Curious about interoperability, encryption, management and standards?

Here's the lowdown on wireless technology today. By Joel Conover

10. Cost

Thanks to integrated chipsets from Intersil Corp. (formerly Harris Semiconductor), Lucent Technologies and other component manufacturers, the cost of developing and delivering a wireless solution has dropped significantly.

The result is a PC Card solution that is on par with that of wired Ethernet. The products we tested for this article list for $179 to $249 per card? A price that can be easily justified for home office or mobile users. Wireless has become cost effective; for its PowerBook line of notebooks, Apple Computer even has a wireless module that costs just $99, a price we expect most vendors to hit within 12 months.

9. Performance

Wireless performance has nearly quadrupled over solutions based on proprietary or even 2-Mbps products using 802.11. Single-card performance can reach 6 Mbps, which is more than sufficient for the average business user. Much of this is thanks to the 802.11b high-rate standards body, which was driven primarily by Lucent and Harris. The 802.11b standard uses a technology called CCK (complimentary code keying) to encode the wireless data in a format that fits within the 802.11 DSSS (direct-sequence spread-spectrum) FCC rules. CCK is what allows these wireless devices to operate at 11 Mbps. Of course, CCK is not without its trade-offs; 11-Mbps products have significantly shorter range than their 2-Mbps counterparts. Fortunately, most vendors have implemented 802.11b products that drop back to 5.5 Mbps, 2 Mbps and 1 Mbps as range increases. As you inspect our performance charts (see graphics from main story), it is easy to identify where the cards under test dropped down to lower rates to support increased range.

8. Interoperability

The 802.11b high-rate wireless standard is the best thing ever to happen to the industry. In our labs, we found that every one of the products we tested was capable of interoperating with products from competing vendors. The fact that we needed no special engineering support to make any of these products work together tells us that this technology has finally jelled. Rather than 11 vendors delivering 11 wireless products, there is one industry capable of delivering a wireless Ethernet solution.

The efforts of the IEEE and the Wireless Ethernet Compatibility Alliance (WECA) and work being done at the University of New Hampshire are making wireless interoperability a nonissue. The work of these groups guarantees 802.11b will be the future of high-speed wireless. Without these groups and participation of the wireless networking vendors en masse, 802.11b would just be another shot-in-the-dark technology. WECA's WiFi (Wireless Fidelity) branding scheme is your guarantee that the wireless products you buy will be interoperable.

7. DS Reigns Over FH

There is an almost religious war raging between the DS (direct sequence) and FH (frequency hopping) camps. However, because of the technology used in 802.11b, only DS solutions can operate at 11 Mbps. Fortunately, DSSS offers superior range and performance, and today's technology makes it affordable.

DS technology uses a chipping code to spread a signal across a larger chunk of spectrum. A typical 2-Mbps DS system uses 11 chips to spread its signal, resulting in about 22 MHz of spectrum utilized. Note that there's 83.5 MHz of bandwidth, so you can get three clean DS channels out of the 2.4-MHz ISM (industrial, scientific and medical) band.

In contrast, FH transmits across a set of very narrow bands. If one or more of these bands are polluted with noise, the radio simply skips that channel and moves on.

Because the radios spend only a small amount of time in each band and there are plenty of bands (79 bands of 1 MHz in the sequence), FH theoretically can scale better and offers greater immunity from interference in most cases. DS gains a slight performance advantage over FH, because both the channel bandwidth is greater and it doesn't have any latency while switching channels.

Although this latency is very small, it does add up to a performance hit. To its detriment, DS does not scale as well when it comes to the number of nodes in a given area, and there will be more susceptibility to signal interference as you scale a DS network compared with when you scale a FH network.

DS radios once also were very expensive to build, compared with the relatively cheap FH radios. FH radios rely on the host driver software to take care of some MAC (Media Access Control) layer issues. Specifically, a packet cannot be spread across a hopping boundary; thus, the MAC layer must make sure the packet will fit in the remaining time slice of the current hop. If the packet cannot fit in the current hop, then it must wait to transmit until the beginning of the next hop.

In contrast, DS radios must deal only with a high-speed chipping code. The speed necessary to accomplish this chipping was once cost prohibitive; however, today's low-cost ASICs and DSPs (Digital Signal Processors) have made it easy and cost effective to build a DS wireless solution.

The 802.11b high-rate standard uses multiple chipping codes and other special processes to achieve its high-speed rate. Because of these multiple codes, the signal-to-noise tolerance is not as high. The result is decreased range at 11 Mbps. The religious war may be drawing to an end, however, because 802.11b high-rate wireless can be implemented using only a DS radio. So even though FH technology has some merits compared with DS, DS will be the long-term winner.

6. Encryption

Thanks to an accommodating standard and well-written drivers, it is easier than ever to pick up on a wireless Ethernet conversation and listen in as you see fit. In the lab, we were able to sniff packets out of the air with ease.. Not only that, but while wandering outside near the windows of our call center, we could tap into the corporate network.

Encryption shouldn't be an option on your shopping checklist. WEP (Wired Equivalent Privacy) is an option that sits on top of the 802.11b standard, but almost every vendor is supporting it in some fashion. Although we didn't do a complete interoperability matrix, our lab tests show that WEP appears to be just as interoperable as the 802.11b standard itself.

But beware: WEP can be a management nightmare. Because it uses a shared key system, you are responsible for distributing passphrases, hex keys or ASCII strings that represent your wireless-encryption key. If that key is leaked, your data is compromised. The difficulty of changing keys varies from vendor implementation to vendor implementation, and though not complicated, the keys can be hard to disseminate securely.

One vendor, NoWires Needed, has developed a unique public/private 128-bit key exchange algorithm called AirLock. With AirLock, every session is encrypted, and AirLock's operation is transparent to the user.

The AirLock system uses a public key exchange to negotiate an encrypted session; private keys are established once the initial encryption has taken place. The result is a highly secure wireless network with absolutely no need to disseminate keys to end users. If only WEP could work that well.

5. Management

Wireless management and configuration tools are, plainly put, pathetic. In more polite terms, the state of wireless network management leaves much to be desired. Most vendors offer only paltry device-configuration tools, with no concept of enterprise management, monitoring or capacity planning. Of the products we tested, the best of the bunch offered Web-based device management.

This is simply unacceptable. For one thing, it is nearly impossible to push an enterprise change out to multiple access points. Trying to accomplish this task via a Web browser is tedious, even if the Web interface supports the option to configure multiple access points?and few do. Many of the products lack even a Web or serial-based interface. Instead, you must individually configure the devices via a (semi-)proprietary SNMP tool.

These tools don't offer much in the way of capacity management or network monitoring. To successfully manage a wireless network, your NMS (network management system) solution must be aware of mobile devices and users. Only one vendor? Symbol Technologies?has anything that approaches the needs of a true enterprise management solution, and that solution is proprietary to its access points. Vendors need to deliver effective device management before this technology can truly be deployed in the mainstream.

4. The Reseller Market

The wireless market may look big, but the number of players is really pretty small. There's Lucent, which sells to Apple, Enterasys Networks and Intermec Technologies, among others. Likewise, Symbol provides equipment to Intel Corp. and 3Com Corp. There are only a few other top-level providers, such as Zcomax Technologies. Intersil supplies reference cards and chips to a number of vendors, including Compaq Computer Corp., Farallon Communications and Zoom Telephonics, though Intersil doesn't deliver directly to the retail channel. And NoWires Needed delivers access points to Compaq and BreezeCom, as well as delivering its own access points and cards directly to the market. Cisco's Aironet series belongs solely to Cisco, though Dell Computer Corp. and Aironet (prior to being owned by Cisco) had a reseller (OEM) agreement. And in this market of OEMs, the number of radio manufacturers is even smaller.

The radio, of course, is that tiny bit of hardware inside your PC Card that makes this all work. Intersil and Lucent comprise nearly the entire radio market.

3. Wireless Is Not Ethernet

One of the most interesting aspects of wireless Ethernet is that it's really not Ethernet at all. Although 802.11b shares some common aspects of an Ethernet network, in reality, wireless is a CSMA/CA (Carrier Sense Multiple Access/Collision Avoidance) network. The 802.11b standard has a raw bit rate of 11 Mbps. The wireless MAC layer is only 70 percent efficient to begin with. This is primarily because of the CSMA/CA and Random Backoff protocol, which is a necessity in a wireless environment. In a wired network, the adapter can listen to the wire and hear collisions as they happen. But in a wireless environment, the transmitting signal is so locally overpowering that collision detection isn't an option. Thus, CSMA/CA must be used to guarantee that collisions are minimized, and data is delivered reliably.

CSMA/CA utilizes the RTS/CTS (Request To Send/Clear To Send) protocol to notify other workstations that a transmission is about to take place. This four-way handshaking minimizes the number of collisions and makes sure that hidden nodes are aware of transmissions across the entire wireless segment; however, this method introduces significant overhead on the network. The wireless MAC is also significantly more complex than a typical Ethernet MAC. Wireless includes four MAC addresses instead of the two found in an Ethernet header.

Finally, to maintain backward compatibility with 802.11 devices, the 802.11b wireless device transmits the preamble and a portion of the packet header at 1 Mbps. This accounts for a significant amount of additional overhead, as the preamble here is significantly longer than an Ethernet preamble. The overall result is a network that can be at most 70 percent efficient (allowing a maximum data rate of about 7.7 Mbps).

Losses can be eliminated or reduced by improving the strength of the primary signal, reducing the time it takes to discern ghost signals from the true signal and reducing the amount of time it takes to sample diverse antennas. All this adds up to a wireless standard that really isn't Ethernet at all. Vendors also can "push" throughput at the cost of interoperability. In the case of point-to-point devices, eliminating the 1-Mbps 802.11 legacy transmit speed can significantly improve performance. Likewise, if a network is known to be point to point, the Random Backoff algorithm, interframe gap and preamble can all be minimized to maximize throughput. It is important to note that in a point-to-point application, 802.11b compliance is more of a liability than a benefit?the price you pay in throughput is not worth the gains of a standards-based wireless service.

2. Other Wireless Standards

There are at least a six working groups and wireless organizations dedicated to wireless technologies other than 802.11b. For example, HomeRF is building a home networking protocol and standard for all sorts of home-based cordless devices, and the group is petitioning the FCC for rules modifications that will permit high-speed FH using 5-MHz channels. Other groups include Bluetooth, HiperLAN2 and 802.11a. Bluetooth and 802.11b will coexist, even though they share the same spectrum, and there may be the potential for some interference (resulting in lower throughput). Bluetooth is designed as a peripheral interconnect wireless point-to-point protocol. HiperLAN2 is a next-generation technology that will deliver 54-Mbps wireless access in the 5-GHz spectrum.

Like WECA, the HiperLAN2 group will be responsible for guaranteeing interoperability of these devices. The HiperLAN2 Global Forum is the equivalent of WECA for the HiperLAN2 standard. HiperLAN2 is being driven by ETSI BRAN (Broadband Access Radio Network). Interestingly, IEEE 802.11a shares the same physical layer as HiperLAN2 technology. Only the MAC is different. 802.11a is all but complete on paper, thanks to the work of these two groups. The HiperLAN2 standard will take longer to finalize because the working group must design an entirely new MAC protocol. Vendors are fence-sitting, with one foot in each camp, waiting to see which product will receive universal acceptance.

One important thing to consider with 802.11b is that it is an end-of-the-line technology. Upgrading to 5-GHz technology will be much like converting from an Ethernet network to FDDI. Your existing access points may have upgradable radios (removable PC Cards), but chances are that the network interface to the wired LAN won't be able to keep up with the 54-Mbps data rate. That means new access points. Thus, don't buy 802.11b with plans to upgrade to faster 5-GHz networking in the immediate future. That will be a forklift upgrade. But you shouldn't wait for 802.11a either?affordable 802.11a products are at least three years away.

1. New Wireless Products

The wireless industry is responding to the growing market by providing new tools and appliances to take advantage of wireless deployments. In addition to testing the 11 solutions sent for this article, we also had the opportunity to test two wireless accessories designed to work in conjunction with enterprise wireless infrastructures: SpectraLink Corp.'s NetLink Wireless Telephone System and WRQ's NetMotion 1.0 software.

The SpectraLink solution is an 802.11 DS wireless VoIP (voice-over-IP) telephone system. Working in conjunction with your existing wireless infrastructure, the NetLink DS solution provides 802.11 DS telephones that can be used anywhere you have wireless coverage. We tested the NetLink solution in conjunction with the 802.11b products in our review. The NetLink hardware is a VoIP gateway into your wireless network. The solution comprises wireless phones and an Ethernet VoIP gateway. We found that the solution was stable and functional, but it used a significant amount of bandwidth on the wireless network. On average, a single call used almost 150 Kbps when a call was in session. That doesn't make for a highly scalable solution.

While the NetLink solution is a snazzy technology, a maximum of five clients per network is not realistic for an enterprise setting. There is no reason this technology couldn't use 64-Kbps encoding to accomplish the same job. WRQ's NetMotion is designed to deal with the transient nature of wireless connections. Many connection-oriented applications don't deal well with the temporary loss of signal that can occur when a wireless client moves out of range. WRQ's NetMotion attempts to solve that problem by acting as a proxy agent for your wireless communications. TCP sessions are held open by the proxy agent (a dedicated Microsoft Windows NT workstation). When a wireless user roams outside the coverage area, the proxy agent maintains the application session until the user returns to the wireless coverage area.

In our brief tests of NetMotion 1.0, the software seems to live up to its promises, at least for the telnet sessions we attempted through the NetMotion proxy.