WAR vs wrap [Archive] - Valemount Networks Support Forums

lonnie

08-18-2005, 09:30 AM

I wouldn't say the trend is away from that model but you have to realize that I get tired of dealing with the same old issues related to a customers choice of motherboard or radio. We know this is not an easy task since we have done it lots. The result is that we always offer the items that work on our website.

People want more reliability and performance and that is hard to do with run of the mill PC architecture. We build remote Solar Powered repeater sites here since that is the only way to get our network expanded, thus one of the most important issues for me, personally, is small size, low power and high performance.

Can you blame us for moving from the PC and WRAP architecture when the new WAR does EVERYTHING I want in hardware? VxWorks is not as easy toi use as Linux, thus it took a long development time, but it is smaller, faster and has some features that makes it more powerful.

We have just announced the official beginning of v3 and that should signify that we are continuing to support your choices.

It will be a huge step forward but we will use other gear here. It will be WAR boards everywhere, and eventually (I hope soon) even faster boards so that I have to swap everything out yet again. It is progress and technology and we love them both.

lonnie

08-18-2005, 09:34 AM

One HUGE reason for using the new equipment as AP and Client is the 5 and 10 MHz channel widths and spacings. At 5 MHz bandwidth you have more raw throughput than your 802.11B radios but you have 11 non overlapping channels with OFDM modulation for improved interference and multi-path rejection.

Imagine a tower with 11 Access Points in 2.4 GHz, each one connecting 30 to 50 customers and then you see why the new equipment will save your business.

With that sort of sectoizing you can really begin to reuse channels. It is tough to map out 3 channels and keep them separated, but with 11 channels to choose from you can have your directions well chosen to not interfere with your other radios.

Skaught

08-18-2005, 12:41 PM

There are many ways to approach this. I presume from your post that you build networks first and fix problems later. Often i have found success in designing the network based on the equipment i want to use. often a feature/platform/os looks good just on paper.

I have been building wireless networks since the mid nineties. I have learned you can never build a network around a specific product as they tend to be discontinued after 2-4 years leaving you in a lurch.

tog

08-18-2005, 01:02 PM

QoS for VoIP has become very important to me and I'd like to put a WAR board or two up in place of existing WRAP access points almost entirely because of that, but I cannot until StarVX at least supports WEP.

With WEP I'm at least wearing a fig leaf. Without WEP or WPA my AP is naked.

For me to put up an AP in most places I seriously only use:
Atheros support
bridging and static routing
WEP or WPA

Also, for many isolated low-cost microcells I put up a DHCP server, but for the couple of places I have in mind I do not yet need that.

The fun will begin when I get WEP ;)

And what fun it will be, I bet burning through some foliage here in the jungle with 4X cloaking will be neat. I wonder if that would harm the transmitter on those high-powered Ubiquiti cards.

Skaught

08-18-2005, 01:58 PM

Imagine a tower with 11 Access Points in 2.4 GHz, each one connecting 30 to 50 customers and then you see why the new equipment will save your business.

To do this I simply have to find a <$300 CDN Star-os CPE solution. The lowest I think I can get it down to right now is about $380 and that will limit the number of installs I can do quite a bit. Right now I am paying $209-240 per entire install plus cable and mount.

At 100 installs a month that $80 equals $8000 monthly that I would have to make up from somewhere. Right now my clients pay $300 for hookup and I cannot raise that.

lonnie

08-18-2005, 04:37 PM

I'll see what we can do. I don't try and predict future pricing until we get closer to release. Our hardware platform is quite a bit more costly than the base that others are using, so hopefully our features will justify a bit of a premium.

We'll have increased static routes (250) and WEP by early next week. RIPv2 is already there. The release is in testing now. We have one leg of our backbone all on WAR boards now and they are taking many, many GiGs of traffic with no issues. We feel confident they are solid.

palmczak

08-18-2005, 08:11 PM

Lonnie, I can appreciate the amount of support you provide. Often times you offer much more support than the hardware manufacturer(s) who really own the problem.

I am a small-to medium size WISP. Approx 450 wireless year round clients and another 250-350 seasonal clients. I live in tourist town, our county population is only approx. 9000 year round residents. There are more than 18,000 second homes and hotel beds that is the reason for the seasonal number. Being 1/3 of my company and the whole technology division, supporting the numbers I have is really more than I can do, I can honestly tell you, you are preaching to the choir when you talk about "the same old problems" After 3 consecutive 6AM calls from one customer with the same Hotmail problem I diabled his account refunded his money and called my competitor and warned him!

I do not blame you, not one bit, and the new channel width/spacing is a great feature, but as Skaught pointed out that shiny CPE in the StarVX wrapper is not affordable (at least for me), when we have to compete with $50 cable and DSL modems, not to mention companies that have bottomless marketing budgets.

I also love progress and technology. I am impressed with the claims of the StarVX as well as the vision for it.

Just how compatable will the future versions be? Will a V3 x86/wrap get the atheros features and the channel spacing/width?

Lonnie & Tony I want to say thanks for a great product and I can can appreciate your efforts. I will continue to use your products and will use The VX platform as I see it fit in to my network.

Joe

lonnie

08-18-2005, 10:22 PM

The v3 will be an advancement over v2 but it will not have the features that we have in the StarVx driver. We are reserving those advanced features for our very own hardware/software brand.

Don't rule us out for the CPE platform. It all comes down to volume and if we can hit the magic number we will be in the running for the CPE market.

The mistake being made in trying to get the $50 CPE is that you lack the features to control the system. If you cannot control what is happening and you cannot deliver the speed and reliabilty customers want then a cheap CPE is hurting your business. Any of the cheap CPE units we have ever tried have ended in disgust and failure and have been replaced with our own WRAP units. Sure they cost money but they do have a nice payback. Once the cost is recovered you want less trouble and the cheap units never give that.

Just a thought to keep in mind.

tog

08-19-2005, 04:57 AM

In general the CPE's capability/features has not been too important as long as my AP can handle things and show me what's going on. I use really dumb D-Link bridges with nearly worthless management and it doesn't bother us much and it provides us with a really cheap (Atheros even) CPE that simply bridges and can do WEP and WPA.

ninedd

08-19-2005, 10:20 AM

..a really cheap (Atheros even) CPE that simply bridges and can do WEP and WPA.Hi Tog. What Atheros based model have you had the best success with? We have a small trailer park we'd like to offer a lowest possible cost micro-pop solution for.

Skaught

08-19-2005, 11:54 AM

For me, bridging, ethernet surge supression, +23dbm, >15dbi antenna and RTS are mandatory in a CPE.

Some very nice to have are:
802.11e
adjustable ack timing
watchdog
CBQ
A utility to configure it that can traverse routed links

palmczak

08-19-2005, 05:10 PM

The v3 will be an advancement over v2 but it will not have the features that we have in the StarVx driver. We are reserving those advanced features for our very own hardware/software brand.

That is what I was afraid of. I have been locked into to many Propretary networks. Not going there again.

You might as well change the name to CanobreezStaroxim.

Thanks for being honest, that does set you apart from many other manufacturers.

lonnie

08-19-2005, 05:49 PM

Skaught

08-19-2005, 06:23 PM

Thanks for being honest, that does set you apart from many other manufacturers.

Definitely! I have been going in circles with a certain company for some time now and getting any information is nearly impossible and when you do it is often incorrect.

VNC will not blow smoke up your ass.

ninedd

08-19-2005, 10:15 PM

That is what I was afraid of. I have been locked into to many Propretary networks. Not going there again.I don't think you have to really worry. At no point did Lonnie say that StarOS would not work on standard hardware, nor did he even say that StarVx would not work with standard clients. All he said is

1) that the new StarVx hardware/software would ALSO have some EXTRA proprietary features that you COULD use if you wanted to take advantage of them, and

2) that not every single StarVx feature would be possible in StarOS ver 3.

I understand your concern, I also wish all the new StarVx features could magically be in my existing StarOS AP's that are in the field. Frankly, I also worry that there won't be the time or resouces to develop both StarVx and StarOS at the same time (in which case, StarOS will probably suffer). However, nothing to panic about I don't think.

a) StarOS v2 works well
b) StarOS v3 is in development for WRAP/PC Hardware.
c) StarVx has extra features, and is still standard.
d) StarVx also has proprietary features if you choose/need/want them.

it's all good...
- Todd

palmczak

08-21-2005, 12:04 PM

<RANT ON
The concept that StarVx is more proprietary than StarOS is simply not logical.
RANT OFF>

How can you say this? Everything, from the switch to the VX platform to the new StarVX only features, means more proprietary than StarOS built on Linux. The simple fact that the VX Platform is only available preloaded on the Gateworks boards from you = more proprietary!

In itself, this is not bad. However, it is a deviation from the old practices, and represents a fundamental shift in the business model of VNC.

v2 works very well and we will make v3 work even better. At no time have I EVER promised anyone that we would add every asked for feature.

Correct, but we never new that there would be 2 separate, non compatable platforms until the VX announcement. I believe that I, along with many others, were under the impression that while VX was taking priority the StarOS/x86 platform would get the features at a later time.

The point I want to make is this. The single most attractive thing about StarOS was the ability of the network owner/operator to choose his/her own destiny. The simple ability to choose virtually every component, based on what works for the operator, is extremely compelling. If a simple supply chain issue arises and Wrap boards are out of stock, the choice to use ITX or Soekris is available and valuable. I have several Waverider sites and I can't tell you how many times I have not been able to get CPE's when I need them. I bet most other WISP's have this problem as well. There is no "one size fits all" solution. StarOS gave us WISP's immunity to vendor/supply chain issues. Since many of us cannot afford to stock pile hundreds of parts for the day we need them, the old StarOS model was a better fit accross the board.

I am no shrink, but I believe the audience that views this forum and uses StarOs are "hands on" people. They flash their own cards, climb towers, choose antennas carefully and enjoy, and take pride in doing so. What StarVX represents is an exceptionally capable system, that offers no or very little choice. (especially if you use the extra features)

Do you GUARRANTY there will never be a shortage of StarVX components? if the answer is no then how can we be expected to base a whole network on it? As I stated before, I will continue to use Star products as I have a great admiration for them, but the areas where the VX platform fits into my network are few at this time.

Joe

lonnie

08-21-2005, 06:23 PM

Did it ever occur to you that we have some features BECAUSE of the VxWorks? Perhaps we cannot bring some things to Linux.

There is no problem if you choose not to use the new Vx platform. That is EXACTLY why we will have the v2 and v3 platforms. You will have your choice of 4 platforms from us, if you consider the CPE version to still do a job. We use it to this day and many of our customers do as well. So, with 4 platforms I think we are giving some options.

The hands on guys can go on choosing their hardware and their software. We will carry on developing but don't rag on me if I choose a better platform than you have access to. Don't get on my back if that platform performs better (why else would I choose it?) or has some unique features.

What more can we do? I'm fresh out of miracles.

tog

08-22-2005, 02:05 PM

Inet2000: Using D-Link DWL-G810

With a D-Link PoE adapter you can put them inside rootennas and have yourself a nice cheap CPE. Not manageable, but they work nicely for cheap residential accounts. They even allow you to force the transmit rate to 11mbit and enable the atheros Super-G features and even turbo (which you don't want to be doing outdoors)

I like everything you're doing Lonnie, and as always, we've got StarOS v2 which is here right now and works fine. Did I mention I'm hurting for some QoS for voip, though? :)

The FCC has given us small US ISPs a good non-stop rogering, we need stuff like voip and other things besides just providing access in order to survive.

mp3turbo

08-23-2005, 12:49 AM

Lonnie,

I would be interested in knowing possible Atheros performance delta for StarVx and StarOS on WAR and x86 platforms. x86 CPUs are fast enough, so this question is basically about features tuning etc - what are your expectations?

And what about performance difference between v2 and v3 on the same x86 platform?

thnx, mp3turbo.

lonnie

08-23-2005, 01:43 AM

x86 and a general purpose OS like Linux gets what it gets because it has to have code to handle every possible hardware combo.

The IXP-420 and VxWorks have a single hardware platform and a single radio type (Atheros) so there is no glue or compatibility layer. Every layer you can strip from the code increases performance and reduces size.

x86 is not nearly as efficient as the IXP-420 and consider that the 4 port clock is 2X the WRAP speed. Then you get into power drain - the WRAP with 2 cards draws about the same power as the WAR with 4 boards.

I hate to say it, but the WRAP is on its way out. The IXP platform is superior in every respect and once we get some volume going it will be cheaper per radio and mbps as well.

If you want to continue to use the WRAP board I have no issues. You can buy them directly from PCEngines and load StarOS v2 or v3 when we release it. You have choices and we are not holding you back.

Mark

08-24-2005, 03:34 AM

The v3 will be an advancement over v2 but it will not have the features that we have in the StarVx driver. We are reserving those advanced features for our very own hardware/software brand.

Just a thought to keep in mind.

Just when I had great hopes, I read you're going to screw us all big time.

I have been a VERY faithful user and loyal to you guys.

I've put every dime I had into star-os hardware and software-wise and now...

"Go screw yourself, we're going to obsolete everything you just built on".

Thanks for nothing, Lonnie.

lonnie

08-24-2005, 08:42 AM

Mark

08-24-2005, 10:38 AM

A wee bit of an over reation, don't you think? Just consider that v2 has all of the features you NEED, and most of the features you WANT. It is solid and inexpensive. If you want more then we will have v3, which will borrow some code from the StarVx and will go a long way to completing the items you WANT that we have had a hard time adding to v2, because of the overall layout and early design choices we made.

>>> I understand the OS change. Now, look at it from my POV.

V2 to V3 - requires physical access to the board to upgrade - and a complete re-program while doing it. At least a good half-hour. Why? "we don't want to invest the time to make it a seamless upgrade". Why's that? It is EOL, obviously. And I think I'm under-reacting.

StarVx gets it features from a COMPLETELY different operating system. It is a Real Time OS called VxWorks. It is fine tuned to the hardware and allows us complete control of every piece of code in the system. Sorry, but we can make the code do things in VxWorks that is near to impossible or actually is impossible in Linux. They are each different designed for different tasks in mind.

I will ask -- how is providing the CPE version, the Router version, the Server version, and the soon to be released v3 version obsoleting everything you have built?

<<<< When I'm told that we now have powerful new tools at our disposal, such as the "11 sector tower", except to find out that all my CPE are no-go, and there is no compatible CPE, and there will only be one if you get around to it, and if WE pay for a rather expensive version to begin with until you can get the price down with volume. Oh, wait, I've heard that argument before... It was why I was buying WRAP based CPE... Now there's no WRAP boards at all coming from you, and all that investment in hopes of the "big deal" is nada due to a platform switch. Not only are you not searching for a cheap X86 CPE board, you've just locked any else's research out as well.

Do you REALLY think people are going to do that twice??? Come on...

tony

08-24-2005, 11:48 AM

As one of the lead programmers, I would like to clarify a few things.

We have not abandoned StarOS, or x86 in any shape or fashion. StarOS v2 is still available, and is still being worked on. It has grown to the point where we would like to include some very exiting capabilities, however early design decisions have forced us to rewrite much of the code base. The result of this effort will be v3 (followed by v4, v5, etc..)

The WAR platform is simply another product we offer, and is not directly related to StarOS, or it's ongoing development. Much of the functionality we are adding to the WAR platform will be available in StarOS v3 (such as WDS, DFS, WPA2, Virtual SSIDs, etc.). As for the channel bandwidth reducing option, this was requested, and funded by a 3rd party for inclusion in the WAR platform. Unfortunately, agreements restrict us in releasing it outside of the WAR platform at this time.

I will now return to the shadows, and continue on with my other duties.

lonnie

08-24-2005, 02:04 PM

Mark, when we switch from Linux to FreeBSD it cannot be seamless. Part of the previous design problem that limited expansion was the config storage. That is being redone. Thus the "break the old" way of upgrading.

Believe me, I searched long and hard for a better x86 platform. They exist but at a huge cost and then they also do not have the low power. Our new platform has much higher performance and it draws less power.

As for everyone's x86 research you still have not accepted that the x86 is alive and well. QUESTION: If it was EOL why on earth would we be taking the time to get v3 out to you guys? I mean, seriously, if we were going to EOL StarOS why would we put time into v3? I realize you are upset but use logic to see what we are doing, rather than trying to second guess what we are not doing.

We are not hurting anybody by not supplying WRAP boards. My decision to stop stocking them does not spell the end of WRAP board production. When we first began with WRAP boards I had 65% of their total production but now I am likely under 5%. That translates to a lot of places to buy WRAP boards.

I am not sure what you mean by the "big deal". The WRAP has dropped as far as it can due to the cost of the CPU and RAM but it does not have the performance we need. What would you do? Well, I found another platform that has the performance and now we are working to get the price down. I had to start somewhere. The WRAP has adequate performance for most ISP users and it is quite affordable. I can understand if you do not wish to go down the new path since the old path serves you quite well. I will add again, that v3 will server even better.

sten

08-25-2005, 06:28 AM

Mark, when we switch from Linux to FreeBSD it cannot be seamless.

And i read on another list (wisp@tog.net) you used to run FreeBSD/PicoBSD....
I am a die hard BSD man myself. Why the choice to suddenly base StarOS on Linux? You dont have to go into why you would want to go back to BSD - That is far too obvious to waste time discussing. IMHO ;-)

tony

08-25-2005, 08:00 AM

StarOS uses linux mainly because FreeBSD had very poor pcmcia (and no cardbus) support back in the day. We have been thinking about moving back to BSD for quite some time now, and seeing that we are faced with a StarOS rewrite / redesign anyhow, we decided to make the move at long last.

sten

08-25-2005, 10:53 AM

Wow, you answered even when i tried to be as arrogant as possible. Cool! Oh yeah, i understand the PCMCIA/Cardbus issue. :D If we were to produce software for our hardware products today it would be NetBSD but not for any other reason than it runs on a toaster (http://www.embeddedarm.com/news/netbsd_toaster.htm). Otherwise it would be definitely be FreeBSD.

lonnie

08-25-2005, 02:58 PM

We have played with NetBSD and it is a good choice as well, and we might even use it. The choice between FreeBSD or NetBSD can be made late in the cycle since it is not too difficult to port between the two. Or we could do both. Guys want choice and that would give it.

I did not really catch your arrogant tone. I guess you have to be Linux weanie to come off as arrogant. BSD folks are just normal folk and it shows no matter how hard they try to be arrogant.

palmczak

08-26-2005, 10:30 AM

I think we all hear you, when you say that V2 is not going away. And I am excited about V3.

But, I know from experience that it is impossible to serve two masters as well as you serve 1.

sten

08-26-2005, 01:22 PM

...Unless you have a mad good whip. I'm not familiar with StarVX' build environment but i'm willing to speculte out of boredom. Much of the code can be reused unaltered between the two build environments. :D

lonnie

08-26-2005, 02:02 PM

You re correct, much of it usable between VxWorks and FreeBSD. There is effort to port, but we know what we are doing and the basics work the same. Just the very low level code needs porting.

titan_wireless

09-18-2005, 08:25 PM

To do this I simply have to find a <$300 CDN Star-os CPE solution. The lowest I think I can get it down to right now is about $380 and that will limit the number of installs I can do quite a bit.

Skaught,

Have you found a StarOS CPE for under $300 yet?

Regards,

Skaught

01-22-2006, 03:58 PM

Nope, I am going to go with the solution you are developing for me in conbination with a free Linux OS on the CPE. The $40 for star-os kills the budget on CPEs and I only need 5% of the features STAR-os has and a few it does not.

ninedd

01-22-2006, 07:46 PM

neil

01-23-2006, 07:12 AM

Yes with 5Mhz and 10Mhz Channel capabilities, I don't think this can be beaten for price/performance

ninedd

01-23-2006, 10:08 AM

Well, I'm not evangelizing any particular CPE, just that the price barrier seems to be an ever-moving target. A couple month's ago, sub-$300 was the requirement, and now that the price is there, I'm sure there will be those that say it needs to be $230 or there is simply no business model of any way, shape or form that $300 can possibly work for. :)

Don't get me wrong, I'm a HUGE supporter of lower cost CPE's and I do certainly understand the importance of getting it cheaper and connecting more customers as a result. If brand S's CPE is $100 more than brand T's CPE, and you put 50 customers on an AP, then that's $5000 extra for that AP by the time you're done. There's lots of ways to do that math and lots of business models to consider.

I was just commenting on the Sept conversation about there not being a <$300 Staros CPE. That seems to be solving itself, but wondered what Skaught was trying for a solution and at what price and with what results.

Skaught

01-23-2006, 07:05 PM

I have Tasos over at titan helping me out with a 5ghz CPE with the following specs.

+26dbi antenna
CM9
WRAP
FreeBSD/Madwifi/some of our own magic (auto backoff etc)
An enclosure good to -40C and made of metal (allthough I have no chamber to properly test this but our insulated WRAP backhauls seem happy year round)
POE with surge supressor (without it we would suffer due to hundreds of lightning storms we get)

And so far I have it below my $300CDN price point but I have yet to finish building one to test it but each compent on it's own is tested. (I am using staros for my testing but my production unit will not let star-os fit in at $40 USD) For say $5-10 I might consider a star-os CPE based solution but I am not sure as it still missing a few features I want.

I will run Star-os on the AP and on commercial clients. I will initially be using WAR baords for the AP and commercial clients but I may move to nano ITX as it will work with anything.

luka

01-24-2006, 05:31 AM

Nano-itx is a quite expensive and will stay such some time. Also, maybe it is better option to find some flat panel 23 dbi antennas? They have wider frequency range, are strong enough, not big and looks better than grids.

tog

01-24-2006, 09:13 AM

I have tested FreeBSD 6 and -CURRENT (FreeBSD 7) on WRAPs and I have had some issues, BUT I did get to test out WMM/WME capability and that worked nicely!

When I use if_bridge under FreeBSD with the ath driver, I seem to get mbuf corruption, it's very nasty and crashes multiple times a day.

The one I have up that does no bridging and is only routing is solid 100% no crashing.

In addition to the problem with if_bridge, the ath driver in 6 and 7 has some new bugs that 5.x's ath driver did not, i.e. the ath interface timing out and needing to be ifconfig down/up'ed.

All in all, it was promising but it had too many problems to actually use.

Also keep in mind if you want to use it as CPE, there is no solution for bridging it that I know of. StarOS has a nice solution currently in v2 with the proxy arp translation stuff and an even nicer solution coming up in v3. We have no similar solutions at all available for FreeBSD that I know of, but Sam is working on adding WDS support and he told me the number of WDS associations will be unlimited, not limited to 6 or anything like some commercial access points I've used.

StarOS v3 is supposed to have WDS or something WDS-like for use on the CPE side so we will no longer have the proxy arp translation thing. I'm not sure if it v3's WDS will be standard/compatible WDS or if it will be proprietary.

lsf

01-26-2006, 09:38 AM

Since you are looking into freebsd as a CPE OS, you might want to take a look at www.pfsense.org (http://www.pfsense.org) as a CPE it will do a lot of nice things, as a AP it will most likely not.

phendry

01-26-2006, 12:06 PM

Have you tested the wireless performance of pfSense on a WRAP board? Would be interesting to see what sort of speeds you can get.

Skaught

01-26-2006, 07:46 PM

Specs seem to thinnk it would need much more ram than a wrap can offer.

tog

01-26-2006, 08:48 PM

Yes I messed with it for a little bit, it works fine on a WRAP with 64M of RAM.

I ended up rolling my own plain vanilla compact flash distribution of FreeBSD to mess with, though.

David L. Vrablic

01-26-2006, 09:14 PM

I can understand if I set this super sector up in a remote area where there is no one else. But how does this work with hundreds of Road Runner Linksys AP's all playing at the same time in a given area. It can't free up the available spectrum used by others.
I understand that the cloaked channel can not be seen but I just don't see how that interreleates to existing non cloaked gear.
Someone shine a light in my direction please.!

Skaught

01-26-2006, 11:04 PM

I have never really had issues with home APs. Even other 802.11 ISP APs are usually not too bad if they have few clients. Now non 802.11 is another story.

David L. Vrablic

01-27-2006, 08:06 AM

Got it, Just because I see their beacons they most likely arn't making enough noise to overcome my SOM?

Legionx

02-06-2006, 01:10 AM

If there is a single MiniPCI PCBA with IXP422 266mhz cpu, 32mb ram and 8mb flash with a price below USD50, will this be interesting enough? Can't seem to find this on the market...

lonnie

02-06-2006, 08:49 AM

Skaught

02-06-2006, 06:21 PM

Well, if you're buying from Valemount - the StarOS is already installed and built into the price. Wartenna CPE's are $250 - $260 with a Power Supply, and I think they were $230 on the YSYL sale.

http://www.staros.com/prices.php

That looks to be sub-$300 doesn't it?

But that is at 2.4, I need 5.x at 24/26dbi gain. The WARtenna also lacks the EMP protection I need. And I am talking in CDN.

The solution that the guys at Titan are putting together will do all that and for less.

budi

02-08-2006, 10:33 AM

Dear Skaught,

What's type of surge protector (ethernet) do you plan
to use in your CPE ?

We're lacking good & comparatively cheap ethernet surge
protector for our CPE.

Thank's
//Budi

Legionx

02-09-2006, 02:39 AM

That would be a very hard price to meet even with an order for 100,000 boards. I am willing to look at anything that comes close to that price though, so if you know of something, please forward me contact details.

i'm working on the BOM now, trying to design a reliable board. it should not exceed USD50 if i do everything carefully. IXP422's price very high. if i succeed i'll definitely post updates

tog

02-09-2006, 02:57 AM

Legionx

02-12-2006, 08:12 PM

Keep in mind Gateworks did a really nice job with the whole design. It's a very high-quality SBC and when I have good software for it I will be really happy to have them as my access points.

If you can put something more barebones together for CPE purposes, we're all thrilled at the idea, but I can't imagine a $50 board being as nice :)

WRAP boards are more than enough horsepower for CPE and I've seen them as cheap as $85 so that's probably a good baseline comparison.

$85 you might be referring to PC Engines? $50 is a challenge, but that will work out somehow with very careful design for single MiniPCI.

My current plan is IXP422 266mhz 32mb ram 8mb flash. cpu's more expensive compared to IXP420 266mhz (doesn't have AES/DES/3DES). if the price doesn't meet then i'll go down to IXP420 266mhz. any suggestions on specs you want to see?

luka

02-13-2006, 03:53 AM

My current plan is IXP422 266mhz 32mb ram 8mb flash. cpu's more expensive compared to IXP420 266mhz (doesn't have AES/DES/3DES). if the price doesn't meet then i'll go down to IXP420 266mhz. any suggestions on specs you want to see?
Since atheros has crypto hardware support, you can stay with 420. More ram would be fine, ie 16/64.

tog

02-13-2006, 11:35 AM

Maybe I'm wrong, but I kind of gathered that his idea was to provide you with a similar/compatible hardware platform to the WAR board that was cheaper. I think this conversation is about something that's for Valemount's benefit unless "Legionx" is intending to sell these SBCs himself directly.

That having been said; Legionx, you might want to just discuss your progress privately via email with Lonnie, even if your intent is 100% for Valemount's benefit just to show Lonnie a cheap hardware platform he can sell.

lonnie

02-13-2006, 01:06 PM

I guess I did say I would be interested in a $50 board, so carry on, but be advised we are not sitting still either and I have several other hardware platforms under evaluation.

We had to start some place and the IXP filled a need for more power than WRAP but just as nice to use for remote and tower installs. I don't think they are appropriate for the final Client unit, even though they are awesome for that. We have been using them for all of our own customers in the last 3 months. A dual WAR in a Rootenna gives about all you could ever want, except of course price, and that is being handled with another unit.

Bossman

02-14-2006, 02:04 AM

lonnie

02-14-2006, 09:20 AM

The proper operative, IS, should be used.

Bossman

02-14-2006, 04:59 PM

The proper operative, IS, should be used.

OK, so when are we going to be allowed to see what IS being used? Again either I'm missing something or it's still not public.

I believe that VNC is well aware that a number of us are working on our own hardware trying to make a buck and/or do things our own way. It is evident by this thread that many are looking elsewhere and expressing their frustration. Every time you guys make a sharp left turn we lose ground and money on what we've been working on.

We've dealt with VNC for years and have been happy. I personally don't think going to a cheap CPE will slove all of our problems as an ISP. We've seen the service calls get to a point where it cost us more than we ever could have tried to save. For the stability and WRAP related issues we have no concern with moving to the VX platform.

At the same time many of us get the feeling that you want to have your cake and eat it too. You tell us that if we want a cheaper CPE to just give you money and have faith. That's hard for us to do.

lonnie

02-15-2006, 12:24 AM

David L. Vrablic

02-15-2006, 07:50 AM

I am sure we are not alone but I wanted to add my plight to your thoughts for the future.

We have taken your advise at STAR developed.
The first CPE's were the Teletronic CPE with Lucent cards.
Then a mix of Prism based as you recommended the change.

We went flat out and built many $300 ASUS 680E mini ITX units with PCMCIA adapters and loaded them up with the HP235 Prism 2.5 cards.

Now all new CPE's are WRAPS with CM-9s .
So comes the rub. To make them work with all the improved features I had to change the AP's to Atheros.

Adding to the "lack of options" at this time, support for Prism has come to a halt......
That means to me "Get rid of it now" before I really have no place to go.

This week I had to make two emergency swapout units to get customers back on the air because of the Atheros AP radio changes.
I am sure there are more and they are scattered 360 from a sectored site.
Signals are teriffic, associations are solid. traffic slows down or is intermittiant with stops and starts.
One sector would not even pass traffic with a Realtec Aeronaut 250.

Granted I am being forced into doing something I need to do but it hurts with time and expense at a very bad time.

I think what would make things easier for me in the interium untill a lower cost CPE with Atheros compatability comes along is..
A Atheros PCMCIA card that I could just plug into the adapters of the existing units that have been working all along.
I understand that isn't going to happen in V2 at all, so thats out.

I just got caught with the compatibility issue of Prism AP's with Atheros CPE's and mixing new equipment with the older deployments.

No one has answered my question yet about applying the new preamble settings to the problem. Or what the interaction might be.

I don't need any long distance, fat pipe backhaul gear as all links are in place and providing 24 to 54 meg.
I am not going to even mess with VX to make my matters worse,
So that leaves me waiting for the V3 "WAR" client or whatever and I know it is at the bottom of the list.

This problem cast some huge doubts on HotSpot operation also as the Atheros seems to have issues with other vendors radios...

This is an issue as we are in the process of converting all our corp sectors into combination fixed and hotspot service.
With Prism not supported and EOL for Star I can't use them.

That is the position I find myself in at this moment.
Now excuse me as I go beat my balding head against something hard!

lonnie

02-15-2006, 09:06 AM

Prism is EOL. There has been no development in the firmware ever since Intersil got out. Same with Orinoco. They are good cards but the software is dated and we are seeing slight variations in prism2.5 that do not like the old firmware we load for AP mode.

The thing is, I am not responsible for Aeronaut Realtek Client radios. I told the guys making them not to do it since it would be a huge support problem down the road. I guess they're now at that point. Put some pressure on the maker of that unit to bring it up to current standards. The long/short preamble should fix it. We released a beta with it. v3 will have it.

PCMCIA is a nogo. I will NOT return to the days of getting blasted becasue some cheapskate got a bunch of PCMCIA adapters on eBay and wah, wah, wah they don't work and it is all our fault. You saw all the flak I used to get. We were expected to commit our time so some guy could brainlessly purchase "flushed out" gear.

We go through a LOT of time and money to evaluate products. That is why the things we have sold have withstood the test of time. When I sold a good PCMCIA adapter it was my expectation people would support that effort, but the eBay factor came about. Also, other resellers bought and sold based on the $$ and did not care about quality, reliability or compatibility. Then everybody cried that I should support their efforts. My response is "Support your own efforts. I support mine".

When you buy products from another company, DO NOT expect me to go out of my way to support them. We have our own products we are working on and any time we spend supporting someone else's crap is just a huge waste to me. If they can't fix their own gear that should tell you something. Why do we get flak because of some other equipment's shortcomings? It is not fair, it is not right, and it has to stop. The guy who makes the unit is the guy to fix it. The standard for radios has become short preamble. The only cards I have seen, that we have supported, that require long preamble are the ancient TT 2 mbps cards. So, whoever is using long preamble on modern cards is pretty dense. Enough said.

Why do you assume anything is at the bottom of our list? I certainly did not say that.

David L. Vrablic

02-15-2006, 10:45 AM

Oh please don't read anything negative into what I said.
I made my choices based on almost no information.
I will have to live with those decisions.
I was just telling you that I for one am between a "Crack and a hardplace" for the moment.

I agree and am well aware of what you are going through.
Remember this is Dave V. I have been with you since the TT CPE daze ;-)

I only have about 6 of the Realtec units and I know they will not work in my system. I expected to change those out right from the beginning. They are all in the cabinet now. They will not go back into the system.

I am pleased that the EOL PCI / PCMCIA adapter kludge is behind us.

Also Pardon my assumption but I figured if you were working full speed ahead on the WAR / V3 project - WRAP V3 Upgrade I kinda thought a CPE would be down the list a bit.

I have an order for 20 CPE units waiting on my desk at this very moment but I am not going to order anything more that isn't Atheros / Star compatible that will do client mode.
I may have to get more Wraps with CM9's for the time being but I keep watching the Star hardware page.

I just have to have client gear ASAP.

Bossman

02-15-2006, 01:55 PM

I'm not sure what your point is Arthur. Of course I am aware that lots of guys are working on their own hardware and trying to do things their way. It should come as no surprise that we are as well.

I'm not trying to give you a hard time either. Similar to Dave, we have been with you since before the TT CPE days. Loyal is really an understatment with many of us. We also tried to do a good job on the reseller front for you but the discounters selling the cheap gear seemed to get most of the business.

I agree that the crappy hardware didn't do anyone any good. We for one only sold quality product and we've stood behind what we sold and never sent any customers in your direction for support that was hardware related. What most people don't know is that we were involved in getting the WRAP started and like you got cut out of the loop.

We are doing what we can to support VNC and not put our own business at risk. I think the VX route with specific hardware is one of the best things to happen. I was tired (not nearly as tired as you) of people going out for the cheapest product too and then complaining about problems.

We're just finding it hard to guage where we stand with VNC. We've expressed our interest to you but without tons of cash to send your way we can only do so much. We get the choice of being a customer only or being a customer and competing with you for sales. You said yourself that even if we bought in volume that it would be hard to make any significant headway while competing with you.

I guess I have no point. It may be best if I just keep my opinions to myself from now on. I don't want any hard feelings and this thread has enough tension.

lonnie

02-15-2006, 03:40 PM

Dave, you got me out a real pickle with TT in those unforgettable days, and I truly appreciated it then and still do now.

All I am getting at is that people cannot expect us to support their choices. Heck we threw away about 20 DWL900AP+ units because they were crap. We tried them, and they seemed to work, but the reliability was not even close to being there. We build a pretty decent AP unit and if some cheap POC does not work with it, too bad. Move on. We have units that do work.

Now, for the v3 development being top priority, and the unit I am not talking about, well v3 is the ONLY priority at the moment. Just take the logic one more step and realize that v3 is crucial to ALL future products. Once v3 is complete, we not only have support for IXP-42x but we have x86, PPC and MIPS. Each one will require a month or so of final porting and testing, but essentially we arrive at a single code base for almost any processor we choose to support.

If you accept that bit of logic, then you know that all we need is the right hardware and we are right at the point that I don't want to talk about any further.

Things are really close but still not at the point of taking orders. As I have said before, we are really starting to have fun now.

Oh please don't read anything negative into what I said.
I made my choices based on almost no information.
I will have to live with those decisions.
I was just telling you that I for one am between a "Crack and a hardplace" for the moment.

I agree and am well aware of what you are going through.
Remember this is Dave V. I have been with you since the TT CPE daze ;-)

I only have about 6 of the Realtec units and I know they will not work in my system. I expected to change those out right from the beginning. They are all in the cabinet now. They will not go back into the system.

I am pleased that the EOL PCI / PCMCIA adapter kludge is behind us.

Also Pardon my assumption but I figured if you were working full speed ahead on the WAR / V3 project - WRAP V3 Upgrade I kinda thought a CPE would be down the list a bit.

I have an order for 20 CPE units waiting on my desk at this very moment but I am not going to order anything more that isn't Atheros / Star compatible that will do client mode.
I may have to get more Wraps with CM9's for the time being but I keep watching the Star hardware page.

I just have to have client gear ASAP.

lonnie

02-15-2006, 03:58 PM

Arthur, I am not at all upset with you. We have known each other for a long time, and some day we have to actually meet.

I am at the point in my life when I have to stop providing my ideas and support so that other people can take my concept and build a big business from it. It is really impressive the number of dimwits I have helped over the years. All of them had more funding so of course they could always buy more and get a better price. When price rules I always lost out and just got to watch them eating my lunch.

Not any more.

I do value loyalty and I did (and still do) offer you a shot at doing something with us. The dimwits are on a blacklist and will not get access to me or my products as we move forward. They have no loyalty and I want to eat their lunch for a change.

A really great guy who worked with my Father used to say "Justice rides a slow mule". I have seen how things usually get made right, but it takes time and you have to keep going forward. As many people know I simply will not quit. Some people call it stubborn. I call it doing the only thing I know. It is fun and we make a good living. What more can you ask for?

I'm not trying to give you a hard time either. Similar to Dave, we have been with you since before the TT CPE days. Loyal is really an understatment with many of us. We also tried to do a good job on the reseller front for you but the discounters selling the cheap gear seemed to get most of the business.

I agree that the crappy hardware didn't do anyone any good. We for one only sold quality product and we've stood behind what we sold and never sent any customers in your direction for support that was hardware related. What most people don't know is that we were involved in getting the WRAP started and like you got cut out of the loop.

We are doing what we can to support VNC and not put our own business at risk. I think the VX route with specific hardware is one of the best things to happen. I was tired (not nearly as tired as you) of people going out for the cheapest product too and then complaining about problems.

We're just finding it hard to guage where we stand with VNC. We've expressed our interest to you but without tons of cash to send your way we can only do so much. We get the choice of being a customer only or being a customer and competing with you for sales. You said yourself that even if we bought in volume that it would be hard to make any significant headway while competing with you.

I guess I have no point. It may be best if I just keep my opinions to myself from now on. I don't want any hard feelings and this thread has enough tension.

Mark

02-15-2006, 05:23 PM

I am at the point in my life when I have to stop providing my ideas and support so that other people can take my concept and build a big business from it. It is really impressive the number of dimwits I have helped over the years. All of them had more funding so of course they could always buy more and get a better price. When price rules I always lost out and just got to watch them eating my lunch.

Not any more.

I do value loyalty and I did (and still do) offer you a shot at doing something with us. The dimwits are on a blacklist and will not get access to me or my products as we move forward. They have no loyalty and I want to eat their lunch for a change.

A really great guy who worked with my Father used to say "Justice rides a slow mule". I have seen how things usually get made right, but it takes time and you have to keep going forward. As many people know I simply will not quit. Some people call it stubborn. I call it doing the only thing I know. It is fun and we make a good living. What more can you ask for?

Lonnie, this looks a lot like "we got beat selling hardware before, so now we're going to run everyone up a blind alley and screw them on hardware this time"... I hope that's not what you mean.

You guys had the world's most awesome concept... Provide a software platform that allowed a universal interface and set of features (well, there was SOME hardware specific features) for a wide array of hardware, so that people could put solutions together from a multiplicyt of vendors and products.

Now, it seems you're trying to be the next Canopy or Alvarion, where you have captive hardware/software products. I don't really understand this switch. There are a LOT of companies out there making "solutions", which are just plug in and use. They have a lot of square pegs to pound into our round holes. If you follow this route, you will become a "we have one too" company, one we'll remember with a certain sentimental fondness.

You gave us a shapeable peg to put into almost any hole.

But it seems you're not content to stay in that role, you want to control it all, and make everything you sell worthless without you, and what you write worthless without what you sell.

How many people got burned on Karlnet?

The best thing about star-os was that it added value to many other products. Now, many of us are facing change, and wondering just how "open" our future is with you, and if you're trying to blind alley us into something we'll regret... Like the karlnet stuff.

yes, I know, you're asking us to be patient. It really hasn't anything to do with patience, in some ways. For many of your customers, decisionmaking time is determined by markets and competing products. If you can't provide us a roadmap to a "safe" future (staros always had a "safe" future, becuase it could be used even if you went away, and the hardware could be used even if it didn't have star-os on it), you're going to hamstring yourself a bit.

Many of us are in the valley of decision. What we put up now has a definite life span. It has to get paid for. It needs to have a lifespan. It has to stick around for a while.

And when you're planning on the future, and a 400 or 500 client deployment, things like "$20 upgrades" for specific functionality are no laughing matter. Or $40 or even $60 price changes. I read a lot of "so what's $20?". In the long run that's $10,000 to me, or could be. That's not a decision ANY of us should make lightly. Some of us must make such decisions NOW, not later.

Give us a commitment to staying that "safe" value adding company and product you were and we'll stay with ya.

So you talk about loyalty... It's a two-way street. It's not all about "trust us, and then we just do what we feel like, even if it screws you over". We need you to be responsive to us as well.

Heck, you've never seen a demand from me for a deadline. Nor a complaint because you missed one. My complaints have always been when you laid out a future path and then swerved from it. Like, promises of QOS and source routing. Or remaining that amazingly flexible platform for diverse network needs, or trying to become the "set solutions" company.

lonnie

02-15-2006, 07:20 PM

Mark, are you saying that StarOS v2 suddenly is no good? I reject that assertion entirely. You are free to choose your own equipment, just don't expect me to take responsibility for your choices. We are investing a lot of time and money into the WAR platform and I feel it is worth it. Coincidentally so do a LOT of other people.

We have awesome drivers that are reliable and high performance. That is what we are all about. I must say I am astounded over the ruckus about the IXP board and whatever else we are working on. It is so sweet to have a system doing VDS and routing that can pull 30 mbps and still not hit 100% CPU. It is nice to have a hardware watchdog, just in case something goes wrong. It is also nice to have only 8W of power draw for our Solar Powered sites.

Did you read about the site we just took down that was an OLD TT CPE with Orinoco card and amplifier? It got one connection of a 4 port WAR and signals (unamped) are higher and throughput is WAY better. The old CPE was running 1.06 and it was still doing a job that people appreciated. Sure it did not have Hotspot, p2p matching, Atheros, 5 & 10 MHz channels, and a host of other things, but you know what -- it was flawless and just ran and ran. I'd take that any old day over bells and whistles.

Lots of you guys do not have the basics of networking figured out (not you, because I know you understand it) but they know they need the latest whizz bang features and they need high power cards. I can run a very good Internet service without the new features and I do not need high power. As for the high power, it amazes me that guys will beat me up over the price of our hardware and software and then in the same breath tell me they are not wanting our CM9 so they can pay more for high power cards they do not need.

As for QOS and source routing, it is in v3. It works. Now all we have to do is get a user interface wrapped around it and we can deliver it. It sounds to me like you are complaining about a missed deadline, because we have those features, just a lot delayed. I also fail to see why we are still not amazingly flexible. Sure we have dropped some useless things, but we have better things to replace them with. They are outdated.

Anyway, I have no time for all of this. We are developing new hardware and software, like it or not. I grow weary of this whole thread and topic in general. I hate being attacked for some perceived future wrong. At least you could wait until you see the new stuff before attacking me over its failings.

Bossman

02-15-2006, 10:49 PM

Time for me to step over to the other side of the fence on this post.

Now, it seems you're trying to be the next Canopy or Alvarion, where you have captive hardware/software products. I don't really understand this switch. There are a LOT of companies out there making "solutions", which are just plug in and use. They have a lot of square pegs to pound into our round holes. If you follow this route, you will become a "we have one too" company, one we'll remember with a certain sentimental fondness.

yes, I know, you're asking us to be patient. It really hasn't anything to do with patience, in some ways. For many of your customers, decisionmaking time is determined by markets and competing products. If you can't provide us a roadmap to a "safe" future (staros always had a "safe" future, becuase it could be used even if you went away, and the hardware could be used even if it didn't have star-os on it), you're going to hamstring yourself a bit.

Many of us are in the valley of decision. What we put up now has a definite life span. It has to get paid for. It needs to have a lifespan. It has to stick around for a while.
I've had similar thoughts about the Canopy/Alvarion comparison, BUT... the only time I've had a Star-OS upgrade problem was on a PC system... never a WRAP which before the IXP was the recomendation of VNC. You can't say the same for Canopy or Alvarion upgrades. Do the others work with wi-fi if we want... no. Do the others take all the crap in a forum like this and still deliver product? Hell No!

The argument that the hardware can't be used is a little weak. That could have been the arguement for the WRAP board before everyone and their dog started to write for it. There are a number of companies currently writing for the IXP platform, and there will be more to come I'm sure. Heck even big brand names use this platform because it does what it does so well.... which by the way is why it was chosen by VNC.

As for life span, I still have customers running an old 2 Mb TT somewhere and it's fine for what they need. Now if you consider the WRAP's I have running Star-OS, I can't remember where most of them are because I haven't seen them since they were installed. I have no doubt that they will run far into the future and perform quite well. If Lonnie says V3 will come to the WRAP I believe him.

We have awesome drivers that are reliable and high performance. That is what we are all about. I must say I am astounded over the ruckus about the IXP board and whatever else we are working on. It is so sweet to have a system doing VDS and routing that can pull 30 mbps and still not hit 100% CPU. It is nice to have a hardware watchdog, just in case something goes wrong. It is also nice to have only 8W of power draw for our Solar Powered sites.

Lots of you guys do not have the basics of networking figured out (not you, because I know you understand it) but they know they need the latest whizz bang features and they need high power cards. I can run a very good Internet service without the new features and I do not need high power. As for the high power, it amazes me that guys will beat me up over the price of our hardware and software and then in the same breath tell me they are not wanting our CM9 so they can pay more for high power cards they do not need.
Anyone that doubts the performance of a WAR can talk to me. In the past 8 hours I've run about 60 GB (yes, 60 Giga BYTES) through our upgraded 15 mile WAR link and the ping time during the sustained 15 Mbps +/- throughput are averaging 8 ms. NO amps, no magic, just a network built the way Lonnie has preached. The old link even with an amp and WRAPs is 10 times more solid than what we had before, but can't hold a candle to the WAR gear. I have guys amazed at what we are doing because they barely see that with big name gear that is WAY more expensive.

Yes, it's hard to wait. Yes it's hard to trust. Yes some people are making decisions. Lonnie is staking his reputation and his livelyhood on his decisions. I too am nervous and frustrated at times, but the WAR works... plain and simple. In a few weeks We'll have about 100 miles connected over multiple hops and I'll gladly share the results to any of those with doubts.

rafamous

05-03-2006, 07:16 PM

I'm pretty new around here but maybe that's why I can come in with this perspective.

I say keep kickin A$$ Lonnie.

You have to do what keeps you getting out of bed in the morning with a spring in your step. If rehashing the old stuff is yawning you then screw it. It's your business. It's your vision, your ideas.

That's what got me into the wireless. It's always something new. Trying to figure out these crazy puzzles to get customers super speeds with rock solid connections. Winning new customers over from the cable company, yeah baby! Meeting new people all the time. Braving tornado winds while climbing freakin insane 500' towers. Let the others watch TV. I gotta feel the sun on my face, the rain on my back, break new ground. There's nothin like it.

Steady and strong, there's new frontiers to be conquered!

Still not as cool as Skaught. Dang it.