Forum

Devin is who he is because of passion. He is of course one of the main reasons why CWNP is what they are today, and he was faithfully vendor neutral while at CWNP.

Over the last 5 years he and I had numerous discussions of the pros and cons of each vendor. I don't recall a mention of a favorite, even in private conversation.

Devin now works for Aerohive. He isn't vendor neutral anymore but he is still objective. Was his statement marketing fluff? I'm not one to judge, but if he can back up those statements, it isn't fluff. I'm not weighing in whether Aerohive's gear is what he says it is. But he is passionate about the product which is why he is at Aerohive.

Should we expect everyone that ever works at CWNP to remain vendor neutral their entire lives? Of course not. He is as free as anyone to promote their gear with passion and enthusiasm.

GT

That wasn't my point GT.

I know Devin's heart and passion went into CWNP. That is all great. His stint there was great, CWNP has grown into quite the org and a respected testing and certification organization. And for that, Devin should be praised for his work to that effect.

It was met with some sour taste when he went to a vendor, but again, a man's gotta eat. Meaning, a person has to choose their job, choose to support family, etc. The offer from Aerohive was probably a damn good one, and he might have been foolish not to accept it.

And then everyone understands that he now speaks with the bias of a vendor (so do I of course.) My heartburn comes from the fact that I, as a person who works for a vendor, has gone out of my way to make posts that are neutral as possible--because we aren't here to make partisan shots at one another. We are here are lovers of technology, some of us just happen to work for vendors, and that's good.

So I, and probably many others here, have gone out of our way to avoid the blah-blah marketing babble-flush that comes from many vendor mouths. In fact, Devin has blogged about this in the past!

My point, and I know you want to defend him GT, is that the statement I quoted was most out of line in the forums in general (at least to what we strive to be, do and post here). Obviously, others agreed and I don't believe they were from vendors.

If every vendor were to come in here and post with marketing techno-speak and said pointly, "I am the best!" no, "I am the best!", no "We can do that and so and so vendor can not!" then every post here would be reduced to uselessness.

In my opinion, Devin should know better, and shouldn't be given a pass just because of his prior involvement with CWNP. Devin, and I hope you know I'm not picking on your personally. Just your post bro...that's it. I hope you can see where I'm coming from.

Hi bjwhite,

I dont find anything wrong in what devin claimed .
He said he is up for a test and that is how every product needs to evaluated ...

Every product wud have USP and they would be the best in that .We should leave it to the customer .


Stop pulling devin lets pull Marcus .If someone says that their product is good let them give facts to the community.If Marcus says one vendor is better let him come with a through analysis . one liners are not enough :) I wanna pick someone :).

Thanks
Wirelesswizard

Did Marcus claim any vendor was better than the others?

Meanwhile Devin said this:
"Aerohive (see banner to the right of your screen) is the most scalable, most resilient, and most cost effective architecture available in today's market. It's Wi-Fi that Works - Period."

You honestly don't think that sort of outright "we are the best" statement is out of line for these forums?

In the spirit of full disclosure, I work as a Principal Technologist for Xirrus. Rather than simply say ???¡é?¡é?????¡­?¡°we are the best???¡é?¡é???????? (which I believe), I will tell you how our solution is architecturally different than other Wi-Fi manufactures and how that benefits our existing customers. Xirrus aggregates multiple radios into a single platform (either 4, 8, 12, or 16 radios). All radios leverage a system of high-gain directional antennas.

There are two main advantages to this: coverage and capacity. From a coverage perspective, the FCC (in the United States) regulates the amount of output power for access points. A high-gain antenna focuses this limited amount of output power and pushes it out further in single direction. Therefore, by using an array of directional antennas, the Xirrus arrays have more range per radio (and coverage per device) than any traditional thin-AP system. More coverage per device translates into using fewer devices for a given area. This further translates into significant cost savings in the form of fewer cable pulls, fewer switch ports, less electrical power utilized, etc.

From a capacity perspective, each additional radio provides bandwidth to the available total. This is exceptionally important with Wi-Fi, because it is a shared medium -- the total available bandwidth is shared between all the users in a given physical space ???¡é?¡é?????¡é?€?? and more users generally equates to less bandwidth per user. Increased proliferation of wireless devices and the increased complexity of applications exacerbate the need for additional capacity.

The Xirrus solution is centrally managed, but processes data in a distributed fashion (at the edge). This means all encryption/decryption, VLAN tagging, Quality of Service, etc. happens at the edge, rather than a centralized controller (which could be a potential bottle neck). This is similar to lessons learned with Ethernet routing & switching. In the early days, organizations routed all their traffic at the core (which is similar to controller based WLANs). As the Ethernet market matured, organizations began to push more and more intelligence out to the edge switches, as it is more efficient (similar to distributed WLAN architectures, such as Xirrus).

I think there are also many business process benefits to working with Xirrus, but those are outside of the scope of this discussion. Obviously, we would also like the opportunity to prove the value of our product/company through a ???¡é?¡é?????¡­?¡°bake-off???¡é?¡é???????? or similar process. Best of luck with your decision -- I am available anytime at douglas.haider@xirrus.com if you have questions or concerns.

Respectfully,
Douglas

So, is this what this forum is going to be like in the future? An infomercial?

There is nothing wrong with commenting on a specific feature of a manufacturer with regards to an issue, it is the "selling" or unwanted solicitation that other forums don't like, and have banned.

On the one hand, I agree with cjoseph about minimizing the marketing propaganda. Let's keep it technical please. On the other hand, and this is true of Douglas' post, solutions overviews from a technical perspective are helpful. The whole point of this forum is to discuss technology, so now is your chance to probe Xirrus to find out more and to try to poke holes in their positioning. Take advantage of the opportunity to learn here.

So, since both Aerohive and Xirrus have represented here, I have some questions.

Both of these vendors claim that intelligence at the edge is beneficial since controllers can become a bottleneck. My question is (this is a real question), how many deployments actually experience bottlenecks at the controller? By not oversubscribing controllers (which have robust processors and memory, BTW) with APs and using link aggregation to prevent bottlenecks, at present, this point doesn't seem too convincing. Do you guys see this being different in the field?

For Xirrus, I have another question. Everyone is always saying that we need to balance the RF link between the AP and client so that both uplink and downlink traffic are reliable. Do you see any problems with high transmit powers and directional antennas on the AP and low transmit powers and omnidirectional antennas on the clients?

I'd like to hear some arguments, but properly designed systems don't need balanced links.

Xirrus is using static beamforming which improves the signal to the STA and also allows the STA to be better heard. Ruckus uses dynamic beamforming, but the end results are somewhat similar.

The more power that an AP / antenna can transmit, the better (barring health issues). The key is controlling that RF output to only go where it needs to go and not causing unwanted interference.

Omni antennas are terrible. I use the example of the wire. Why is the wire so good? Because the signal only travels down the copper an no where else. What if we could control the RF in the air to only go to one point? Well, that isn't being done yet, but Xirrus and Ruckus are the best at it right now.

Ruckus and Xirrus use two different methods to achieve the a similar result. Better signal, range and data rate to the STA while not causing massive amounts of RF energy to propagate elsewhere.

I have to get going for now.

GT

This is getting way off topic from Aruba, but...
In response to GT's comments about balancing the link, I agree, but I have another question as a rebuttal. BTW, I'll be the first to admit that my knowledge of RF is deficient. I'm not dave1234. :)

I definitely see that directional antennas improve both gain and receive sensitivity, but usually not equally. For example, Eric Stresen-Reuter said in the CWNP/Ruckus webinar that Ruckus beamforming can provide up to 9 dB of gain and usually improves receive sensitivity by about 2-3 dB. That's a pretty significant difference. When most APs have better radios (i.e. Rx Sensitivity) than clients, it seems to me that the link is becoming less and less balanced, so at some point, you would see clients connecting where downlink traffic looks good, but uplink is less reliable.

I'm not sure if Ruckus' directional antenna technology is the same as Xirrus' in this way. So, do fixed directional antennas provide gain and Rx equally?

Ruckus, Xirrus, and Meru (they encourage high transmit power on the AP) would all seem to have issues here, but I've not seen anything that demonstrates link balance problems for these three vendors.

Educate me, please. :)

Definitions of receive sensitivity for radios can very often be somewhat vague as to how the manufacturer defines that parameter along with all it???¡é?¡é?????¡é???¡és associated ???¡é?¡é?????¡­?¡°test set-up variables???¡é?¡é????????.

Take for instance a microwave radio link [ non-802.11 ] transmitting a T-1 carrier. From point A to point B. Usually we would not just take a T-1 and send it directly into the modulator. Some overhead would perhaps be added [ for example alarm information that would allow the technical staff at the other end of the link, point B to see if the radio at point A was indicating alarms, and vice-versa ]. Perhaps order-wire data would be added [ usually you have a mini-phone system at both ends of the link to allow engineers to talk to each other ]. This data would be perhaps scrambled [ to allow for energy dispersal of the transmitted waveform ], modulated and upconverted [ changed from say 70 Mhz to 5 Ghz ???¡é?¡é?????¡é?€?? double upconversion would be used to prevent what is known as image channel interference ]. Then amplified and filtered. The final RF signal would be radiated from the antenna, pass through [ hopefully ] free space and then be received by the receiving antenna, filtered, amplified, downconverted, demodulated, descrambled and then the overhead data would be picked off and the actual T-1 would be sent on it???¡é?¡é?????¡é???¡és way. This would be a continuous stream of data. Many radios are fitted with Bit Error Rate test sets. An engineer at point A could inject a stream of test patterns, repeated, from A to B. This would be done normally during commisioning or downtime. The pattern may actually replace some customer ???¡é?¡é?????¡­?¡°time-slots???¡é?¡é???????? or could be injected at a lower rate in overhead bits so as not to disturb customer traffic.

The engineer at station B could the read the bit error rate of the continuous data stream. With sophisticated equipment, the engineer could read values of RSSI [ in dBm ] SNR etc along with BER [ Bit Error Rate ]. This BER parameter is key in determining receive sensitivity for a system.

In the lab, the manufacturer would set up a test system and have the receiving radio receive a particular signal level with a particular S/N ratio [ it is actually a carrier-to-noise to noise ratio, but I???¡é?¡é?????¡é???¡éll stick with popular terminology ]. Perhaps he will have adjacent interfering carriers, perhaps not. The BER would be measured. The values of receive signal level and S/N [ for a particular modulation type/coding ] would be adjusted until a particular value of BER [ say 1 bit in error in every 1,000,000 bits or 1 in 10 to the 6 ] is achieved. We could then say that the receive sensitivity [ in dBm ] was ???¡é?¡é?????¡é?€??x dBm for a y Mbit/s carrier having a coding rate of z. In order to get around the whole variable coding/modulation business, in real life, we can use a parameter called the Eb/No, but I won???¡é?¡é?????¡é???¡ét get into that here.

The manufacturer would then publish tables showing the receive sensitivity.

I don???¡é?¡é?????¡é???¡ét how the Wi-Fi manufacturers do this for 802.11, as I have not read any documentation on this. Usually the physical radio itself will have a fixed sensitivity for a fixed data rate/mod type etc. This is a function of the quality of the electronics inside [ mainly due to noise caused by semiconductor impurities etc ] etc.

Usually the radio antenna is not considered. In other words, we need an antenna to pick up a signal in real life, but the manufacturer could easily inject a test carrier directly into the radio receive port without the antenna.

Dave