The current release of BBHN to be found in the firmware directory is 1.1.2.

On the home page I read, "The bug in release BBHN 1.1.2 announced in August causes OLSR Secure to crash. It is more widespread than originally thought"

and,

"we have posted a BBHN experimental version 3.0.0 with the Secure module removed.

...

We encourage you to move to this interim release as we continue our troubleshooting."

Question - are there any released / stable versions of BBHN firmware?  As far as I can tell there's nothing but  "Not recommended for use" (1.0.0), "crashes" (1.1.2) and beta versions (3.0.0b02).

Mark K5LXP

Albuquerque, NM

Mark

You have a valid question.  I will tell you that in my opinion, you would be best served with the 3.0.0 beta.  Yes there are still some bugs to be be worked out, but all of the released versions to date have had some bugs. 

My understanding is that as soon as the Development team can get some more feedback to help verify that their are no "major" issues in 3.0.0 that they will release it.  The more people that are willing to use it now and help validate or find new problems, the better.

As has been explained on these forums many times, the protocol has changed from 1.1.2 to 3.0.0.  The numbering change is to better reflect the protocol changes.  3.x.x nodes will only connect with 3.x.x nodes.  This means that your entire mesh needs to be on the same protocol.  Some people will not want to upgrade. 

3.0.0 supports a lot of new features over 1.0.0.  This is especially true with the Ubiquiti devices.  You have M5 and M900 devices now supported.  I believe that the 900 Mhz devices are going to open a lot of opportunities with BBHN.  There is the ability to use different channel widths.  Device to Device linking is a big plus for major hubs with multiple nodes or nodes on different frequencies.  Being able to reset a password with a 5 second reset button press or restore to freshly flashed state with a 15 second reset button press has been a blessing to me as I have tried to get some tunnels working.  The Aimer function is a big help to get antennas pointed in the right direction and peak signal strengths.  There are a lot of really cool things in the 3.0.0 build.

Unfortunately,  while the Linksys nodes running 3.0.0 will connect to a UBNT node running 3.0.0, all of these new features are not available or, in the case of Device to Device linking, need to be manually configured.  There are multiple reasons why but the big one is not enough memory.  An other is OpenWRT which is the underlying software for BBHN has dropped WRT54G support from its latest versions and many of the new features use code from the newer versions of OpenWRT. A third reason is the actual hardware will not support the new features.

I believe the 3.0.0 beta version is the best choice out there at the moment.  I have been running it here in Waco since it came out.  I updated to 3.0.0b2 when it came out. While I have seen a couple of OLSR crashes, the watchdog has restarted the node and our network has stayed up and been very stable.  My home node has been up for over a week with no OLSR issues.  That will probably change now I posted this. 

Clint, AE5CA

Hi Clint;

Thanks for the detailed reply.  In light of the info above and what I'm reading on the forums, I think I'm going to stick with, and recommend to others, V1.0.0 for the Linksys routers.   Beta testing and bug chasing can be challenging and fun but not when doing mesh demonstrations, or deploying a network.

It seems the "exciting" developments are all on Ubiquiti but they're not quite ready for prime time.  Maybe at some point when it can be demonstrated the BBHN software and Ubiquiti hardware is stable, I would invest in some Ubiquiti products.  For now, a Linksys running 1.0.0 is pretty stable and useful enough with some simple updates for most purposes today.

Mark K5LXP

Albuquerque, NM

Mark

I can appreciate your hesitation to move one to the 3.0.0 firmware.  But I do have to take exception with a couple items you said.  

• The Ubiquiti hardware is the real deal.  It works and it works well!  I cannot recommend that anyone deploy a BBHN node in an outdoor location using Linksys.  
• I would say that most of the issues  we have seen in beta testing the newer firmware are problems with the Linksys side not the Ubiquiti side.  The Linksys firware is getting difficult to support.
• The OLSR issue is present in both the Linksys and the Ubiquiti firmware.  It is also present in all the earlier versions but did not manifest itself as much.  Why it has popped up now is the the question the development team is chasing.
• It is significantly easier and cheaper to deploy a node using a NanoStationM2 than is is with a Linksys WRT54G.  The performance is significantly better as well.

My recommendation to the person just starting out today is get a pair of NanoStationM2's, and run the 3.0.0b2 firmware.  Of course 3.0.0 should have a released version in the next few weeks.

Clint

Ultimately I believe version choice had to be decided in each regional area by it's members so I won't go into what you should or should not do in that regard.

I will take a moment to back up Clint's statement that the issues exist in the software, many of these lines have already been traced and confirmed to exist in 1.0.0.

Most of these reports are being found out on Ubiquiti first and foremost because the Ubiquiti beta team is the most active group I've seen for sending reports in (this is not to say that other groups are inactive just that most reports I am seeing are showing up from them). The beta team has hundreds of hours each this year alone testing. Bug wise I've seen mostly Bugs that hit both or are Linksys based (big issue in 1.1.x with the built in switch)

At the moment no real Linksys team exists, the Ubiquiti team has been picking up the ball and pulling the Linksys builds forward and fixing bugs that have existed for years once they are told about them so bugs do exist in 1.0.0 (see code repo)

The biggest item I can see is that networks have gotten bigger now with those running Ubiquti because they can do things now they couldn't before. Several issues were tracking to traffic levels related which means 1.0.0 can have the issue if your net grows.

Known bugs with routing exist in 1.0.0 So if you do stick with it DO NOT plug in the WAN port unless you are OK with everyone accessing the internet or network plugged into the WAN port (even with mesh GW disabled). This is just one of the many issues the Ubiquiti team has fixed since they started looking at the project.

Ultimately though Clint is right the better tested builds are when we put them in an early release phase the more stable the final build will be and we can avoid issues like the Linksys switch and NAT issues in 1.1.x

This is just my opinion, but frankly none of the BBHN code is production code or even release candidate level.  That doesn't mean it won't be someday, but it's not today.  Having a major change in the development team and consecutive releases of fairly buggy code does not instill user confidence in new releases.  You may say those days are behind and even if true, it's still not field ready today.

I'm not trying to be a snot, I'm just offering why I'm not anxious to jump on 3.x.x.  Mesh is not my passion, so I have no interest in beta testing or playing musical software versions every few weeks.  My perspective is specifically as an end user.  As such I won't sit on the forum and throw rocks either.  I'll just wait and see what comes along and when I see something that looks promising, I'll jump in.  For today, 1.0.0 works perfectly for what myself and others in my area are doing with it.  As far as advocacy, it's a tough sell to describe how a beta version should work someday in front of a group of prospective users.  You cannot show something they can't reproduce with success themselves. In terms of deployment, it's unrealistic to expect users to "upgrade" field units in established networks every few weeks, reconfigure, and rediscover what works and what doesn't.  Having disparate hardware platforms and incompatible software versions is only serving to delay reaching the critical mass of nodes required for mesh to see any useful level of adoption.  1.0.0 was polished enough to attract my attention but the follow through hasn't happened yet.  

I understand, and agree that Ubiquiti is a better radio and hardware platform.  But because the Ubiquiti code is not solid yet, I see little value to spend money, time and effort to replace a working solution with an unknown one.  BBHN is just a little too "green" for me to make a sizeable hardware investment in right now.  For all I know, the the current line of Ubiquiti models that run BBHN today may not be the same models or even the same brand BBHN supports a year from now.  

I see a day when the Linksys platform will not be developed further.  I lament losing the low cost and ease of deployment they offer.  Conversely, should the Ubiquiti (or another) platform succeed, then it's not much of a loss to repurpose the Linksys boxes with OpenWRT (or other) and move ahead.  But, as a user, not until I see BBHN become stable.  To me, that's a shift from revisions for just getting it to run to revisions that offer improvements and optimization.


Mark K5LXP
Albuquerque, NM

Again, won't tell you what to go with as the following statements are equally true for version 1.0.0 as they are for 3.0.0b02  so it it wouldn't matter what version you chose.

I can say as the guy doing most the coding a this time, the Linksys isn't the stable code base... its got issues in it, and unlike the Ubiquiti base there is zero chance for them to be fixed by BBHN in some cases (Flaws inside the onboard switch module for one, and flaws in the WIFI driver which is closed source for another as example)

As noted, the issues your seeing in 1.0.0-1.1.2 are around 60% Linksys only issues (never affected Ubiquiti) and 40% affect both equally (making it 80% Linksys unstable, 20% Ubiquiti)  -- Doesn't really go well for calling "Linksys" stable and the Ubiquit code "unstable" when you look at it from a known flaws standpoint.

Again, choose the version your area is most comfortable with. In the end its each area that has to deploy the devices so they need to be happy with what they are running  and I do not intend to force my view on that into whats a local decision.