CWDP Certified Wireless Design Professional Official Study Guide

Client devices and applications (which are covered in the next chapter) are the most commonly overlooked and under-engineered component in WLAN designs.

Ref. 06B2-A

At the same time, the number of client devices typically exceeds the number of infrastructure devices by a significant amount.

Ref. 56DD-B

you must also understand the client devices that will be running on this WLAN and even the protocols they use to communicate.

Ref. 5B9C-C

understanding and profiling your client devices is one of the single most important components in driving the infrastructure design.

Ref. FC2B-D

It’s also important to know what type of security modes are supported, what application protocols are used, whether the applications require multicast, and how many of these devices may be running at any given time.

Ref. 3ED1-E

Wi-Fi isn’t the only device that uses the 2.4 GHz band.

Ref. 7CEE-F

The other usual culprits for producing 2.4 GHz interference are microwave ovens, wireless video cameras, and cordless phones.

Ref. 04E0-G

4.9 GHz band is also available for licensed public safety

Ref. 2233-H

The challenge for WLAN designers is when client devices support different and potentially incompatible PHYs.

Ref. B961-I

802.11b operates at 2.4 GHz and only supports four data rates, also known as PHY rates—1, 2, 5.5, and 11

Ref. 31F0-J

If your APs are set to 802.11g-only modes or if a single OFDM rate is set to basic or “mandatory,” it will eliminate every 802.11b device from ever associating to the AP. Basic rates are data rates that all client devices must support in order to join the Basic Service Set (BSS).

Ref. 5DD7-K

We recommend that you design a WLAN for both 2.4 GHz and 5 GHz. This provides spectrum redundancy and resiliency in the face of interference. Operating on two spectrums using clients balanced between the two provides up to double the network capacity per AP.

Ref. 71D2-L

Picking up nearly any infrastructure WLAN hardware device, you will see that not only are 1 and 2 Mbps supported but they are specified as basic rates.

Ref. 12E7-M

Briefly, basic rates are rates that must be mandatorily supported for a client device to join the BSS.

Ref. A298-N

nasty side effect of making 1 Mbps a basic rate is that the 802.11 standard specifies that certain 802.11 frames (e.g., management, broadcast, and multicast frames) will be transmitted at the lowest basic rate.

Ref. 9401-O

consequence. The longer the transmission takes, the higher the likelihood that the frame will be interfered with—and the longer other devices that want to transmit need to wait to use the channel.

impact of low data rates

Ref. 00D3-P

Furthermore, let’s use a similar scenario, but this time with two differences: we’re sharing a wireless medium with other users and different data rates are available for use (making other assumptions, of course). Next, we have a device transferring a large file at 1 Mbps PHY rate and another needing to transmit notices that someone is already using the wireless medium (WM). The second device capable of using high PHY rates defers until the WM is available. The second device has to contend for the same airtime while the other device hogs it up because it is using such a slow transfer speed.

important idea

Ref. 2DF5-Q

The WM is capable of higher efficiency, but because the slower rates are supported and being used by even one device, it basically takes the available airtime away for others. In other words, it is like a bad neighbor.

Ref. E14C-R

A good understanding about PHY rates will pay off in all your wireless designs and troubleshooting

Ref. 4ED7-S

To disable these rates, you must test your client devices to ensure that they still operate properly. If you find that a client device has problems when they are disabled and you still have high signal quality, likely a client driver or firmware update will resolve the issue.

Ref. 37A0-T

To be blunt, if someone is still making 802.11b-only capable devices nowadays, they just simply don’t care or they are dealing with some strange constraint. Using an 802.11g device would provide the same or better performance but at least with the ability to use higher PHY rates.

Ref. 57BE-U

Furthermore, the physics of RF dictate that the propagation of higher frequencies do not travel or penetrate as well through most obstructions and the signal decays faster over distance. Therefore, the range is shorter at 5 GHz than it is at 2.4 GHz.

Ref. 2586-V

The biggest challenge with Clause 19 (802.11g) has been backward compatibility for DSSS-only capable devices (802.11 and 802.11b). Simply put, legacy devices simply can’t understand it. To DSSS-only devices, Clause 19 is nothing more than gibberish.

Ref. 1493-W

For example, if a DSSS-only device wanted to transmit on the WM and an OFDM transmission was in progress, the DSSS device might not be able to detect that an OFDM

Ref. 5F85-X

To prevent this, Clause 19 specified certain protection mechanisms for these DSSS-only devices. What these protection mechanisms do is essentially transmit a DSSS transmission that the WM will be in use for a specific time period. Since every 802.11 device must honor these types of medium reservation protocols, the DSSS devices do not transmit during the reservation period specified by the protection mechanisms. Production mechanisms include RTS/CTS and CTS-to-self.

Ref. 4D4D-Y

Nowadays, most organizations have eliminated most of the 802.11b devices from their environment and these protection mechanisms are not as relevant. However, many devices may still use protection mechanisms even when no 802.11b device is present, which causes unwarranted

Ref. ED8A-Z

For many of us using 802.11n at home, going really, really fast until your PC’s traffic arrives at your 5 Mbps broadband connection has some limited appeal. Yes, doing file transfers between PCs and other similar activities will allow you to realize the benefits, but there is another major feature that isn’t marketed nearly as well as it should be: multiple-input, multiple-output (MIMO).

Ref. FC53-A

In certain environments—such as warehouses, outdoor container yards/shipping ports, and multifloor, high-density indoor facilities—a great deal of multipath typically occurs.

Ref. F202-B

an additive signal processing technique to mathematically increase the received signal strength. This technique is called maximal ratio combining (MRC).

Ref. 9A46-C

As for clients, or even an AP for that matter, you need to pay close attention to the type of 802.11n radio that the device supports.

Ref. 3E76-D

The first number is the number of radios; the second is the number of discrete antenna elements.

Ref. AF0E-E

3x3 design, which employs three radios and three antennas.

Ref. 9D1F-F

What is critically significant about the meaning of this value is understanding the capabilities of the client. It reflects the capability of both speed and receive performance. For example, a 1x2 802.11n client has only a single radio but two antenna elements. The number of radios dictates transmit performance that affects uplink speed. For example, a single radio 802.11n design can transmit up to 150 Mbps. Each additional radio adds a theoretical 150 Mbps increase in speed. Therefore, a three-radio 802.11n client such as the Intel 5300 mentioned earlier can transmit up to 450 Mbps (150×3 spatial streams).

Ref. 4D23-G

The number of radios dictates transmit performance that affects uplink speed.

Ref. D1DA-H

The number of antenna elements dictates the maximum possible speeds the client is capable of in the downlink direction. For example, the 1x2 radio mentioned earlier is capable of 300 Mbps (150×2 spatial streams) in the downlink or receive direction. Using the Intel 5300 again as an example, it can receive up to 450 Mbps in the downlink direction.

Ref. 5348-I

More antenna elements allow for more receive spatial streams, and if the antenna elements are separated appropriately, they can greatly increase the stability of received signal quality to nearly eliminate the negative effects of multipath distortion.

Ref. 3902-J

In other words, all 802.11n isn’t created equally and as a designer you must understand this concept well.

take home

Ref. ECCC-K

Remember that when looking at 802.11n clients, both the 2.4 GHz ISM band and the 5 GHz UNII bands are available for use. This means that you will find some clients that only support one band whereas others support both. Pay particular attention to these details.

Ref. 2F47-L

WLAN radios are similar, in some ways, to human beings. Some people can hear better than others.

powerful metaphor

Ref. F9F9-M

In this section we will explore the important characteristics of WLAN radios and how they affect your overall WLAN design.

Ref. 24BE-N

WLAN radios, on the other hand, are usually manufactured up to approximately +/− 3 dB or higher of variance.

Ref. 4ACB-O

you remember that a difference of 3 dB is half (−3 dB) or double the power (+3 dB), this is a significant difference.

Ref. A610-P

The placement of the antennas and the polarity are equally important.

Ref. 13B6-Q

The types of antennas used by one manufacturer could be quite different from another.

Ref. D953-R

Even the placement and orientation of the antenna elements in the lid can change the performance.

Ref. FA83-S

that all devices are unique in their own

Ref. 5D28-T

clearer than they would be at a loud volume. At loud volumes when an amplifier has too much to do, distortion often occurs. At the same volume when there isn’t much sound being

Ref. 0797-U

How big of a problem is this? Usually not much. Part of the

Ref. F0E7-V

reason is that when you are transmitting high PHY rates, you typically have high signal levels, which means the client should be close enough to the receiver for a transmission to still be heard.

Ref. B899-W

So, the problem usually is only visible if you are running an AP at full power and the clients cannot reproduce high PHY rates at the same transmit power levels as the lower PHY rates.

Ref. 3765-X

Many consumer and business hardware purchase decisions are driven by cost. While you generally get what you pay for, if you don’t understand the benefits of one system over another then you are not likely to pay

Ref. 01A2-Y

Imagine a customer who is about to make a large, strategic investment in 802.11n. It is easy for executive management to tout that they are offering 802.11n to employees and key stakeholders, but out of ignorance they didn’t spend the extra $15 per client radio upgrade in their laptops and ended up with a 2x2 802.11n chipset versus a 3x3 one. The 3x3 chipset would result in better radio performance in both throughput and stability.

scenario

Ref. F82A-Z

It may also mean that the client with the lower receive sensitivity has a lower PHY rate that is used versus the one with better receive sensitivity operating at a higher PHY rate.

Ref. B9BC-B

Radio spec sheets should be referenced before performing an RF site survey or designing a new WLAN.

Ref. 6A6A-C

For this particular radio, this tells us that we will have a variable transmit power based on the data rate.

Ref. DB17-D

If the AP that this client connects to is using a full 100 mW at 54 Mbps PHY rates and the client can only perform at a maximum of 20 mW using the same PHY rate, it should follow that the area that 54 Mbps can be heard by a client is larger than the area that a client can transmit and be heard at 54 Mbps. Therefore, the uplink range of the client for the same data rate is smaller.

Ref. 6D11-E

When devices are used in their intended function, it is possible that some performance degradation can occur. For example, a Wi-Fi–enabled smart phone worn in a pocket using a headset will perform differently versus it being held directly to the user’s head. These factors must be considered in a WLAN design.

Ref. 030A-F

baseline of uplink RSSI information while using the client ideally at a distance (testing should rarely be done right next to an AP). For your baseline, the client should be placed more or less in free space without obstructions. 3. Once a baseline is established, use the client in the intended fashion it will normally be used. 4. Now, start taking readings to observe variance to the baseline. You may want to also rotate the client’s orientation to the AP in cases where the device might be worn, such as with a Vocera communications badge, or a device that experiences a variety of antenna polarity changes. Ninety-degree increments would be fine to get a gauge. 5. If the data shows a high degree of variance in the use cases, angle, or orientation to AP, factor this into your design. For example, let’s say you are getting satisfactory signal readings near the edge of the building from the APs placed in the hallways when you surveyed with a professional surveyor utility.

Ref. C511-G

Because the combination of that particular device facing away from the AP and a person standing between the signal propagation to the AP lowered the signal strength by approximately −10 dB, you might experience problems with these devices when used in production at the edges of the RF cells. Some examples of device-specific use cases that should be considered include the following: Handheld Computers This includes ones used by warehouse staff—possibly at different elevations if forklifts or man lifts are used. Vocera This is a VoWiFi device that is worn in the center of a person’s chest. RFID The signal of this usually tiny RF device might be obstructed depending on how it is mounted to assets or people. Mobile Workstations These are sometimes referred to as computer or workstations on wheels (COWs or WOWs). Vehicles Vehicle mount antennas require careful placement and are often more sensitive

Ref. 13A2-H

Some of the most significant MAC features, those that play an important role in client design, will be discussed next.

Ref. 9AAF-I

Designing your WLAN and client devices to support efficient roaming using strong security is the holy grail of WLAN design—no compromises.

Ref. B3DF-J

For instance, consider a network design where only 5 GHz is available but the client device favors 2.4 GHz, or vice versa. Using the first example, on every roam, the device scans and probes 2.4 GHz before it does 5 GHz and adds time to each roam.

Ref. AC96-K

Many client adapters will scan using passive, active, or both methods to discover access points. If your design incorporates a three-channel plan—for example, channels 1, 6, and 11—it would be quite useful for optimize your client scan behavior. This aspect is especially important

Ref. 7173-L

you are using a VoWiFi device that has high sensitivity to roaming delays.

Ref. 124A-M

If your voice device is attempting to roam without knowing what channels your infrastructure is operating, the time involved in brute-force scanning channels you don’t even use can cause unnecessary delays and unhappy end users. Figure 2.6 is an example of a client capable of channel optimization.

Ref. 56A8-N

Some devices allow you to control how aggressive or conservative the signal value will be that kicks off the roaming process.

Ref. BF98-O

to an AP and start walking down a hallway, pass another AP, and the client would drop its connection and then finally reassociate. This “sticky” type of behavior would spell disaster for your end users. The issue is quite simply a client issue.

Ref. B96F-P

We have learned a lot in the WLAN industry over the years and the methodologies that vendors are using have improved—sometimes drastically—in only a few years.

Ref. 5144-Q

Designing and supporting client devices starts with understanding and analyzing the type of PHY support available. You learned that it is important to know what PHYs are supported and which one performs best in a given situation.

Ref. FFF1-R

TCP has built-in mechanisms for handling difficult network conditions that can result in lost transmissions, duplication, and segments delivered out of order.

Ref. 4E05-S

Many handheld computers used in retail and warehousing applications as well as

Ref. 2C4E-T

some medical systems will use a Telnet type of protocol.

Ref. 9ED4-U

Sure, some of you are already thinking you know exactly what RSSI metric or SNR value your design target at cell edges should be, but we’re asking you to forget about all that for a moment so you can explore how many vendors have generally come to the same value.

Ref. 51D1-V

Rate shift boundaries Roaming characteristics and target RF metrics Cell edges and transition

Ref. F8B9-X

Knowing the factors involved with rate shifting is critical to understanding the primary factors governing enterprise WLAN designs.

Ref. 5DE2-Y

Essentially, as the signal strength gets weaker, the less fidelity can be deciphered from the received signal. Therefore, the radios shift their PHY rate to a less intense modulation or maybe the coding rate decreases.

Ref. CF9E-Z

This usually does the trick for the other end to decode the transmitted message at the other end of the RF link.

Ref. EC3D-A

That section in Chapter 2 also explored the time factor differences involved in WLAN transmissions at different PHY rates. While VoWiFi traffic isn’t very bandwidth intensive, it requires reliable transmission and reception of each end of the voice communication. Even a small amount of data can result in a lot of airtime usage as PHY rates degrade to lower levels. As other conversations and traffic might be present, the longer transmissions involved in the lower PHY rates create RF contention. Furthermore, RF contention means that devices need to wait until the WM becomes available before they can transmit. When waiting is incurred, jitter starts to enter into the communication link, resulting in unfavorable voice quality or voice delays.

Ref. 7DFC-B

It doesn’t always mean that contention is playing a role.

Ref. 1F08-C

that when devices perform rate shifting, it is usually because some of their transmissions aren’t being properly heard.

Ref. 2340-D

Also per the standard, when the transmitter didn’t get an ACK frame, it implies that the transmission wasn’t heard by the receiver of the message (unless a block ACK policy is negotiated, but we will defer that discussion for now). What’s

Ref. 9C88-E

The methodology for that policy is out of scope of this topic, but as more retransmissions start to occur, the backoff timer keeps

Ref. F20D-F

doubling. As more consecutive retransmissions occur, the longer the backoff timer value becomes, which can result in substantial delays from a voice quality perspective.

Ref. CF3A-G

1. Assume the connection begins at the highest level of quality using 802.11g and the PHY rate is operating at 54 Mbps. 2. The mobile device starts to walk away from the AP, and as it is sending all of its traffic, the signal gets weaker and weaker. 3. As the signal reaches a certain point, eventually the 54 Mbps PHY transmissions will not be decipherable by one or both ends of the communication link. When that occurs, no ACK is sent, and perhaps the device simply tries again at 54 Mbps. The client will perform this for a device-specific number of retries.

Ref. D8CF-H

sauce” of each vendor, let’s assume that the device tries up to two additional times at 54 Mbps, which can be regularly observed by several devices.

Ref. 039A-I

time is still ticking and nothing has come through to the other end of the link. 5. Now, the device sends a 48 Mbps frame and still doesn’t get an ACK. 6. The device then changes to 36 Mbps, and then it finally gets an ACK. Likely both sides then rapidly flush their buffered frames that have been continually recorded as you have been talking.

Ref. 508C-J

Once the RF conditions stabilize even at a lower PHY rate, voice quality will start to improve again. As long as the WM isn’t overloaded, there should be plenty of available transmit opportunities, even at middle- and lower-level PHY rates, to sustain a normal conversation where users may not notice any degradation of quality.

Ref. 1CAD-K

As PHY rates continue to drop, the time to transmit the 802.11 frame becomes longer. And, as the 802.11 frame is retransmitted multiple times, it has a combined result of greater jitter delay and even packet loss.

Ref. E8F0-L

Generally speaking, with modern voice devices, transmit power can be assumed to be 25mW.

Ref. 51B6-M

With that assumption, let’s take into account the knowledge you gained from Chapter 2 regarding radio characteristics. Since we assume that voice clients transmit at a maximum of 25mW, the AP transmit power should be reciprocal. We’ve just determined the maximum AP transmit power for the survey.

practice

Ref. 037F-N

you will be using an automated RF management mechanism, you can even back off this transmit power by an additional 3 dB (and, therefore, use 12.5 mW of AP transmit power) to give yourself

Ref. 4DA7-O

headroom and still stay within your link budget if it is ever warranted.

Ref. ED78-P

If you refer to a spec sheet for an AP, you should see specific receive RSSI values in dBm indicating where the radio will need to change to a different PHY rate.

praxis

Ref. 1EE9-Q

Although this process is harder than designing for voice, the same principles apply; you need optimized RF cell edges, good-quality roaming algorithms, and the ability to keep PHY rates high.

Ref. 77D2-R

Assuming that the video clients will operate in a voice-optimized network, the next best payoff when designing for video is to focus on the WLAN radio the video endpoints will use. That means focusing on the antenna types and placement, radio technology, MAC feature support, QoS, and other similar factors.

Ref. D1DF-S

Retransmissions are a performance killer. If you are experiencing retransmissions of 802.11 frames, you need to determine the root cause and seek to eliminate it.

Ref. 0787-T

As you have read, designing for real-time applications over 802.11 isn’t child’s play.

Ref. 7EA1-U

In our previous discussions, we have assumed that clients have the ability to roam to an AP when they need to do so. That is a really big assumption.

Ref. 6BD6-V

Yes, producing documented deliverables stating design objectives and exclusions will protect you financially and legally, but customers often have convenient memories.

Ref. 0BCD-W

The best recommendation we can give you is to design for real-time applications today.

Ref. 7EFF-X

It is your responsibility to explain the pros and cons of data only wireless networks and how this

Ref. CBA6-Y

Another consideration is how these devices are powered. If a Power-Over-Ethernet (PoE) switch were to fail, it can mean an outage for quite a large area.

Ref. 6BB4-Z

One design method that can be used is to stagger adjacent APs onto two different PoE switches so if one switch

Ref. 2509-A

fails, at least half of the APs will still be online and cover the same physical area. The coverage will be spotty, but it is far better than a complete service outage.

Ref. A7E8-B

The closer you are to the transmitter, the higher the change in signal strength. For example, looking at Figure 3.3, when a transmitter is 5 meters from a receiver, it reads a signal strength of −60 dBm.

Ref. DBE2-C

meters and 80 meters away, there is a change of only 6 dB. This example of the inverse square law illustrates the need to have the receivers as close to the transmitters as possible in order to obtain good location accuracy.

Ref. DA35-D

RF fingerprinting is a way of mapping x,y locations on a map based on recorded RSSI values.

Ref. 4C44-E

Everyone Else Because AeroScout tags do not actually associate to the WLAN, AeroScout tags use a specially formatted 802.11

Ref. B157-F

For example, if the tag needs to transmit, it will do so on channels 1, 6, and 11 if that is how the infrastructure is currently configured. This allows the maximum amount of APs to hear that tag wherever it may be located.

Ref. F971-G

What we recommend, based on experience, is to start with a sound voice-grade design as a baseline.

Ref. 0F76-H

This is valuable information that you can put to use right away in your own projects.

Ref. 3821-I

Deploying wireless in healthcare environments, multifloor hospitals in particular, is one of the more challenging tasks for Wi-Fi designers.

Ref. 2FEA-K