Product Focus

BridgeWave Gigabit Wireless Bridges

In past issues of the column, we have focused on bridges operating in the 2.4 and 5 GHz band, using 802.11 technologies. These bridges wring the most out of802.11, but even with 802.11n, there is a limit to the capacity they can provide. BridgeWave’s gigabit wireless bridges operate in the 60 GHz and 80 GHz frequency bands using custom modulation methods to provide gigabit speeds for a true alternative to fiber.

Unlike 802.11-based bridges, where the link speed is half-duplex, and so is split between the upstream and downstream side of the link, BridgeWave’s solutions are true full-duplex, so the aggregate speed of the link is 2 Gbps (1 Gbps up and 1 Gbps down).

The 60 GHz band is totally unlicensed and can provide gigabit speeds up to about 3/4 mile, or 1.5 miles with a high-gain antenna. The 80 GHz band requires “light” licensing (much cheaper and faster than full FCC licensing) and provides gigabit speeds up to about 4-7 miles.

Fiber is not just about throughput, though. It’s about reliability. BridgeWave’s links can be engineered to the desired level of availability: up to 99.999%. BridgeWave’s 28-year Mean Time Between Failures means that the equipment will probably outlive the usefulness of the link (ten-gig wireless, anybody?). Finally, all BridgeWave radios undergo HALT/HAAS testing before they are shipped. This means that they are put into an environmental chamber and cycled through temperature extremes and vibration all while powered on and maintaining a link. This means that any units with manufacturing defects fail in the factory, not in the field.

Essential Wi-Fi

802.11’s Network Access Method: CSMA/CA

When two 802.11 devices transmit data at the same time, on the same channel, within range of each other, their data is likely to be corrupted. In this issue, you’ll learn about Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA), an algorithm in 802.11 that compensates for this occurrence.

As its name suggests, the goal of CSMA/CA is to avoid collisions—that is, occurrences when two devices on the same channel, within range of each other, transmit at the same time. CSMA/CA in 802.11 is often compared to its Ethernet counterpart, CSMA/CD, which uses collision detection instead of collision avoidance. All other things being equal, collision detection is a better strategy, because it detects the actual occurrence of a collision and takes action. Collision avoidance does its best to make sure that collisions don’t happen, but it can’t directly detect that one has occurred. So why doesn’t 802.11 use collision detection? Because in a wireless network, it is impossible to directly detect that a collision has occurred. Wireless signals lose strength rapidly as they propagate through the environment. There’s no way for a station to detect the very-weak signals of other stations everywhere in its coverage area, so there’s no way for a station to know for sure whether its transmission is causing a collision. Collision avoidance is the best 802.11 can do.

CSMA/CA starts with “Carrier Sense.” This means that, before a station transmits, it tries to detect whether another station is already transmitting. If a transmission is already occurring, the station enters a mode called deferring. The station continues to defer until the original transmission completes.

802.11 requires a certain minimum period of silence between each transmission, known as the inter-frame spacing (IFS). Once the air has gone silent, the station waits the IFS.

Now the “Collision Avoidance” comes into play. Imagine that the network was silent. What are the chances that two stations, purely through chance, would start to transmit at the same time, and have a collision? The chances are pretty small. Now, imagine the scenario where one station is sending a long transmission. During that transmission, any station that wants to transmit will enter the deferring state. This means that, at the end of a transmission, there are likely to be many stations deferring. If all of those stations were to wait the IFS and then immediately transmit, a collision would be guaranteed.

Instead, what 802.11 does is this: if a station is deferring, then after the air goes silent and after the IFS has passed, the station chooses a random time period, known as the backoff timer, and continues to wait at least that long. This means that if there are multiple stations deferring, they will all pick different random times. One of them will get to go first and the others will go back to deferring. The collision will have been avoided. Of course, there is still the possibility that some other station, that was not deferring, will spontaneously transmit and create a collision, but as previously mentioned, this is relatively unlikely.

Once a station chooses a backoff timer, it continues to count down that same backoff timer, even if it goes back to deferring because some other station’s backoff timer has expired. This means that, eventually, every station will get a chance to transmit its data. From this perspective, 802.11 provides fair access to the medium.

 Technology and Engineering

Point Coordination Function

In “Essential Wi-Fi,” we discussed 802.11’s Distributed Coordination Function (DCF) and how it provides fair access to the medium while reducing the probability of collisions. But DCF is not the end of the story. In this article, we’ll discuss the other side of the coin, Point Coordination Function (PCF).

Although DCF provides fair access to the medium, a station can be denied access to the medium in the short term. For example, imagine that you have a packet to transmit. You detect that another station is currently transmitting, so you defer. After the air goes silent, you wait the IFS and then you choose a backoff timer. So do several other stations who were also deferring. Let’s say that, just by chance, you chose a particularly long backoff timer. You’ll end up waiting while all of the other stations transmit their packets. In the long term, all stations will choose the same average backoff timer, because they’re all using the same random algorithm. But in the short term, a station that chooses a long backoff timer will end up waiting for a long time, at least in computer terms.

For many types of application, this is not a problem. When we say, “waiting for a long time,” we’re probably not even talking about seconds. We’re probably talking about tens or hundreds of milliseconds, depending on the length of the packets that are being transmitted. When doing a file transfer or downloading a web page, it doesn’t really matter if a few packets go very quickly and a few packets go a little more slowly, as long as the overall transfer is sufficiently fast. But for applications like Voice Over IP, there’s an expectation that all packets arrive with approximately the same spacing. There’s even a name for the degree to which the timing between packets varies: jitter. And if jitter is too high, call quality suffers.

The Point Coordination Function (PCF) is 802.11’s answer for applications that need regular, guaranteed, priority access to the medium. The way it works is this.

Every so many beacon intervals, the AP will initiate the contention-free period (CFP). It does this by sending a special beacon frame that causes all of the stations in range to go silent. Any station that attempts to transmit a frame during this period will detect that the network is busy and will enter the deferring state. Next, the AP sends CF-Poll frames to stations which have indicated that they want to participate in point-coordination. When a station receives a CF-Poll frame, it may send a single data frame. Stations indicate whether they want to participate in point-coordination when they associate with the AP. When all stations have been polled, or when the beacon interval is over, the AP sends a CF-End frame, which indicates that the contention-free period is over. All stations resume normal DCF operation, using CSMA/CA to determine which one gets to transmit. This process repeats itself periodically, every so many beacon intervals.

To Infinity... and Beyond!

Wi-Fi, You and the Law

Arguments for encrypting your Wi-Fi network abound, but here’s a new one: a man in Buffalo, NY, got a visit from Immigrations and Customs Enforcement after his IP address was used to download illegal materials. It was later determined that the homeowner was innocent, but that was probably cold comfort after having his door kicked in, being thrown on the ground, and having guns pointed at him. Sharing your Wi-Fi may be a neighborly thing to do, but leaving the network unprotected opens the door to all sorts of activities for which law enforcement may try to hold you responsible.

So long, and thank's for the fish!