Wireless designs come with a plethora of nuances, specifically around requirements and past experiences. You can look at a floor plan and start to randomly place access points in multiple locations until the floor is filled with APs. The next logical step is to take a step back and ask, “what are the requirements”? This is when things become interesting, as each stakeholder will have a different idea of the requirements.
Gathering upfront what some see as trivial requirements is undoubtedly the first step in designing a “working” 802.11 network.
- How high are the ceilings on this floor(s)?
- Are there clean “flattened” PDF files along with AutoCAD drawings where measurements can be obtained?
- What type of ceilings does the building/floor contain? Are there reflected ceiling plans?
- Are there areas with limitations/restrictions where access points (APs) cannot be mounted?
- Are there areas that require outdoor wireless coverage?
- Will robotics be used at the site?
- Is wireless coverage expected to be ubiquitous? If yes, does this include bathrooms, janitor’s closets, elevators, maintenance rooms, etc.?
- What are the expected client devices? OS X, Windows, RF guns/scanners, printers?
- Is this a new or existing site?
- How high are the ceilings in the warehouse/distribution center? How high is inventory stacked while being stored? What kind of inventory is kept in stock?
- What is the address of the site?
- Is Personal protective equipment (PPE) required? If so, is training also required before entering the facility?
- Are there any unsecured areas in which an AP should not be mounted?
- Will robotics be utilized at the site?
- Does the site heavily utilize forklifts, cherry pickers etc.?
The bottom line is this, gather as many requirements as possible. Once the “basic” requirements are gathered, you can continually expand on them. Ask many questions upfront, as it will save time later when there is a change in scope. Trust me; the scope will change.
In this blog, I will look at how the requirements listed above can be used for two different designs. The first design is a “typical” office space, and the second one will be a warehouse/distribution center.
Both designs will touch on predictive modeling then post validation surveys.
The first set of requirements will be based on the following.
- Office space consisting of two floors:
- Second floor – 36,453 square feet
- No high-density requirements
- Primary/secondary RSSI of -67dBm with a SNR of 25
- No more than one access point in any hallway.
- 25-30 users per access point
- Cisco 9130 access points
- Client devices – primarily company-issued laptops with guest access to personal mobile devices
- Only (14) Cisco 9130 access points can be deployed
- 5GHz and 2.4GHz network (disable 2.4GHz radios where necessary)
- 5GHz – 20MHz channel width
Phase 1 – Predictive Modeling Design
Predictive AP Layout
As seen in the layout above the APs were placed inside of the offices. The design logic is to “allow the environment to help with the RF design. Regarding roaming, the walking path of users should be taken into consideration. This is where past experiences, both negative and positive, will play a role. APs should be strategically placed so that the STAs have reliable information as it pertains to roaming.
RSSI / SNR of the Predictive Model – Primary Coverage – 5GHz
RSSI / SNR of the Predictive Model – Primary Coverage – 5GHz
Note: The predictive modeling was designed with the 5GHz band as the primary network. The 2.4GHz will “fall into place”. Some of the 2.4GHz radios can be disabled as required. Design around the customer requirements, their budget or other constraints.
Phase 2 – Pre-Install Validation via AP on a Stick (APoS)
This can be debated for years; my personal thoughts are that you should perform AP on a stick (APoS) validation surveys when possible. One of the most challenging verticals to perform an APoS validation survey is Health Care, especially on patient floors. As wireless professionals, we “love” to validate our designs as it makes the most sense. However, it will be extremely difficult or impossible to access patient rooms or operating rooms in most cases. Therefore, you must use common sense and discretion as it pertains to APoS.
AP on a stick validation survey requires much forethought; again, this is one of those areas where past experiences will be the foundation for a successful survey. However, below are a few items that should be taken into consideration.
- Depending on the type of environment – Is Personal protective equipment (PPE) required? Is PPE training required? What are the PPE protocols? If you were fortunate enough to be working between March – December 2020 at the height of the covid-19 pandemic you would have a better understanding of the PPE requirement. Hospitals do not have the luxury of shutting down or ceasing their operations. I worked on multiple Hospital projects in 2020 that required APoS, and it was mandatory to wear PPE and provide the results of a PCR test.
- Will an escort be provided? Security escorts are a must when working in most environments. Ask yourself, is wireless connectivity required on Psychiatric wards? The answer most likely will be yes. In those cases, there will be areas where access points cannot be mounted due to patient safety concerns. One of the requested requirements for the Psychiatric ward was “you have to disable the LEDs on the APs.” Disabling the LED makes it difficult when performing an APoS as the LED is usually used as a system status indicator. Disabling the LEDs is also an issue once the access points (APs) are in production. Nevertheless, it is a valid requirement and should be adhered to. The security escort will be able to de-escalate potential issues that will arise on a Psychiatric ward. As a safety precaution you will most likely need a co-worker standing next to the wireless equipment at all times.
- Access to rooms – wireless professionals prefer to walk into every single room in to collect valid RF data. Valid data will result in a true RF statistics, this are 100% accurate, but there is a possibility that every room will not be accessible. Sometimes keys do not work, or in Hospitals, patient rooms will not be available. It is critical to document rooms or areas that were not accessible at the time of the validation survey. Most importantly, these items should be clearly outlined and defined in the scope of work. When the requirements are not met, the APoS data will not reflect the predictive modeling. Gaps in coverage will always be a concern to customers, especially when they see the Ekahau report. If these items are not outlined up front, you will have to revisit each room, floor or site and try to access the rooms. This will take 3 or 4 visits. Again, try to call this out up front before the scope of work is signed.
- Equipment – APoS equipment should be lightweight, and Power over Ethernet (PoE) equipment should last between 6 to 8 hours. Many vendors offer kits that are optimized for traveling via air or land. You can take a look at https://wifisurveykit.com/ or https://wifistand.com/.
Note: The items above are not meant to be exhaustive. You should find a solution that fits your needs. As I mentioned earlier, everyone will have their own experiences and opinions. There is no one correct way over another, multiple solutions that work exist.
Phase 2 – Pre-Install Validation via AP on a Stick (APoS) – Part 2
After you come up with a plan and procure the equipment, the next part is the best. You get to validate your design against the customer’s requirements. As mentioned many times in this blog, find what works best for you and stick to that methodology.
One other essential item that I missed in part 1 is scheduling. You must schedule the site visit one to two weeks in advance. If you show up ready to perform the validation survey without notifying your point of contact, you will not be able to complete the survey. Your project manager is responsible for scheduling, so make sure that the person aware the scope of work.
Once onsite the next step would be to set up the tripod and AP in the first designated area on your design. The aim is to walk around the access point. You should walk as close to the AP as possible, the adjacent rooms and up to 60 feet if possible. The further you walk from the AP, the more accurate the data regarding RF propagation, SNR, data rates, throughput, etc. If you walk five to ten feet away from the, then move it to the next location your data will NOT be accurate. The AP will show less coverage that is actually being propagated.
If you are required to collect uplink and downlink data, you can set up a local iperf3 server. The server can be wired or wireless.
Both devices should be on the same SSID. In my opinion, it is better to do an internal iperf test between the two devices for the APoS. Then, once the network is up, you can perform additional test.
Another critical aspect of the APoS survey is “freezing” the access point. The logic behind freezing the access point is simple and can be found using the link below.
Reference from Ekahau.com
To simulate a complete network with a limited number of access points in the pre-deployment survey phase, use the Freeze Access Point tool. The Freeze Access Point tool allows a single access point that has been placed in several locations to be treated as multiple access points. Using Freeze AP, you can individually visualize and analyze each location where the AP was placed. If Freeze AP is not used, you would simply see one AP detected, and be able to visualize one huge coverage area. Moreover, it allows simulating a complete network in terms of overlap, channel interference, and data rate, by using just one or a few access points.
Note – by default, when you freeze an AP in Ekahau, it will turn blue. Below I will compare predictive modeling vs APoS.
A Cisco 9130 access point was converted to an “autonomous” AP / embedded wireless controller.
The settings on the 9130 were configured similar to the Cisco Wireless LAN Controller (WLC)
Min – 11dBm – 12.5mW
Max – 14dBm – 25.0mW
The APs in Ekahau were also configured for 25mW
Ekahau Predictive Modeling – AP 10
Ekahau APoS – AP 10 – you can clearly see why pre or post validation is critical.
Ekahau Predictive Modeling – AP 12
Ekahau APoS – AP 12 – Confirmation is a beautiful thing.
Ekahau Predictive Modeling – AP 2
Ekahau APoS – AP 2
Ekahau Predictive Modeling – AP 8
Ekahau APoS – AP 8
Full APoS Survey – RSSI Primary Coverage -67 – 5GHz
Full APoS Survey – SNR – 5GHz
Full APoS Survey – RSSI Secondary Coverage -70 – 5GHz
The takeaway is – validate, validate, validate via AP on a stick (APoS) when possible. When APoS is not possible, you MUST gather as much of the requirements upfront and design to them.
- Clean floor plans
- Clearly defined areas that require wireless coverage
- Clearly defined areas that do not require wireless coverage
- RF configuration requirements
- AP model
- Number of APs
- Wall types, inventory, equipment, etc.