Approaches to planning and designing a developed Wi-Fi network

As practice shows, in typical Russian reality Wi-Fi is usually referred to as a simple and undemanding technology. And the most common approach is to estimate the required number of access points by eye, order, and then understand ...
Unfortunately, the results of such an approach are deplorable, and even with high-class equipment one can observe extremely unstable and poor-quality results in the network service.
On the other hand, the requirements for wireless access are constantly growing, the range of heavy services is growing, which can already be qualitatively transmitted via Wi-Fi or provided using Wi-Fi. Much of the projects already require the development of "capacity", and not "on coverage." Low-power terminals, such as smartphones or RFID tags, are required.
All this leads to one thing - it is necessary to pay serious attention to such an important part of any wireless project as a field survey of the object (Site Survey). Usually this does not cause issues for 2G, 3G or WiMAX networks, but Wi-Fi - “it's simple” and “you shouldn't bother”, and then - “oops, the network works crookedly”.

Start with analyzing your current wishes and discussing the future development of infrastructure within the company. It is better if this is a “brainstorming” of the leaders of the IT department, marketing, operations, etc. Otherwise, if the decision is made only by "techies", then after launch it often turns out that after a couple of months, the management or "sellers", seeing an article about Wi-Fi capabilities somewhere, want to have it themselves, and the network is designed for other requirements, and it all starts over again (including a new facility survey, design, search for places and ports for additional equipment, etc.).

Sample discussion questions:
- at what frequencies will the solution work?
')
- what are the conditions for deployment? (floor plan, features of building structures, ceiling height, etc.?)

- Is work expected in a complex radio environment? (a lot of interference, for example, in machine-building workshops, etc.), although this is necessary for checking during the radio survey of the object;

- how many users (of which active - simultaneously consumed services) are expected on the network?

- how will users be distributed over the network? (It is important to consider the possible zones of concentration and the expected number of users in these zones)

- what services should be provided to users at the start of the network and how it will develop in the foreseeable future, for example:
... Internet access (at what speeds per user, minimum requirements for a strip on the border of a cell),
... access to local resources (to which resources and at what speeds),
... voice services (VoIP over Wi-Fi),
... streaming video with multicast,
... location for Wi-Fi clients and / or indoor RFID tags,
... building a developed radio security system,

- what client equipment will be used on the network
(laptops, tablet computers, smartphones, barcode scanners, RFID tags, etc.).

- what are the requirements for user authentication and what are the most appropriate methods to use and in what cases? (for example, second level 802.1x or third level authentication options using, for example, a web portal)

- Is guest access required and what features should be considered here? (for example, flexibility in approaches in managing guest accounts and the widest possibilities or maximum ease when creating new accounts, etc.)

- Are there any requirements for aesthetics in the room where the Wi-Fi network will work? (This is important for understanding whether it is necessary to use only points with internal antennas or, if external antennas are mandatory, then which antennas are best chosen from the point of view of aesthetic perception).

After agreeing on network requirements and collecting all possible input data, the next step is worth taking (not mandatory, but which will greatly facilitate future work):
- perform assessment planning using any software package that allows you to take into account floor plans, configurable building elements (walls, doors, windows) with variable attenuation coefficients of the corresponding radio waves, parameters on cell borders, types of equipment and antennas used, services on the network and etc. Such a tool allows you to build a model that is exactly as accurate as the input data. But, unfortunately, the absolute accuracy of the input data cannot be ensured by definition, just as it is impossible to lay down the state of the radio environment and conditions for multipath propagation, attenuation of signals, etc. Therefore, the results of the work of such a tool should be treated with caution.
The question immediately arises - then why is this step necessary?
Answer: the use of such software allows you to do the following:
1. To form a fairly qualitative assessment of the required number of access points and, often, the corresponding beautiful proposals. But this is an estimate, not final values!
2. Create an ideal model for placing access points directly on the map of the future coverage area of ​​the real network. This data can be used very effectively as a basis for conducting a field radio survey. You can place real access points directly to the places indicated in the preliminary design on the map, take the current readings and tailor the ideal model to the real one.
Such a software module is integrated, for example, into Cisco's well-known WLAN control systems: WCS / Wireless Control System and the new Cisco Prime NCS / Network Control System. By the way, these systems are in demo versions (full-fledged system), available for 30 days. You can download from the Cisco site.

Here, it is already possible to discuss the topology of the transport infrastructure to support the intended topology of the Wi-Fi wireless network, as well as the presence of data network ports (better with support for 803.3af / PoE-Power over Ethernet) in the coverage area and 220V sockets if ports are not available everywhere with roE.

Further, already prepared enough, you can proceed to the field radio survey. Here, in addition to the information described, it is necessary to use a spectrum analyzer for the frequency section under consideration and special tools (software + radio unit with antenna) for performing radio surveys. Immediately make a reservation that it is extremely important to perform not just a passive survey (Passive Site Survey) with the collection of information about the signal level, etc., but to do an active survey (Active Site Survey), in which data traffic is constantly transmitted and readings are taken. really achievable data transfer rate in a large number of coverage points. IPerf is often used to generate test traffic. Thus, a real picture of speeds is obtained regardless of the signal level, which is very good, because linear dependence does not always work - the higher the signal level and SNR, the greater the MCS (Modulation & Coding Schema) and the higher the achievable transmission rate for end users.
The survey toolkit produces many companies, such as AirMagnet (absorbed by Fluke Networks), Ekahau, etc ...

Important nuances:
- during radio survey it is important to use exactly those access points and antennas (if with external antennas) that will be used on a real network,
- during radio survey, it is important to perform tests using the worst user device (by radio characteristics), which is expected to be seen and served on this network. Often these are smartphones or RFID tags.

After completion of the radio survey is performed:
- the final design of the Wireless Network,
- the final design of the transport infrastructure,
- the final design (or refinement) of the central site of the entire network,
- analysis and design of the integration of the developed solution with existing systems,
- recommendations for the future deployment of the solution,
- the quantity of the necessary equipment is specified and it is possible to proceed to a reasonable order.