Ping Helps Keep Mesh Alive

Ping command in most operating system has been a method to detect an IP address availability. In the last few years I have use ping command to help keep my MeshAP alive.

The problem arises from MeshAP lack of activity. The MeshAP are periodically used. When this happens I noticed that MeshAP sometimes looses connectivity between nodes; however, this is only true when the density of the mesh is weak. When MeshAP are linked together with "circle routes" or with "dog legs" they tend to have intermittent lost of connectivity. Generally speaking the observation is the nodes then do not contact WIANA and give an indication that they are not up. However they are functional.

To fix the problem I use ping and the gateway address to keep the MeshAP alive. When this happens 90% of the time with in 30 minutes the MeshAP is once again communicating with WIANA and everything is normal again.

For the other 8% of the time an "unblocknode" command is issued to closest live link. Radio interference tends sometimes block a node. Even if you try to ping the gateway you find that the packets are not getting through.

The remainder 2% of the time a reboot of the MeshAP may be in order.

Old Article Same Problem

Well about a year ago I started this blog and it has turned out to be very successful. As I reviewed my web statistics I noticed that the top three items on my web are Mesh Topology, Blogger, and Atom for the blogger. In fact since I let everyone know about my atom it is overwhelmingly became popular.

But I started to review many articles and asked people to update them. There have been many interested parties but nothing has become of it. I'm worried that information may become stale and may need to be removed; however, after reviewing many of the articles I noticed that they are still valid today as when I wrote them.

On of the articles that sometimes is popular is Attempting to connect to gateway. Now I know this wasn't the best article but it does explain some problems that are the result of poor implementation. As you know Wireless Mesh is fairly new and deployment has no basic methodology. One can argue that point to point deployments have decades of experiences and that each mesh node should be laid out that way.

My view was a little different based on other people writing. That Mesh nodes can be deployed any where and any how as long as the antennae have a clear line of site without Fresno interference. Well this is not the case.

I wondering if roof top deployments are the way to go or is there a better method. I explored bring the MeshAP to people level, such that 5 feet or 1.5m from the ground may be better for everyone.

But back to the statistics. I noticed that there has been a big drop in viewing all the articles I once written. I should actually be very glad that not only have 250,000 of you read my articles last year but people are still coming to my website.

I will keep writing as long as you keep reading. I'm experimenting using videos to help illustrate my point. It is not hard or expensive to produce but it takes a little more time. The problem will be is viewing them. Maybe I search for a video blog.

I going to ask for feed back. If you can let me know if the articles are good, bad, or simply ugly please let me know. feedback@moskaluk.com or post to the blogger.

Thanks

Mesh where no Links were before

Well it is cold enough and my radio signals are once again bouncing of the atmosphere connect the two furthest nodes together, see map. Is it coincidence that I just changed out one of the radio cards and started to get a new connection? The samething happened last year about this time. Mild weather in January follow by tremendous amount of radio skip. The question does occur why only two nodes? And is it truly radio skip or reflection from tall buildings in the area?

The new radio card sure boosted the power and the receiving such that a steady signal can be viewed. When I changed out the old 200mW radio card the signal fade to the point that it was not receiving or sending. Back in the Lab it is once again working properly.

If this was radio skip rather than reflection then all the nodes would start to try to mesh with each other. I believe this really is atmospheric combined with reflection. For example. I have two units that have a clear path between them but due to trees and small buildings they do not mesh. If it was atmosphere skip then these two nodes would start seeing each other. Currently they don't even detect the radio card signals (sigspy).

Connecting between two node that are reflecting off major structure is never good. Signals are not that consistent and tend to end up with an intermittent connection. Now saying this I do have numerous large building near me and some of my connections due reflect off these buildings. They are steady but the signal are never that strong.

Radio Card Life Span

I had a problem with one of my MeshAP's radio card. It pop out of it's PCI slot such that a PCMCIA adapter with a 200 mW radio card. I notice that the signal was starting to fade. After a months of loosing power the radio card lost many dbi's. Switching the radio cards I saw that the newer card was yet again stronger. As I brought it back to the lab and tested the radio card, it once again started to work.

Scratching my head, yet again I asked a few people about it. A number of other observers indicated that Radio cards loose power over time and then simple stop working. This concerns me as radio cards in my opinion should last a long time not a few years.

The new question is how many times will I have to change radio cards? What is the life cycle of these units? When should I expect failure to occur. Testing the wireless mesh can answer these questions.

Can the answer to these question be with a simple maintenance approach? I remember seeing a maintenance department change all the light bulbs in every floor of a building. So when I talk to the manager of the maintenance people he said that getting calls about burnt out lights was his number one problem and that two individuals full time job was to change light bulbs. I ask him how this was saving money? He said that in two weeks he changes all the light bulbs and frees up two people time for a year. Is this the answer to changing out radio cards?

Should we expect a life expectancy of 1 year and then change them out? I will review the economics but I think I will have to add this to the financial model.

Shake Rattle and Short Circuit

Today was a warm day for Toronto. Way above seasonal temperature I believe it was 11C Normally about -15C. Good day to go fix a MeshAP that I lost for the last 23 days. Fortunately the person is letting me use their chimney as a MeshAP tower. Well I got up on top of the roof and sure enough open the Nema enclosure to find that the radio card had dislodge itself from the of the PCI adapter on the motherboard. Normally I wire down and glue each component down but I wanted to see how long it would take to shake, rattle and roll. I can tell you it must have been a really strong wind storm because the CF-Flash card was also off. Wind vibration can really knock your units around even if they are bolted to a Chimney. I saw some sparks as I tried to boot the unit back up and sure enough it kept working.

You live and learn!

Street Lamps go Wireless

I just read an article that is very exciting. It has to do with consolidation of various types of technology. Consolidation of technology is relative known in the computer world, i.e. with networks, computers, TV, Radio, print are making up the fabric of the world wide web. However, I have notice other consolidation of technology that are some how becoming pratical. Who ever heard of a street lamp, WI-FI access point and solar panels? Well as per my blog I have been working on a solar panel MeshAP (WI-FI) connection. I saw the benefit of having wireless mesh having it's own power. But to share that power with a street light and use the pole as a transmission tower as well WI-FI connection is outstanding. It is like "killing two birds with one stone." But in this case it would be 3 birds.

I do however see problems with the technology. The solar panel would have to be very powerful not only to power a street lamp but also to run a access point. It can be done but we forget that some street poles also carry communication lines, electricity etc. I like the idea of solar panel street light but I think it may be too soon for such great ideas.