Thursday, January 24, 2013

Sometimes I think we're not speaking the same language!

Listening is a key skill for IT consultants, but I find myself more often trying to understand what the customer really expects when they say they need a 'site survey'.

A 'site survey' can mean quite a few different things and there isn't a standards body to formulate and finalize standard descriptions of site survey terms. 

Active, passive, predictive, validation, simulated and budgetary are all words I've heard used to describe the type of site survey a customer wanted.

The first pieces of information I need to get are: 

  • Does the customer have an existing wireless deployment?
    • Do they want their current RF coverage documented, then augmented?
      • If augmentation is desired, existing access point power would need to be reduced to ~25mW before documenting the RF coverage
      • Augmentation may not be the best solution due to the potentially vast differences in their current access point hardware and antennas vs. the features available in today's access points.
    • Do they want to deploy a new wireless coverage design and disregard their current deployment?
  • What does the customer intend to do with their wireless infrastructure?
    • What types of client devices do they anticipate?
    • What kinds of applications will be used?
    • Is there the potential for high-density user counts? 
    • Is any level of real-time location tracking required? 
Each deployment is different as a result of the building type, building layout, client use cases and customer budget. There are a great deal of questions that need answers before you can accurately define the type of site survey the customer needs.

Defining the different site survey terms is often problematic. Everyone seems to have their own idea of what an 'active' or a 'passive' site survey means. These are the terms I use to describe site surveys:
  • Active (also known as AP-on-a-stick)
    • Digital floor plans are obtained and scaled to a known distance.
      • If the floor plans do not have a scale indicated, the scale of the drawings can be obtained by Google Earth measurements or by using a laser distance measurement tool when on site at the facility.
    • The RF engineer brings in the actual access point the customer will purchase, powers the access point with a battery pack and documents the coverage of the access point in the actual installation location with the site survey software of choice. The access point installation locations can be surveyed one at a time and then merged, or multiple access points can be surveyed at a time (depending on the hardware resources available to the RF engineer)
      • This type of survey will not give you accurate data on the RF noise floor that will be present once the final access point infrastructure is installed and operational, as there are a limited number of access points transmitting during the survey period.
    • A Spectrum analysis should also be performed to determine if there are any major fixed sources of interference present in the facility.
  • Passive (also known as validation)
    • Digital floor plans are obtained and scaled to a known distance.
      • If the floor plans do not have a scale indicated, the scale of the drawings can be obtained by Google Earth measurements or by using a laser distance measurement tool when on site at the facility.
    • The RF engineer walks the facility with the survey laptop detecting one or more SSIDs in use at the customer site.
      • The RF engineer may not have access to network credentials to authenticate the the customer's WLAN and would only document the RSSI/SNR detected by the site survey application.
      • If the RF engineer has network credentials to associate/authenticate to the WLAN, additional testing of throughput/load/voice quality could be performed at the customers request.
    •  Spectrum analysis should also be performed to determine if there are any major fixed sources of interference present in the facility.
  • Predictive
    • Digital floor plans are obtained and scaled to a known distance.
      • If the floor plans do not have a scale indicated, the scale of the drawings can be obtained by Google Earth measurements of the outside wall of the building, or by using known distances common to corporate facilities (doorways are typically 3 feet wide).
    • RF obstacles are drawn in the RF simulation application
    • A software model of the access point's RF coverage is used to place the access point at a given location in the floor plan drawing and predict what the coverage for that access point would be given the RF obstacles drawn by the RF engineer.
    • Once the desired coverage level is achieved, a bill of materials can be generated
    • NOTE: Predictive surveys are not 100% accurate and a post-deployment verification survey should be performed to ensure the desired coverage was achieved through the predictive survey process.
  • Throughput/load test
    • There are several methods to performing throughput testing (this list is just an example and is not all inclusive)
    • The RF engineer would associate/authenticate the wireless client to the customer WLAN with credentials provided by the customer.
    • The RF engineer would also need access to a wired port to do end-to-end testing from a wired client to a wireless client.
  • Voice quality test
Hopefully the more we all talk about the different types of site surveys and learn from one another about the words we use to describe them, we'll all encounter less RFPs/RFQs asking for validation surveys for wireless infrastructures that aren't installed in buildings that aren't built yet.



5 comments:

  1. Thanks for this post, I've also seen a couple of your videos, great work!, this clarifies a lot, I'm currently looking for someone who can help us troubleshoot a WiFi issue and recommend what hardware is needed to improve it in general, I believe I'm looking at a passive+predictive site survey, I'll try to get in touch with you to see if you can help us.

    ReplyDelete
  2. For the AP on a stick surveys, I prefer the term - pre-deployment survey, while for the validation - post-deployment survey.
    The active/passive survey is more used to explain whether the surveying client associates to SSID, or just listens for beacons, and can be done both for pre and post deployment methods.

    ReplyDelete
  3. Jennifer,

    I think you mis-used some of the terms there a bit.

    An Active survey can be used pre-deployment or post-deployment. All it means is that you have an active connection to a SINGLE access point during the survey collection. With this single connection you get better data, but ONLY on that one Access Point.

    A Passive survey, on the other hand - also can be used pre-deployment or post-deployment and it collects data from ALL access points, not only single one at a time. This allows the WLAN Professional to see co-channel interference, something that is missing from the Active surveys.

    I like to say you MAY want to do an Active survey, but you MUST always do a Passive survey or you'll miss all the CCI, and only have a sparse data collection of ONLY the AP's in the Active Survey.

    Validation surveys can and should have Passive and Active if you've got the time and money. Pre-deployment surveys need to have Passive and Active as well.

    Keith Parsons

    I've got a white paper with more detail on conducting Passive and Active Surveys here: http://wirelesslanprofessionals.com/wp-content/uploads/2010/06/Seven-Rules-for-Accurate-Site-Surveys.pdf

    ReplyDelete
  4. I meant to include a paragraph on an 'active survey' of an existing deployment. This has been described to me as a survey where you associate to an SSID on the existing WLAN and gather information on that WLANs coverage area. You could also do additional tests once associated/authenticated to the WLAN (throughput etc).

    The number of times where I've had the ability to associate to an existing WLAN and document it's coverage (outside of an open guest WLAN) is pretty small. Most often the customer has more than one SSID they want mapped. In those instances I do a passive survey and collect data on all the SSIDs that can be detected by my survey laptop.

    If my pre-deployment survey is the ap-on-a-stick variety, and I've only got one ap to survey with, how would you go about doing a passive survey of the ap-on-a-stick? Usually when I'm doing the ap-on-a-stick survey, the customer is replacing the existing infrastructure. Therefore a spectrum analysis is sufficient for documenting rogues/interference.

    ReplyDelete
  5. I'm not too into doing full Active Surveys to an SSID in a post-deployment phase because your adapter can only be connected to 1 AP at a time. This means that as your walking around you might be gathering heat-maps of the AP your connected to at -75dBm (before you roam) while there are stronger signals available.

    Doing a passive for a complete RSSI, CCI and ACI picture usually works for me as well as a spectrum analysis and some static performance testing.

    As far as pre-installation surveys go, I rarely do a passive to gather CCI/ACI info from neighboring signals. As far as I"m concerned, there is not much you can do about it and you should deal with it in a post-insatllation phase because the environment can change between pre and post. As Jennifer said you can get enough info from a spectrum analysis to know how usable the freq band is.

    Justin Cetko

    ReplyDelete