Monday, September 10, 2012

Onsite Pre-Installation Site Survey vs. Predictive Pre-Installation Site Survey


Onsite Pre-Installation Site Survey vs. Predictive Pre-Installation Site Survey
In contrast to wired LANs, a Wireless LAN deployment isn’t that straight forward. Wired LANs can have a range of Ethernet switches, routers, wiring outlets, servers, etc. When all of these devices are configured for a certain design they can be deployed in any type of environment, building or country. When properly deployed, the overall performance (throughput, capacity…) will be the same in all scenarios.

Wireless LANs are a whole different ball game. You will also need wiring (not that funny), Ethernet switches, routers, servers, etc. but the big difference is the access layer which is the "air" for Wireless LANs as oppose to an RJ45 wired outlet when dealing with wired LANs. The overall performance of a wireless client depends on where it is physically located in relation to the access point that it is connected to.
Every building has its own specific characteristics in regard to the propagation of RF signals. It is for this reason that you cannot copy one specific Wireless LAN design to another building in contrast to the wired LAN.

With this, we can say that the design of the access layer is a very crucial part of a successful Wireless LAN deployment. The design of this access layer is done through a Wireless LAN site survey. The goal of a Wireless LAN site survey is to determine the amount and correct positions of access points & antennas so that the Wireless LAN requirements (coverage, throughput, capacity…) are met throughout a predetermined coverage area.
There are two ways to do this, the "onsite pre-installation site survey" and the "predictive pre-installation site survey".

The Onsite Pre-Installation Site Survey

During this site survey the Wireless LAN engineer will go onsite with the documented Wireless LAN requirements and ground plans. After a kick off meeting the survey will which includes:
  • ·      A visual inspection of the coverage areas
  • ·      Spectrum analysis
  • ·      Tune/calibrate the measuring device(s) with the weakest wireless client device that will deployed (upstream and downstream)
  • ·      Start measuring the requested WLAN application requirements
  • ·      Onsite documentation of the results
When the onsite measurements are complete, the results of the entire analysis will be documented in the WLAN Site Survey Report.

The Predictive Pre-Installation Site Survey

This site survey can basically be done anytime & anywhere and there is no need to go onsite. The Wireless LAN engineer will import all of the ground plans into the planner software, and will enter an estimated attenuation value for each wall, elevator, stairwell or any other attenuating object. Next, the software will be configured with the Wireless LAN requirements of the applications. Finally, the planner software will disperse the access points’ layout on the imported ground plan. A Wireless LAN Site Survey report will be generated automatically or can be written manually.

So, which is the most preferred way? Let's compare both methods and discuss the pros & cons.

This comparison is based on a Wireless LAN site survey project for a warehouse/factory with offices on multiple floors, and our experience with both types of site surveys throughout Europe.

Onsite Access

Predictive Site Survey pros

One of the advantages of a predictive site survey is that you don't have to go onsite. You can simulate the entire site survey on imported ground plans in a planner software tool. This could come on handy when the facility is not yet constructed or is still under construction.

Predictive Site Survey cons

The accuracy of the site survey highly depends on the details and accuracy of the received ground plans as well as the accuracy of the predicted attenuation of the objects that have been drawn on the plan. Do the ground plans reflect the latest construction works and are all the interior objects/furniture displayed? Getting accurate ground plans from the customer could be challenging with old buildings.
Drawing in the walls and other objects also requires a thorough knowledge of the attenuation of these materials. Let's take an office with glass walls as an example. RF signals pass easily through normal glass but coated glass is a real RF killer.
Measurements are calculated for downstream RSSI only.

Onsite Site Survey pros

Being onsite get's you more involved in the project. Contact with the customer and a walkthrough of the facility gives you a better perspective on the characteristics of the facility and the aesthetic options for mounting the access points. 
Measurements can be done with the weakest wireless device (iPad, iPhone) and can be tested downstream and upstream.

Onsite Site Survey cons

Well, you need to be onsite which could mean travel and hotel expenses, and if the facility is not yet constructed or under construction you can't do the survey.


Time consuming & budget

Predictive Site Survey pros

A predictive site survey will take less time to do, and will therefore be advantageous for urgent & low budget projects. Less time also means less budget!

Predictive Site Survey cons

On the other hand, an inaccurate predictive site survey could be a budget killer (if any is left over). If the Wireless LAN performance does not meet the requested Wireless LAN requirements after deployment, additional access points & cabling could be needed. The same goes if too many access points are installed and need to be disabled/removed. Either one of these scenarios will have a ripple effect on the overall access point layout.

Onsite Site Survey pros

An onsite site survey will be for 99% accurate so that there will be no surprises at end of the road.

Onsite Site Survey cons

An onsite site survey will take more time and budget.


Visual inspection of the coverage area

Predictive Site Survey pros

N/A

Predictive Site Survey cons

The coverage area is only visible on the ground maps.

Onsite Site Survey pros

It is an added value to inspect what the RF barriers are, and what to take in account regarding the building aesthetics and cabling for access point and/or antenna installations.

Onsite Site Survey cons

To be onsite.

Access points #

Predictive Site Survey pros

Fast and cheap way to have an estimation of the number of access points and antennas needed.

Predictive Site Survey cons

Accuracy is the big drawback here. Normally an additional percentage of access points are calculated in the Bill Of Materials. Having extra access points to fill in the gaps afterwards is not the most professional approach; not to mention the additional cabling that will be needed.

Onsite Site Survey pros

Accurate number of the "measured" access points, if done professionally there will be no need for additional access points or cabling.

Onsite Site Survey cons

To be onsite.


Positioning of the access points

Predictive Site Survey pros

No need to be onsite, the planner tool will automatically disperse the access points on the ground maps.

Predictive Site Survey cons

Accuracy. This is especially true in warehouse (storage racks) and factory (robots, objects) environments. 
In 802.11n deployments it's vary hard to predict the effect of multi-path, which is beneficial in 802.11n deployments.

Onsite Site Survey pros

Accuracy. Taking the measurements on a measured location will highly increase the accuracy of coverage, overlap and capacity. Multi-path effect can be measured more precisely. 
Being onsite you can take in account aesthetic limitations if applicable.

Onsite Site Survey cons

To be onsite.

Mounting accessories

Predictive Site Survey pros

I can't think of any here.

Predictive Site Survey cons

It is especially difficult to predict which type of access point mounting accessory would be suitable for each access point location in non-office environments.

Onsite Site Survey pros

While onsite and measuring correct access point locations, you can perfectly determine the mounting accessory needed for the access point. Let's say that the access point needs to be lowered down 6m to provide service between storage racks but needs to be high enough so that the forklifts will not hit the access points when they load the pallet to the highest rack. This will require a specific mounting accessory.

Onsite Site Survey cons

To be onsite.


Spectrum analysis

Predictive Site Survey pros

N/A. unless you could say that the deployed access points will have spectrum analysis onboard.

Predictive Site Survey cons

It is not possible to perform a spectrum analysis that will detect non-WiFi interference with a predictive site survey.

Onsite Site Survey pros

A spectrum analysis can be performed during the onsite pre-installation survey.
The spectrum analysis will scan for non-WiFi related sources of interference. These types of interference could/will disrupt Wireless LAN transmissions because they do not respect WiFi protocols. This is why they are defined as non WiFi related sources of interference. Even if the Wireless LAN access points that will be deployed have spectrum analysis onboard it's highly recommended to do a spectrum analysis in the pre-install phase, due to the fact that you want to deal with this interference as soon as possible. Detection of non-WiFi interference in a post-installation phase will cause deployment delays.

Onsite Site Survey cons

To be onsite.

Overall Conclusion

Taking everything in consideration there are two big differences between both site surveys, the first one is that for an onsite pre-installation site survey, you need to be onsite and have access to the coverage area (which speaks for itself), while you don't need to be onsite for a predictive site survey. The onsite site survey will initially be more time consuming and more expensive (especially if travel expenses and hotel costs are added to the bill).
The second big difference is the accuracy of the survey, which is way more efficient with the onsite site survey. Putting these two together it will be a trade off between the 2 types of site surveys.
You can save money on the site survey by doing a predictive site survey, but loose more money and time if after the deployment the Wireless LAN does not meet all the requirements and results in bad performance, new cabling, repositioning of access points, adding access points. Keep in mind that one bad positioned access point could have a ripple effect on the whole access point layout.
Or, you can go for the onsite pre-installation site survey, which will initially cost more and will take more time, but will provide accurate results with no surprises at the end of the road.

Recommendations

Unless you're Wireless LAN project consists out of small office floors or separate meeting rooms that require non-latency sensitive Wireless LAN applications we highly recommend onsite pre-install site surveys.
Although cost and time will be higher you are sure you've got what you bought.

Regarding the overall cost of an onsite pre-install site survey, there are many manuals and best practice guides out there which speak of the integration of the site survey in overall project phases meaning several visits to the customer.

To decrease the overall cost of the onsite pre-install site survey we recommend to do a "clean cut" site survey, which means that you cut as much overhead as possible during the onsite pre-installation site survey (rogue detection, heat maps, etc.). Most of these measurements can be done by the Wireless LAN infrastructure before deployment and be dealt with during the verification site survey.

Take note that with either type of site survey a verification site survey before deployment is "highly" recommended. That’s right folks it ain't done with the site survey only. A fine-tuning of the RF, data rates, TX powers, etc. to the requested Wireless LAN applications is crucial for a fully optimized Wireless LAN deployment.


Monday, July 9, 2012

Beamforming and How to survey for it?



Beamforming and How to survey for it?

Beamforming is popular these days with certain WiFi venders, this technology is maybe not a big mystery for the geeks in the WiFi industry, although I see and hear a lot of misconceptions around beamforming.

What is it?
Beamforming allows an access point to concentrate RF energy in the direction of the WLAN client/receiver in order to improve the Signal to Noise Ratio at the WLAN client/receiver and thereby improving WLAN performance (Increased SNR = increased data rates & decreased retransmissions).

So basically you could say beamforming affects the radiation pattern of the access point, so with this in mind anything that affects the RF pattern should be taken in consideration when doing a pre-installation onsite site survey, correct?

Before we elaborate on this further let's dig into beamforming first to grab a good understanding what it actually is (anyway how I see it ;-)).

Beamforming comes in 3 versions, static beamforming, transmit beamforming and dynamic beamforming (sometimes they are called different but for this blog we use these names to differentiate).


Static Beamforming:
Under static beamforming we can understand that an access point will use a fixed directional antenna to focus the RF energy (beam) in a certain direction, the coverage area in this direction will extend in comparison with a traditional omnidirectional antenna. Almost every vendor can do static beamforming when their access point supports external (directional) antennas. The directional antenna will be aimed (fixed) towards a certain area to improve SNR for that area, keep in mind that all mobile WLAN clients have small omnidirectional antennas on board that also need to send traffic back towards the directional antenna of the access point (more on that later). So static beamforming will be mostly used to improve SNR for a certain location or to extend a coverage area.
Some vendors like Xirrus, use an array of static directional antennas to improve SNR in 360 degrees, by assigning a different channel to each beam they can be packed together within a single array.


Transmit Beamforming (TxBF)
Under transmit beamforming we can understand that an access point will transmit 2 or more phase-shifted signals towards a WLAN client/receiver in order to achieve an constructive signal at the WLAN client/receiver. The phase-shifted signals as they travel through the air create constructive signal (in phase) points at various locations in time and space, the goal here is to achieve this at close as possible to the WLAN client/receiver and thereby increase SNR.
To do this correct, possible feedback is required coming from the WLAN client/receiver to help pinpoint the location of the WLAN client/receiver.
Explicit and implicit transmit beamforming are 2 forms of optional components of the 802.11n standard and require feedback from 802.11n clients, as till now 802.11n chipsets do not support either form of transmit beamforming.

Although Cisco has his own proprietary form of TxBF called ClientLink, version 1.0 works only with legacy 802.11ag WLAN clients without requiring feedback. Optimizing SNR values on the 802.11ag clients in a mixed environment (802.11agn) will increase airtime for all 802.11 WLAN clients and is beneficial for overall WLAN performance.
ClientLink 2.0 works for all 802.11agn clients (release 7.2 on the 3600 series access points).


Dynamic Beamforming 
Under dynamic beamforming we can understand that an access point will focus RF energy (beam) towards a certain WLAN client/receiver in a way it's kinda like static beamforming but with a big difference that it is not static and uses an dynamic antenna array that can change it's radiation pattern on a frame by frame basis. It selects the optimum antenna pattern for each communicating device in real time, while actively avoiding interference and minimizing noise to nearby networks and devices.
Test results showed that dynamic beamforming had the highest SNR increase in an overall WLAN deployment of the 3 beamforming technologies as described above.
The only WiFi vendor currently offering dynamic beamforming technology is Ruckus Wireless.


Site survey considerations
By understanding the 3 different types of beamforming, what do we need to take in consideration while doing a onsite pre-installation site survey?

This is how we see it.
Note that these are only additional considerations for the onsite site survey concerning beamforming technologies, during the actual site survey other considerations like capacity, throughput and so on need to be addressed also.




Static beamforming: 
Because static beamforming will cover only one certain coverage area following points need to be taken in consideration.
   Always measure with the antenna type that will be proposed in the bill of materials.
   Adapt your TX power settings before you measure (see additional note at the end).

Transmit beamforming:
In this case only cisco has a proprietary form of TxBF, we recommend to follow their recommendations (I really appreciate all the documentation Cisco put's out there) 
   "Although ClientLink increases the SNR and data rates of clients at the edges of the cell, it does not extend the maximum range of a cell. This is because certain packets that must be heard by all clients (such as beacons) are sent as broadcasts, which ClientLink cannot optimize. These broadcast packets effectively become the limiting factor on the cell size. Therefore cell spacing should not be increased when using ClientLink. In fact, to be conservative, the ClientLink feature can be turned off during a site survey." this makes sense to me I would say ClientLink feature must be turned off during a site survey.
   Take note that ClientLink 2.0 is enabled by default, and needs to disabled by a cli command.



Dynamic Beamforming:
In this case only Ruckus supports this technology, I have a tiny issue with their recommendation to do onsite site surveys. I've got this information from their technical guy at a Big Dog seminar.

   Ruckus recommends that onsite measurements need to be done with the WLAN client/receiver in the beam and with the access point at full tx power, note that the beacons are not broadcasted in the directional pattern but in a omnidirectional pattern to communicate with other WLAN clients/receivers. Also sending out traffic in a directional pattern at full power (20dBm EIRP in Europe) raised an big question mark in my mind that is related towards WiFi math and physics, certainly if you want to support smart mobile devices (see additional note at the end). 
  I would recommend doing the site survey with the access point tx power settings at 14dBm IR.


Additional Note: this note refers to the "antenna reciprocity theorem" (damm that’s a hard one if you're native tongue is not English ;-)) which I refer to when it comes to tx power settings and antenna options.

Basically what this theorem says is the following, If an electromagnetic force of some particular magnitude is applied to the terminals of antenna "A" and the received current is measured at some other antenna "B" then an equal current (in both amplitude and phase) will be obtained at the terminals of antenna "A" if the same electromagnetic force is applied to the terminals of antenna "B".

In human terms, no matter with type of antenna is used on both sides if the input power on both ends is the same then, if you can hear me, I can hear you.
Explained in a real world scenario: if you're laptop has a wireless NIC with a lousy omnidirectional antenna on board and is sending at a tx power input of 17dBm (50mw), and I have the same laptop with the same setup sitting 1km away from each other. In this case we would probably not be able to set up a wireless communication because we don't hear each other, if I connect a high gain directional antenna to my wireless NIC and aim it towards you're direction whenever I can hear you, we will be able to set up a wireless link even if you're lousy omnidirectional antenna don't reach me, as long as we have the same tx power input on both laptops (tell this to all the IT managers that for security reasons are eager to block WiFi signals so that they don't bleed into the parking lot).

So lets put this in an access point and a mobile wireless device relationship. 

Consider what happens if an access point has a tx power input of 17dBm with a directional antenna and the mobile wireless device has an tx power input of 14dBm with a low gain omnidirectional antenna, due to the fact that the access point is more powerful then the mobile wireless device, then "reciprocity" is broken and around the edge of the cell the mobile wireless device will attempt to associate with the access point that it will never reach.

Consider what happens if an access point has a tx power input of 14dBm with a directional antenna and the mobile wireless device has an tx power input of 17dBm with a low gain omnidirectional antenna, due to the fact that the access point is less powerful then the mobile wireless device, also "reciprocity" is broken, but there is a big but here ;-), reciprocity is broken in a way that augments, rather than detracts.


So taking all of this in consideration, and if you're WLAN needs to support smart mobile devices (BYOD, which have a limited power input) it's better to have your access points on limited tx power settings, so take this in account when doing the onsite site survey.  

Please reply, if there are any remarks or somebody has additional information regarding this topic.
Thursday, June 21, 2012

Detecting the RF Noise Floor


Detecting the RF Noise Floor

In preparation of the CWAP exam I came across a major misconception of how the RF noise floor is detected and measured that warrants being shared. Before I start let’s define what RF noise floor is.

Noise Floor
The RF noise floor is defined as all the background RF signal that is received in the frequency range that your device is operating in. This could be generated from any type of device whether an intentional radiator, such as wireless cameras, or an unintentional radiator, such as lights or motors. Determining the noise floor is important in in site surveys and troubleshooting because with it you can determine the signal to noise ratio (SNR), which is the measurement in dB of how high your received signal from access points is above the noise floor. Or in simpler terms, how much of the signal from the APs can actually be heard.

RF Noise Detection in Wireless NICs
The wireless NIC is not a spectrum analyzer, and though it can transmit and receive data at impressive rates, the only thing getting past its encoding filter is bits, it cannot see raw ambient RF signals.

So how does the wireless NIC report the noise floor? After all you have seen many different screens from various 802.11 devices displaying either noise or signal to noise ratios (SNR). Well, various vendors have come up with unique ways to guess the noise floor but again; since wireless NICs can only process bits they cannot see ambient RF signals. What’s worse? Each vendor that manufactures 802.11 devices calculates noise in a different way. Each of them have developed sophisticated algorithms for calculating noise based on bit errors and other factors. Some have even figured out how to turn off the bit encoding filters and use RF signals coming through the antennas much like a spectrum analyzer, however they can never encode bits at the same time.

Scenario
Imagine that you are in a room surrounded by a Faraday cage and absolutely no RF signals from outside the room are audible. You turn on your wireless NIC based protocol analyzer. Obviously, there will be no noise to measure.

Now you bring a microwave oven into the room and turn it on. Guess what…. still no noise is detected. Why? There are still no modulated signals and thus no bits to corrupt or harm.

After that, you bring in an AP. Finally you will see noise, but the values will be very low because the AP will only be beaconing meaning that there is only a small amount of modulated bits riding on the RF. The microwave will harm those bits and the wireless NIC interprets the broken and harmed modulated bits as noise.

Finally, you add 4 wireless clients to the room. You get them connected to the AP and start generating data traffic. Now you can see a high level of noise because numerous bits are now transmitted across the RF medium in the form of frames. These frames are all hurt by the microwave oven, with lots of corrupted bits causing the wireless card to report high levels of noise.

Site Survey
So, when you do a site survey you are affectively measuring the signal strength of your access point as well an estimated SNR only the channel you are measuring in.

This causes problems because you do not know which channel a wireless controller will place the access point after the installation. You could of course measure each AP location on each possible channel but you could imagine the amount of time it would take and therefore making the site survey unaffordable.

Luckily, now a day, RF site survey engineers are most likely to be equipped with spectrum analyzers that are capable of seeing raw RF energy from any device in the frequency band.

Bottom Line
A spectrum analysis is the only way to see the real floor noise (or interference) in a certain environment and you cannot depend on the interpretation of the noise floor from wireless NICs.

Source: Certified Wireless Analysis Professional Official Study Guide, Cisco Spectrum Expert User Guide


Wednesday, June 20, 2012

3 important steps to make you're WLAN project successful


3 important steps to make you're WLAN project successful

Throughout the many years that we have been in the business, we have seen many different approaches on how WLANs are deployed. Sometimes they were successful and sometimes they were not. If you are involved in the WLAN project from A to Z the flow is transparent. Otherwise, depending from where you step into the story, you are always dependent on third party information which, is sometimes contradictional with what you are doing onsite. Our part of the story was mostly the pre-installation site survey which is a crucial part of the whole but somewhat useless if, for example, the information regarding the WLAN applications that will be deployed on the WLAN is not correct.

Another thing that we noticed was that after the site survey was executed and documented, the WLAN infrastructure would be installed and no fine tuning or verification site survey would be performed before the deployment of the WLAN applications which, from our point of view, is another crucial part of the whole. For example, during the site survey measurements were taken while the access points are transmitting at a certain static tx power level and after installation the radio management algorithm of the controller (if architecture is controller based) will feel the access points and adjust channel and tx power settings automatically. Depending on the building infrastructure and the access point density, the radio management algorithm could lower down the tx power of the access points below the threshold used during the site survey. This could cause coverage holes in the coverage area. Another example is that the access points are installed incorrectly according to the site survey recommendations and the radio pattern will be different than as foreseen.

With all this in mind, we want to stress out the following 3 important steps to achieve a successful WLAN project.

Step1: Information gathering
I guess all WLAN engineers who are doing site surveys have heard this one "I want a site survey, for everything" ;-).
Step 1 is probably the most important step. Gathering information about the WLAN project, such as defining coverage areas, which type of applications will be used now and in the nearby future and in which frequency, high capacity areas, high utilization areas, WLAN vendor and type of access points, aesthetic restrictions for the placement of access points, and so on. I know….. it is a true hassle to collect all the information and sometimes the customer does not know what you're talking about.
The best way to do this is to fill out the Request For Information document. All of the questions to get the information are in there. You just need to fill in the check boxes.

Step2: The Pre-installation (onsite) site survey
Once all of the information gathered from the RFI you can actually start doing the onsite measurements. A small kick-off meeting with the customer onsite to double check the information won't do any harm.

The goal of the onsite site survey is to pinpoint the amount of access points and/or antennas needed to meet the requested WLAN requirements, as well as how and where the APs should be installed.
A spectrum analyses is also required to check if there is any non Wi-Fi interference onsite that could disrupt the performance of the new WLAN infrastructure.
Rogue access points/signals detection can be an additional option to the WLAN site survey.
All the results will be documented.

Step3: The verification site survey (post installation site survey)
Most of the times Step3 is not foreseen in the project due to budgetary or other reasons or there might be a time gap between the pre-installation site survey an and the installation of the WLAN infrastructure. And I agree, most of the deployments will have no problems running their WLAN applications certainly if they are only data applications, but that does not mean that you're WLAN is performing in the most optimal conditions. When adding more and more applications, performance problems can arise. Anyway, we highly recommend adding Step3 in the process before the project will go live.

The goal of the verification site survey is to optimize and fine-tune the overall WLAN to the requested WLAN requirements. Coverage, retry percentage, tx power settings, data rates, channel reuse and AP installation should all be taken into consideration.
All results will be documented.

Cheers,
Joeri De Winter

About Us

Skyline networks is specialized in a professional range of wireless LAN services worldwide. We offer customized services for every aspect of your WLAN.

The Skyline Blog Authors:

Joeri De Winter
Justin Cetko

Joeri De Winter

Follow Joeri

Justin Cetko

Follow Justin

Twitter feed