Location restrictions are among the most common constraints, usually due to aesthetics or due to the inability to run Ethernet cabling to where you want the APs. Fortunately, there are some useful tricks:
Directional antennas: Most indoor APs come with internal omni-directional antennas, which are generally good for both aesthetics and MIMO performance. When faced with location restrictions, it can be useful to select AP models with external antenna ports and connect directional antennas. This improves both the transmit and receive signal strength in particular directions and minimizes the reception in the other directions. While this approach can be slightly more expensive with the added cost of directional antennas, it can make for a well-performing WiFi network when mounting locations are quite limited.
Wireless backhaul: While running an Ethernet cable may be cost-prohibitive, electrical power is often readily available. In such cases, a wireless backhaul solution is often viable. Most AP vendors support mesh networking to interconnect APs wirelessly. While mesh networking compromises both throughput performance and client capacity, since the WiFi radio needs to support both client devices and wireless backhaul, it can be adequate for non-intensive applications. For performance-critical applications, using additional APs as dedicated wireless point-to-(multi)point links can provide both performance and flexibility, as the point-to-(multi)point link acts like a “wireless wire.”
There also are ways to tackle the constraint of a limited budget. A tight budget almost always leads to a lack of access, as nobody wants to pay to send a WiFi engineer to a site and perform pre-deployment site survey measurements. Predictive modeling (i.e. heat map) software, such as Ekahau or Tamograph, are useful in these cases to calculate how signal will propagate, but this approach has its limits. Predictive modeling is generally based on simplified assumptions about the environment.
Nonetheless, predictive modeling is often good enough, provided you have ample information about the site, such as floor plans to scale, wall and floor materials, and presence of other RF systems, etc. A predictive model coupled with a lack of (or inaccurate) information will guarantee a poor WiFi deployment.
Limited budgets also mean using less expensive APs, but while lower-end APs tend to have fewer diagnostic capabilities and lack certain advanced features, this isn’t actually a problem if those features are not required. Unfortunately, most installers take a backwards approach: Instead of selecting the right AP technology for the job, they are locked into their preferred AP vendor, thus forcing the AP they know into the job they have.
The most successful managed service providers are knowledgeable in the product lines of a few different AP vendors. One strategy that has been implemented in numerous hospitality deployments and other verticals to meet budget constraints is to actually mix-and-match APs based on the application, defying both conventional wisdom and every AP vendor’s advice. In such cases, deploy high-end APs in the conference areas, which are driven by capacity requirements to support a large number of clients in a relatively small area, while deploying lower-end APs on the guest room floors, where the requirements are driven by coverage to provide service to a few rooms per AP.
Finally, channel and transmit power settings are entirely within your control. Unfortunately, most AP vendors overly hype their auto-channel and auto-power features, and installers are happy to cede control over these two critical design parameters to overly complex and poorly performing algorithms. Such algorithms are not part of the IEEE 802.11 standard, and thus proprietary to every AP vendor. Unsurprisingly, some vendors perform better than others. This proves to be a very complex optimization problem for a computer, so no vendor has or ever can be perfect.
Even for AP vendors who do this comparatively well, such algorithms typically perform good enough in simple environments, but are overwhelmed in highly complex and over-constrained environments. Turn these features off and establish a static channel and transmit power scheme across your network to minimize self-interference. One can also disable the occasional 2.4 GHz radio on dual-band APs to ensure that there isn’t too much coverage on 2.4 GHz while providing sufficient coverage on 5 GHz.
Unfortunately, most WiFi deployments are over-constrained, forcing engineers to take shortcuts. Still, that's not an excuse for deploying poorly performing WiFi networks. If you understand the requirements of your project and are willing to employ creative solutions, you can create a design that adequately satisfies the requirements for highly over-constrained projects.
Get live advice on networking, storage, and data center technologies to build the foundation to support software-driven IT and the cloud. Attend the Infrastructure Track at Interop ITX, April 30-May 4, 2018. Register now!