Cisco Cisco Aironet 3500p Access Point Libro blanco

 

 

© 2014 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public Information. 

Page 5 of 11 

Once SAgE processing has completed, the radio samples of interesting pulses are passed to the software level for 

detailed fingerprint analysis. Performing this processing on the main radio CPU would negatively affect Wi-Fi 

performance. To eliminate this impact, the Cisco hardware solution includes a custom processing core called the 

Digital Signal Process

or (DSP) Vector Accelerator (DAvE), which is integrated directly into the access point’s Wi-Fi 

chipset. The DAvE core is able to perform intensive signal processing operations, referred to as “Davelets” - such 

as filtering, decimation, rotation, sync-word detection, and modulation detection - without burdening the main CPU. 

The DAvE handles CPU-intensive signal processing operations that would otherwise be a burden to the main CPU. 

The final processing level occurs in a software module that runs on the main 

CPU and is referred to as “Sensord.” 

Note that because the heavy lifting has been accomplished by the SAgE and DAvE hardware blocks, the overhead 

on the CPU is now very low. The Sensord software looks at the timing and frequency of interference bursts, and 

the discovered attributes of the bursts such as the modulation type and identified sync words. This high-level 

information is then used to perform the final identification and separation of one device from another. This final 

classification step provides the powerful features of SI: telling you the specific source of the interference, where it is 

located, and how it can be mitigated. 

Cisco CleanAir technology includes a robust suite of 20 non-Wi-Fi classifiers. Because the analysis takes place in 

software, the list of classifiers may be expanded as new interference sources become relevant in the market. 

In other words, the solution is capable of detecting any kind of interference that might be introduced in the future, 

and requires only a software update. 

Cisco CleanAir technology classification is able to distinguish several different interferers - either of the same type 

or different types - that are operating at the same time. In fact, CleanAir technology is capable of tracking and 

classifying hundreds of devices and reporting the top 10 simultaneous interference devices per radio. This is 

important, because in the real world the amount of simultaneous RF activity can be quite high. Any competing 

solutions that are not sophisticated enough to distinguish multiple simultaneous devices will quickly fall apart in the 

field and are only good enough for demos and lab tests. 

Interference devices can be transient, either because they are turned on and off quickly, or because the user is 

moving through the floor space. For this reason, classification must occur quickly, before it is missed. Cisco 

CleanAir technology enables access points to classify devices within 30 seconds, and is often able to perform 

classification in less than 5 seconds. (Note that reporting may be slightly delayed when consolidating data across 

multiple access points). 

It is important to not miss a source of inte

rference, and it’s equally important to not report “phantom” interference 

when none exists or to mislabel interference, causing IT to look for the wrong type of device. Cisco CleanAir 

technology is designed to produce low false-detection rates, even in very busy RF environments where hundreds 

of Wi-Fi and non-Wi-Fi devices are operating simultaneously. By reducing false detection, CleanAir technology 

saves IT time.