Atheros driver support

[lonnie]

We have been testing the available code for Atheros cards and we are not thrilled nor impressed. So far we have been unable to get Atheros to talk with us, so we need to try another tactic - throw some money at them.

We understand they will license the code and even provide reference designs for cards if we buy a license. I have not contacted them about this yet, so all I know is the "talk" I hear from others, and that puts the price of this at between twenty and thirty thousand dollars.

[bobbyc]

Is the code that you have the same code that MT has? Or did they most likely throw money at atheros a while ago?
The 1 or 2 references I've seen indicate the atheros cards work great with MT.
There's some pretty good atheros cards out now, available from YDI and eje at fament. Atheros had this article a few weeks ago announcing a newer chipset with better Receive sensitivity.
http://www.smallnetbuilder.com/News_story_328.php
Sure would be nice to have support when those come out of the production line.
Bob C

[rasimoes]

The Atheros driver will have distance tunning, like Agere?

[lonnie]

We will have the distance tuning if we can get the source code right from Atheros. If we have to use the BSD code then, no, the distance will be be fixed, and I am not sure what that would even be.

I am surprised since this has been such a hot topic that very few people are bothering to view this, and even less are answering the poll. Maybe it is not such a priority after all. Maybe you guys only want what is not there. :cry:

[bobbyc]

I'm suprised too. With good drivers, this could be an awesome chipset. Even for those WISPs who don't want to delve into 5.xghz, the 2.4ghz AP/client with distance settings would be great. It would pick up where the Prism cards stop. Take the NL-5354MP AP card for example:
Power : 802.11a 13-18dB (depending on speed)
802.11g 15-20dB (depending on speed)
802.11b 20dB (no matter what speed)
Sensitivity :
802.11a :
-84dBm @ 6Mbps -77dBm @ 18Mbps -70dBm @ 48Mbps
-82dBm @ 9Mbps -75dBm @ 24Mbps -68dBm @ 54Mbps
-79dBm @ 12Mbps -73dBm @ 36Mbps
802.11b :
-91dBm @ 1Mbps
-90dBm @ 2Mbps
-89dBm @ 5.5Mbps
-87dBm @ 11Mbps
802.11g :
-91dBm @ 1Mbps -84dBm @ 6Mbps -75dBm @ 24Mbp
-90dBm @ 2Mbps -82dBm @ 9Mbps -73dBm @ 36Mbps
-89dBm @ 5.5Mbps -79dBm @ 12Mbps -70dBm @ 48Mbps
-87dBm @ 11Mbps -77dBm @ 18Mbps -68dBm @ 54Mbps

-87@11Mbps in 80211.b isn't bad at all
Bob C

[georgew]

This is extremely important for all of us, but getting people to put their money where their mouth is will never be an easy thing.

We need the bandwidth and the channels, not to mention the cleaner frequencies. 2.4 is not usable in town. This technology only works in the woods, we need the 5ghz stuff to have enough spectrum to operate.

[dkii]

I would donate $200+ and pay extra for a license with atheros support. After all it doesn't seem fair if I/we pay $200 and everyone benefits, even those that did not pay, and at the same time, it wouldn't be fair for alot of people to pay extra for a feature they wouldn't use.

That would be ideal, however, right now I don't think there is anything, except for a few things I'm not going to mention, that I wouldn't do for .11a/g support. I would have gladly sent you the $1800 I just paid trango for a link10, and used star-os instead, if it had 5.x support.

[mkolbo]

I would be willing to pay an extra fee per Star OS license to get the Atheros drivers. Sort of like having an unlock code for that feature, or pay extra on top of the base license in order to get the new driver.

[lonnie]

It is looking like we will have a special, extra cost license for Atheros support. BUT hey, we'll wait until we have all the responses before we make any decisions.

[sure-data]

I'd be willing to pay a one-time fee instead of a fee every time.
However, a little on each license would be OK as well.

[WISP]

I would be willing to pay extra to unlock the Atheros driver as needed per edition.

Utilitizing the Atheros driver as needed for specific critical applications like backhauls would be most excelent, but I couldn't see always paying for it when it wasn't needed.

[georgew]

I would be willing to pay extra to unlock the Atheros driver as needed per edition.

Utilitizing the Atheros driver as needed for specific critical applications like backhauls would be most excelent, but I couldn't see always paying for it when it wasn't needed.

The problem is StarOS needs to scrape together $30,000. If people only pay when they use it, then the price to use it will be very high, and the higher the price the fewer people use it, and the fewer that use it the higher the price untill the price to use it is $30,000.

The right way to do this is to raise the price on the base product. If you expect to sell 10,000 licenses in the next year, then raise the price $3. If you expect to sell 1000, then raise it $30.

Pay as you go is a nice theory that is fair to all involved. However the laws of supply and demand are such that they don't always support fair pricing.

On the other hand, if you artificially fixed the price to turn the feature on, that would work... the problem is the price may have to be steep, because only a small subset of people will be using it on a subset of their routers. I just hate to see the product go the way of MT, where the router is a good price, but simple features to allow you to use it are expensive. MT's pricing was enough of a turn-off that I never even tried it. So from the point of view of my personal preferences as a potential customer (pretending I'm not one already), the right way to implement cost recovery of a new driver is to increase the price of everything you ship by a small number of dollars.

For those that think they don't need it and it would be unfair, keep in mind that sooner or later you are going to wish there were more than three channels. As soon as a competitor arrives in one of your markets, you will need a solution. Whether we admit it or not, having more channels in the future is a nice feature we all need. I have short 1 block long links that only work on one channel already. As soon as that channel stops working, I'm out of business in those areas. As this technology finds more and more use, we are going to see more and more of these situations. As soon as it happens to an important cell, not having this feature is going to be very expensive.

The trend is clear, wifi is increasing marketshare expotentially right now. If you do the math, you will see that we can expect the 2.4ghz band to become less than useful soon. So even if you think you don't need the extra channels, you probably will soon. And frankly we will eventually need a lot more channels. I don't see the 5ghz as a solution compatable with the trend. We are going to need a new technology very soon, something above and beyond the current bands.

There has been talk to letting the public use unused TV channels. This would be ideal as these channels are going to do nonlos!

[sure-data]

So, is it $3 or $30 on the increase?

And, since you need the money for the license before you get the code started, when will the increase take effect?

What's the development time?

How are you going to handle "upgrades"? Do we pay a one-time-fee to jump from version "a" to version "b" with no up grades after version a.1.2.3. (StarOS just refuses to upgrade without a new key)?

You are right about the spectrum crowding problem. We are always going to need a bigger pipe. Right now, 5GHz is about the only place to go.

Locally, we have met with private users: schools, hospitals, etc and worked out a channel sharing arrangement. They understand that we get 1 & 6 vertical, they can have 11, 4, and 9 horizontal. They also understand the paragraph in Part 15 about accepting interferrence :wink: . Right now this is working.

There is also the "let's be friends" approach where you tell them: you take YOUR equpment down, move it outside of MY coverage area, then we link the two systems together and now we have a bigger footprint. Then, we can wholesale to each other.

Just let us knowabout the upgrade surcharge so we won't have sticker shock.

Thanks

[lonnie]

We have not made any decisions yet. The first thing we will try is the madwifi drivers and see what they do in the field. They might standup pretty good and we can always help them troubleshoot and make them better.

If we have to spend the $$ we will, and we will come up with a way to charge the fees fairly. I will probably make a special product that has the added drivers. As nice as it sounds to charge everybody a little bit, it is quite unfair to the guys who will never use the new driver code.

I am quite happy with the current price levels and it would mess some guys up who have long term plans based on what they now are getting. This is not a good time in the industry to be raising prices without some added benefit.

[bobbyc]

Lonnie you are too nice a guy. I am not a business major; but I strongly feel that starOS is underpriced as is... I think that with the latest additions to starOS such as PPTP server/client/hotspot/radius/etc, you could raise the prices on the basic 2.4ghz license, while having a seperate more $$ license for 2.4/5.x.... and not piss anyone off.
Bob C

[lonnie]

As some of you know our volume is growing very nicely. THANKS. :D

We are content right now. The only thing we will have to be changing is our free support to help guys re-design their systems or learn the business, but that will be another topic when I get a policy worked out.

[lonnie]

Well - we will be getting Atheros support now. It is a yearly license so that means this will not be a one time price. We will be creating an unlock key for the Atheros driver, similar to the way that the Server enables Proxy Cache and DNS.

It has been a long task to get to this point, but these guys do have the best chips, so it will be well worth it. Our first goal will be p2p and backhaul. We need these things for our own ISP use.

[usairwave]

Question....

With the MT Atheros solution, I believe you can only route between networks. You can't set up a transparent bridge without using EoIP. Do you think Star will have the same limitation?

[tony]

There will be no problems routing, or setting up a transparent bridge with our implementation.

[lonnie]

That is too early to tell for sure, but I doubt if we'll have problems and the bridging will be as easy to use as we now have it (and don't get me started on bridging). We have to wait and see the documnetation first. Then we'll do a driver.

[go.fast]

I was thinking of buying proxim mp.11a bsu and su for a couple PtP links and a couple PtMP links to take advantage of the 5.2 5.3 as well as some 5.7 5.8 stuff.
The proxim BSU is 840.00 and the SU is 400 and something.

Seeing I can save some money by using an Atheros card in my Star-OS boxes, I would think that you could easily charge extra for the OS upgrade that includes Atheros support.
We have about 20 something AP's running Star-OS and are always upgrading OS.
I wouldn't mind paying you for the additional feature. And I surely hope you have some competetive pricing on the cards .
How soon can we expect you to release a driver?
George

[go.fast]

One more thing,
I have 2 sets of 5.3 Pac Wireless 2' dishes waiting to connect up 2 sets of radios that needs to be done in about 3 weeks
Should I buy the radios or wait?
George

[tony]

We should have a working driver ready for testing a couple weeks after we get the information needed. Unfortunately this means it may not be ready within 3 weeks.

[lonnie]

I would say to wait for a bit, but make sure you have a backup plan. You might have to do this job with other gear, but there will be other jobs and we will be ready for them.

[bkehoe]

I'd definatly love to see Atheros support in StarOS. So far, I've had to get seperate pcs (i.e. I have a staros and mikrotik pc at each site i've got a 5.8 link to) and mikrotik licenses to do my 5.8GHz links. Would be more than happy to pay twice the current server license price for staros with atheros support with link distance adjustment.

[tony]

We are currently in development phase for the Atheros support. We are not sure of what features it will contain, yet (such as long distnace adjustments, etc) however these are things we will look at as we progress.


Thanks!

[lonnie]

:D :D :D :D :D :D :D As Tony said - we have the license. :D :D :D :D :D :D :D

Now comes the work. Please be patient as this will require a lot of effort. Expect that we will have some very basic drivers for the first few releases, but by the end of November you can expect to see things like long distance tweaks and a few other little goodies - and NO I will not say anything more. :wink:

[bobbyc]

These few posts are what really interested me in the capabilities of the atheros radio; I just never wanted to buy a MT AP license so I haven't tried it out yet.
http://bruno.pmi.lv/mailman/public/routeros/2003-June/002221.html
http://bruno.pmi.lv/mailman/public/routeros/2003-June/002246.html
http://bruno.pmi.lv/mailman/public/routeros/2003-August/002946.html
http://bruno.pmi.lv/mailman/public/routeros/2003-August/002954.html
http://bruno.pmi.lv/mailman/public/routeros/2003-August/003449.html


OK close this thread and don't bug them anymore :)

[teri]

Any news on drivers I have some dishes waiting to test :P :P :P
was thinking of tring MT But Ive had str since you started
I have a big network Ill throw it right in the middle and see what happens

[lonnie]

The beta release is on track for near the end of this week.

Please ...... the release we are about ready to do is beta code. We have had the driver in development for two weeks so this should not go into the middle of your LAN.

We are confident :D but we not not crazy :twisted: and we are not perfect. Please test it and beat on it where you are not going to bring down your LAN. If you do throw it in somewhere that is critical, please make sure you have a method to quickly restore to your previous system.