Last month, we explored three of the major factors that affect whether an 802.11 station can receive a signal: the signal's initial transmit power, the distance between the transmitter and the receiver, and the data rate of the signal. As promised, this month, we'll take a more in-depth look at these factors and some other factors that affect signal range.
Wi-Fi radio technology connects devices together without wires. The “Wi-Fi” name is derived from the Wireless Fidelity Alliance, a global, non-profit industry association that promotes the growth of wireless local area networks (WLANs). The Wi-Fi Alliance tests and certifies interoperability of products based on the IEEE 802.11 specification. There are two ways the devices in a WLAN can be related. In ad hoc mode the computers in the WLAN act as peers. In infrastructure mode a central radio, called an access point, synchronizes the behavior of the devices that are associated with it. The access point also serves as a portal to a wired Ethernet network whereby associated users can access the Internet, printers, file servers, and other wired network resources.
Multiple access points can be interconnected to form a service set, providing a contiguous area of connectivity coverage for the user community. It is the service set identifier (SSID), that appears in the Windows’ “View available wireless networks” dialog box. Selecting the right equipment to implement a WLAN involves prioritizing a number of equipment features based on the requirements of the user community. The inability of a piece of equipment to meet any one of a number of mandatory requirements will render it unacceptable for use. Determining that a particular operational requirement is mandatory, and identifying the level of capability at which a device meets the specified requirement is done in the context of each unique network’s purpose.
In general, there are four main areas of consideration when determining wireless network system requirements. These are:
Radio Signal Propagation
System Capacity
User Authentication
Data Encryption
Radio signal propagation is an area of consideration that has no options. That is, either the radio equipment provides proper connectivity to the user community in the desired coverage area, or it doesn't. This is unlike capacity, authentication, and encryption which allow for a number of alternative implementations and trade-offs to meet the desired design goals.
For example, consider a requirement to provide sufficient capacity to support 100 users in a given area. There are a number of trade-offs that can be considered. If all the users are assumed to be active at the same time, and if each user is allocated 1 Mbps of bandwidth then the network is going to have to support 100 Mbps. If each user is allocated 256 Kbps (similar to a consumer DSL line) then the overall network needs only to support an aggregate of 25 Mbps. If it’s assumed that only 10 users are active at any given time then the aggregate requirement drops to 2.5 Mbps (a reasonable requirement in an 802.11b network).
Users may have strict authentication requirements. They may be required to use biometric authentication like fingerprint readers. They may be asked to enter specialized ID numbers (as with the constantly changing time stamp numbers displayed on a SecureID card carried by the user). On the other hand, the authentication requirements may be more limited. In any case, there are always multiple layers of authentication that can be implemented, again providing trade-offs in the design process. For example, an open Wi-Fi network may allow anyone to attach, but then an access controller may redirect the user to an authentication server for login. There might be no 802.11 authentication, but the user could be required to use a Virtual Private Network (VPN) connection, end-to-end, between their wireless computer and a corporate firewall. There’s no single “right” answer for authentication.
Data encryption, like authentication, may be delegated to any number of different functional parts of the communication system. Wi-Fi’s WEP, WPA, or 802.11i encryption may be used. Wi-Fi encryption can result in reduced packet throughput in a Wi-Fi device and may be turned off. In this case, privacy can be relegated to the use of VPN tunneling, Secure Sockets (SSL), and web page encryption through HTTPS. The point is that a number of choices are available for assuring data privacy and there’s no single “right” answer.
Capacity, authentication, and encryption all present alternative methods of implementation and each has strengths and weaknesses. Each alternative may have an impact on network performance and cost. Nonetheless, it should be evident that these three areas of consideration present various options that must be quantified and assessed. It will also become evident that the radio signal coverage issues are, to a large degree, absolute. A particular area either has the specified level of signal coverage and throughput, or it doesn't. If the radios aren’t doing their job there’s nothing on the back-end that can make up the difference, the way there are multiple levels of authentication or encryption.
Any trade-offs or compromises related to the radio equipment must always meet the requirement that the user community has the proper connectivity in the specified coverage area. It is in this perspective that we begin discussing some of the radio equipment considerations.
Last month, we said that as signal power increased, range increased, all other things being equal. When considering signal power, it's important to keep in mind the directionality of the antenna. An omnidirectional antenna, such as is typically found on access points and PCMCIA cards, transmits its energy equally in all directions. With an omnidirectional antenna, the range will be the same in all directions, assuming that obstructions (or the lack thereof) are the same in all directions. In reality, there will probably be different obstructions in each direction, so range in each direction will be different. For example, there might be a concrete wall behind the AP and open air in front of the AP. Range behind the AP will be reduced because the concrete wall will absorb more of the signal than the open air.
When range is needed in one specific direction, such as with a point-to-point link between two buildings, a directional antenna is typically used. A directional antenna puts more of the signal's energy in one direction than another, allowing greater range in one direction at the expense of decreased range in another. Depending on the area that needs coverage, this tradeoff may be worth it.
Examples of types of omnidirectional antennas include: dipole and slotted waveguide. Examples of types of directional antennas include: yagi, waveguide, patch, panel, parabolic dish, and sector.
Last month, we also discussed receive sensitivity. We stated that the better a station's receive sensitivity, the weaker signal the station could receive, and the more range the station would have. Connect802 has performed a survey of the receive sensitivity of many models of access point and bridge on the market, and has concluded that receive sensitivity is one of the major factors that differentiates low-end, consumer-grade access points from high-end, enterprise-grade access points. The more expensive an access point is, the more likely it is to have good receive sensitivity. This might sound fairly obvious, but other qualities, such as transmit power, don't follow this pattern.
In fact, at Connect802, we believe that having an AP with good receive sensitivity is more important than having one with high transmit power in terms of increasing range. Assuming a two-way radio link is desired (as it always is in WLAN transmission), the range of a link is limited by the shortest range of the two transmitters. Therefore, if we had an AP with a high transmit power, it might be able to transmit packets to a PCMCIA card several miles away, but the PCMICA card's weak signal probably couldn't push a signal more than a few hundred feet. The effective range of the link is only a few hundred feet. One option might be to increase the transmit power of the PCMCIA cards, but this solution is undesirable for two reasons: first, you have to increase the power of ALL of your cards, which is expensive, and second, increasing the power of a device might require external amplifiers, etc..., which defeats the purpose of having a portable wireless station. On the other hand, if we have an AP with a very good receive sensitivity and only a moderate transmit power, the AP would be able to BOTH push a strong signal out to the PCMCIA card AND receive the PCMCIA card's weak response. This would increase the effective range of the link.
Antenna coverage patterns are denoted by azimuth and elevation charts. The azimuth chart shows a top-down view of an antenna's coverage, while the elevation chart shows a side view of the antenna's coverage. An azimuth and elevation chart will allow you to choose an antenna that maximizes coverage where you need it and minimizes coverage where you don't. This month, we'll learn to read an azimuth and elevation chart.
A manufacturer creates an azimuth chart by placing the antenna in an RF-shielded room and then transmitting a signal through the antenna. Using a spectrum analyzer, the signal strength in milliwatts is measured in a circle around the antenna. If the antenna is omni-directional, such as a dipole, then the signal strength will be roughly the same at all points on the circle; if the antenna is directional, such as a yagi, the signal strength will be higher in front of the antenna than behind or beside the antenna.
The outer ring of the azimuth chart represents the strongest signal strength that was measured, and is typically marked “0 dB”. Inner rings represent a signal strength of some number of dB below the strongest signal strength. Note that the azimuth chart does not take the distance from the antenna into account. The chart shows the signal strength of each location relative to each other location, not relative to some absolute distance from the antenna or some absolute power level.
Reading an antenna’s azimuth and elevation chart is relatively simple. First, determine which direction the antenna was oriented when it was tested. Common practice is to point the antenna’s element at the "zero degrees" (straight up) mark on the azimuth chart and at the "90 degrees" (directly to the right) mark on the elevation chart. Some antennas will have an alignment mark on their base to allow precision alignment when installing. To determine the relative signal strength at some number of degrees of rotation around the antenna, find the radius on the azimuth chart that represents that number of degrees of rotation. The point circumference at which the line on the chart intersects that radius represents the number of dB below maximum that the antenna will put out in that direction.
The elevation chart is created and read exactly the same as the azimuth chart, except that it represents coverage in a circle in the plane that is above, to the front, below, and behind the antenna (a side-view circle) instead of in a circle in the plane that is in front, to the right, behind, and to the left of the antenna (a top-down circle).
The Connect802 Report on Networld+Interop 2005, Las Vegas
The Mandalay Bay Convention Center in Las Vegas was
home to the 2005 Interop show. Dave Kornreich (above)
and Joe Bardwell were in attendance.
The Connect802 team was in attendance at the 2005 Networld+Interop trade show in Las Vegas last month. The biggest surprise was the overall lack of big surprises. Past years have often provided a showcase of technological advances but 2005 seemed to be a parade of improvements on the status-quo.
Unlike 2004, Connect802 was not in a booth at Interop for 2005. (Yes, they appear to have changed the name to simply "Interop"). The venue moved from the Las Vegas Hilton (where the show has been for the past decade) to the much smaller Mandalay Bay Convention Center. Frankly, the Connect802 team was not left speechless by anything they saw. Lot's of improvements on existing technology, and a greater emphasis on Wi-Fi networking than last year; but nothing to rave about.
Some of the exhibiting vendors that we spoke to said the crowd was more of the "decision maker" attendee as opposed to the "kick the tires" casual show-goer. From what we saw it looked like many times it was vendors talking to vendors.
It was interesting that the WiMAX kiosk at the Intel booth was constantly occupied with multiple customers asking the Intel engineers about their new WiMAX hardware. Redline Communications (one of Connect802's WiMAX vendors) was showing their latest 802.16 gear along with their upgraded point-to-point microwave equipment.
There was less RFID presence than we would have expected. We've seen several RFID companies showing around the country at various trade show venues, but there was no great splash of RFID at Interop this year. Our opinion is that the RFID providers are focusing on facilities management, supply chain management, and database exposure and are not getting traction with the legacy IT folks who are the core attendees at a show like Interop.
The Best In Show awards included Siemens Communications with their VoIP OpenScape Telephony Control Link and Meru Networks with their Radio Switch. The Meru offering increases the number of players in the wireless LAN switch marketplace (along with Symbol and Trapeze, used by Connect802). Wireless LAN switching provides some unique advantages for VoIP implementations and we expect to see the growth of the wireless LAN switch technology marketplace follow the growth of wireless VoIP, particularly as 3G cellular integrates more closely with in-building wireless VoIP in the Wi-Fi arena.
FireTide introduced the new 3000-series Mesh Router, offering software-selectable 2.4 and 5.8 frequency bands (and other frequencies as well), and MiLAN introduced a new Access G access point that is part of a centrally managed Wi-Fi system. It looks like Network General (the makers of the Sniffer protocol analyzer series) is going to re-energize their efforts to provide analysis tools to the SMB marketplace. They have a number of excellent products that are of interest to large, enterprise Ethernet networks and, for a number of years, the SnifferPRO Portable product has languished in obscurity. At Interop there was a mention that the price was going to come down for SnifferPRO Portable, putting it in direct competition with WildPackets EtherPeek NX and Network Instruments Observer. The new SnifferPRO 4.8 has finally revitalized the product and we're watching it closely to see what new wireless analysis features may be added to SnifferPRO Wireless.
When we went to the show we had scheduled a number of meetings with vendor representatives. The fact that everyone was in one city made it productive to get together with folks that often were simply a voice on the other end of a conference call. Suffice it to say, we felt that the show itself was very average, but the opportunity to interact with other professionals was well worth our trip.