By Vents Lacars - Sponsored by SAF Tehnika

Microwave radio systems have mostly been and in many cases still are a big unknown to majority of recent players in communication services market, especially the ISP community. Things have changed – emergence of advanced signal processing, data compression, error control, link protection, interference avoidance technologies, integrated Ethernet transport support are present in all modern microwave radios.

However, what remains hidden is not only the fact that in many cases the declared capacity values are nothing short of blatantly misleading for the users, but also the fact that in the wireless industry capacity should not be viewed separately from many factors directly affecting it. Thus, the purpose of this blog post is to unmask most common "tricks" used by certain microwave radio vendors in a technically dubious competition for "the best capacity" prize.

VP Business Development SAF Tehnika marketing@ saftehnika.com

Data transmission capacity – in all its many interpretations – has become an object of worship for many users and vendors. The reasons are rather obvious for such a competitive market – vendors can attract attention with less intellectual effort just by presenting greater numbers to customers. However, as it often happens, the devil is in the details.

First, it's very important to understand that each modern point-to-point microwave radio, no matter how Ethernet-friendly, is still in its essence a wireless signal processing device. Thus, there are at least two levels at which capacity of any given system can be defined and should be evaluated the traffic level.

Second, it is useful to understand the difference between the rate of physically transferred random bits and the bit rate of specific form of data. Considerable number of bits representing a specific data form can be encoded in the transmitter into certain order of physically transferred bits and later decoded in the receiver, which leads to increased bit rate of transmitted data at the dedicated interface port. For example, as the composition of Ethernet physical layer overhead is regulated by specific standards, most of the overhead information can be excluded before transmission and later added at the receiving terminal. Thus, the measured capacity value at the Ethernet port can exceed the capacity of the radio channel. This is the basic reason, why it is incorrect to compare two systems if you're not certain which of the various capacity values the vendor has presented in the product datasheet.

At the signal processing level the ultimate limit for technical creativity of any microwave radio engineer is a law of physics described by C. Shannon and R. Hartley, according to which the capacity of any given radio channel is strictly bound to channel bandwidth and channel noise level (or signal-to-noise ratio, SNR).

At the traffic level, for instance, in case of Ethernet-framed data, the additional challenge is to encapsulate as much useful data as possible before it is transmitted over air. Knowing these objective obstacles there are basically two ways for vendors to improve their product – we call it extensive and intensive ways.

Unfortunately, this is where things go wrong. In their pursuit of "the best capacity" prize many vendors try to maintain a very simplified information environment in which the customer is forced to compare apples and oranges. First of all, fantastic, sometimes even mind-blowing capacity values are more likely based on results of best-case laboratory Ethernet throughput tests in an almost ideal environment with favorable size frames, as well as enabled overhead and payload compression.

My suggestion is - you can start by asking, what capacity of what OSI Layer (and in case of Ethernet - with what frame size) is presented as the product capacity. If any kind of compression technology was applied, it would be also good to know which access traffic model was used for reference. There are several such in industry – all of them are based on average values which can differ significantly according to the traffic composition in your network.

Secondly, it is very common for many vendors to describe maximum capacity as link capacity. By that they usually mean the total capacity of a link or even a network of specific topology consisting of top configuration microwave terminals with XPIC (doubling capacity by transmitting signal in both polarizations) or maximum channel/link aggregation enabled, (usually, up to 4+0, i.e., quadrupling capacity of a single radio), transmitting via maximum bandwidth channel, usually 56 MHz.

Please note that all of these capacity enhancement technologies are great if you can afford them, including acquisition of necessary permits from your frequency regulator. Some vendors do not openly say that declared capacity values are valid only for 80MHz or higher bandwidth channels (obviously, they allow to transmit even more data via a single link). Just ask for estimated price of such a top configuration solution and those will just be the equipment costs. Add all the other related costs and compare them to your available budget and estimated return.

While there's no doubt that the capacity of transmitted data is an important feature of modern wireless communication devices, it should also be clear that it cannot be the only feature to be evaluated by users, who are exploring options for building wireless links. And not just because various vendors present differently measured values or play tricks with numbers. Correct link planning can not be performed without taking care of free-space path loss, multipath interference, rain fade and different other factors affecting signal strength and the transmitted data capacity. Even the existence of a nearly industry-standard Adaptive Coding and Modulation technique, which automatically adjusts radio performance according to fluctuations of the radio signal strength, should make it clear that the capacity in a microwave radio link is not a rock solid value. We've noticed even more questionable attempts to mislead the customers. Some present full-duplex throughput calculations per link by summing up (aggregating) transmitted and received capacities. For instance, if CFIP Lumina can simultaneously transmit and receive 366 megabits of data per second then link capacity of two such radios would be, according to such math, 732 Mbit/s.

And some are grouping E-band or Gigabit radios (60 GHz and above) with "standard" microwave radios (up to 38GHz including) and their obviously greater Ethernet throughput values are presented as maximum values for the whole group.

It is important to know that Gigabit or E-band radios are specific high-capacity microwave systems effective in very short distances. They utilize RF channels of considerably greater bandwidths and can be an option if you need a link covering a distance no longer than few km.

Capacity of a microwave radio system is no doubt one of the most important features – just like an engine of a car. But as in the case with cars where aerodynamics, quality of fuel, tyre treads, weight, electronic control systems and many other less important factors influence the general performance of the car itself, capacity of a microwave system shouldn't be the only factor evaluated during your search for the most fitting solution. Especially when there are several interpretations of what "capacity" actually is and its values are subject to numerous conditions, including composition of the transmitted data, point of measurement and system BER threshold.

Understanding your needs, communicating with the vendor and having at least general insight in the specifics of microwave data transmission systems are all essential ingredients of a quality choice.

Wishing all the best,
VP Business Development
Vents Lacars
SAF Tehnika