded in silicon - the latter taking the shape of PCs, laptops, and any other computing device within the enterprise. Multi-hop architecture may also play here - but it's yet to been seen if the enterprise IT Manager will be willing to adopt this seemingly additional complexity.

I'll use this forum to take my stance one step further: once the evolution above occurs, look for the UWB WPAN to gain traction within the enterprise. If the PC will take over most of the WLAN AP functionality, why not also use it as the central hub of a each users personal area network - best enabled by UWB.

The trick will be to seamlessly integrate these UWB WPANs into the legacy enterprise infrastructure - so they become a seamless extension of the more ubiquitous 802.11 based WLAN (seamless roaming, QoS, etc.). And with the "thin AP" protocols being investigated today, expect the UWB WPAN devices to be manageable over the traditional switching/routing technologies that will soon be managing the rest of the enterprise network - with no prejudice between wired and wireless devices.
jacksullivan66 12/5/2012 | 12:05:28 AM
re: Cisco's Speculative Spec I agree with you, airbb - but I think you miss my point. Right now, Cisco is still selling switches/routers. Was/is/will be... But that's an incremental growth business - not a relatively new and relatively lucrative equipment segment like current, "smart" enterprise class WLAN APs.

My take is that Cisco sees the new WLAN switch / router equipment crowd as simply enabling "additional functionality" that they can add to their current, wired switch / routers. So as this new WLAN switch / router market matures (and it will very quickly), all that new functionality will simply be absorbed into current equipment that Cisco already sells.

So no new equipment segment there. Just incremental growth to an already existing product line.

The "new equipment segment" Cisco would like to milk for revenue is the smart AP market (worth roughly $400M to them in 2002, with high double digit growth projected thru 2003 and beyond). But it's become clear that the "thin AP" movement threatens this potentially lucrative new business line. Maybe not today. Probably not tomorrow. But most likely several years out.
airbb 12/5/2012 | 12:05:26 AM
re: Cisco's Speculative Spec Error in the item 3:

from
"MIS people will never replace wired switch with wireless switch."

to
"MIS people won't likely to replace wired switch with wireless switch or unified switch."
airbb 12/5/2012 | 12:05:26 AM
re: Cisco's Speculative Spec You brought 3 interesting items to discuss here:

1. Cheap thin AP or Expensive thick/smart AP?
As was for PC vs NC, the final answer will more likely to be cheap smart AP. There isn't much of magic in building AP. We'll see many off shore vendors beef'ing up their AP at very low price(watch Taiwan vendors)

2. Is $400M/year lucrative?
Cisco AP volume picked up 2x last year, but revenue was up 30%(see the article in Unstrung). In order to achieve high double digit growth, the volume will need to pick up even more(I do believe this will happen though). This means that AP is becoming commoditized very quickly. For the same reason, we'll get cheap smart AP very quickly(item 1 above).

3. Unified switch for both wired and wireless LAN?
This is a bit more difficult to prove, but remember everyone doesn't drive convertable car.
Let say Cisco somehow figured it out technically, what will happen to their existing wired switch product line? Also, customers then have to pay additional $ to buy beef'ed up switch when some may have use for it while others don't. MIS people will never replace wired switch with wireless switch. The same legacy issue is keeping 10Mbps Ethernet line still a significant part of network infrastructure. You see, nobody in the industry wants unified switch. Many will continue to buy regular wired switch at low price while some will buy wireless switch as need arises and only few will pay extra $ to buy one solution fits all unified switch.

Network infrastructure is like Lego block, not Play-Doh.

What do you think?
Steve Saunders 12/5/2012 | 12:05:21 AM
re: Cisco's Speculative Spec "It appears I'm not the only one who's projecting (and excited about!) a combination of the wired / wireless switch"

Why is this a good idea?
Page 1 / 2   >   >>
HOME
Sign In
SEARCH
CLOSE
MORE
CLOSE