Hey Guys due to a recent poll done on Audi about implementing a rule on CTWUG that all new Backbone links made on CTWUG be dual polarity as a minimum and having 80% of the users saying they agree I would like to hear everyone’s feedback around this.
I feel very strongly about this as CTWUG can’t keep living in the past with technology from the stone age. Yes I know its a cheaper means of making a link, but having grid links as a backbone link is also giving a cheaper experience to everyone else on the network.
I can almost guarantee that the few members who voted no, are either clients or nodes who still run grid links. Thus they don’t know what the difference is between the two technologies.
Making the transition to dual polarity equipment is also one step closer to having a network where we can game and download at the same time with the right QOS in place.
So what I am proposing is the following:
All new OSPF links should use as a minimum the wireless N standard and be Dual Polarity.
All current OSPF links should be given 1 to 2 years to upgrade to wireless N and Dual Polarity.
Grids may still be used on the wug to connect to single Pol sectors or to be used for PTP links not running OSPF.
This would vastly improve our whole network and create an experience that everyone will enjoy. It has become to easy for people to try and take the cheapest possible route to become a backbone node on the wug, and I feel these members are trying to join for the wrong reasons. I would rather have fewer backbone nodes that are way more efficient and faster than having a million backbone nodes which have dial up speed and the latency of trying to shout something to someone on the moon from earth. CTWUG is there to push the boundaries of unlicensed wireless equipment and see what is possible.
Maybe I’m just stupid, but I’m just giving my 2 cents.
Jypels, I agree with you completely. I think too many people are going for the the quantity over quality approach when it comes to managing links. I will much rather see a smaller community running proper links and backbones than having this massive community and all that we are getting is dial-up speeds.
I appreciate that not everybody have the resources needed to run dual-pol backbones (especially if you have 4 or 5 backbone links) but I do think we should start somewhere, and this is as good as any place.
I still think there is a place for cheaper links, but they should be based on properly managed sectors etc. To that end, I also think we should perhaps start a project whereby we donate the unused grids / kits to the areas that actually need it in order to grow the “client-sector” community a bit. I have recently seen some old / junk kit selling for ridiculous prices and I think that that is misinterpreting the spirit of CTWUG being a non-profit organization. So all the backbone links that is being upgraded should be handed over to a central committee (or something) that will be able to manage its redistribution in a sensible manner. I will be seriously pissed off if kit I donated to somebody ends up being sold on the forums etc.
Is there anyway to determine from the librenms setup from spin to see what links are still running single pol on ospf? Not to mention names but maybe quantity for the moment?
Agree…1-2 years upgrade period for current running OSPF links is over enough time to upgrade. But how will this deadline be met. say Jan 2018 all grid OSPF links CUT?
I think we will lose a couple of nodes, but i also have a feeling that these are the nodes who doesn’t donate on dday, or during the year, and those who feel nothing for the growth of the wug in fact I think they won’t even visit the forums and see this thread unless Audi is offline for a couple of weeks or something…
Jypels said “I can almost guarantee that the few members who voted no, are either clients or nodes who still run grid links.” Wrong. I voted no. I was thinking about other people and their links and what the wug is ment to be. I was one of the first to run dual polarity backbone links with my 40km link between Fish Hoek and Gordons bay. Since day one, my sectors have been ubiquiti dual polarity and client kit as well. All my links and in my area are dual polarity.
Jypels said “having grid links as a backbone link is also giving a cheaper experience to everyone else on the network”. A short range grid link can sometimes do more than a long range dual polarity. But besides that, Ctwug is supposed to be first and foremost a research and learning network. Download speed is definitely a secondary consideration.
I agree slow single polarity links which run ospf are a bad idea, but other methods such as OSPF costing can mitigate speed issues with these links and we should look at those methods before we go forcing people to spend money and throw away effort setting up kit, so others who have more money or have come more recently since dual polarity has taken off can have a better wug download experiance.
This is not a simple “more speed for everyone so backbones must go dual polarity” issue.
I guess I feel that the regardless of equipment used we need people on links to fix them when they have issues. This is a problem no matter what the equipment is. This is why I’m working on the monitoring side so we can get stuff fixed and keep it that way.
The other thing is that we need to be as inclusive as we can be. I’d rather not exclude people when we don’t need to.
I also think we need to be clear on which problem we are trying to solve with this proposal and make sure that it will solve that problem.
Yes, we do need to fix links. Gaming is bad with all the link issues. I don’t think duel pol backbones will make any difference … quality over quantity.
I know you voted no, actually was waiting for your post on the matter…
Saying you where the 1st to run dual polarity kit is like saying I was 1st at Checkers this morning, it’s irrelevant to the matter, it doesn’t make you or me better than the guy that came 2nd or 10000th. I applaud you for running dual polarity in your area.
You just contradicted yourself here with the following: " But besides that, Ctwug is supposed to be first and foremost a research and learning network." how can we be a research network if we are sticking to technology of the past?
Then you said (I chuckled here a bit) " a short grid link can sometimes be faster than a long dual pol link. That’s like saying a Toyota yaris can sometimes be faster than a Bugatti Veyron if the Veyron is in reverse. In anycase I will take you up on this challenge, here is a 40KM wireless N Dual Polarity Dish link from Bellville to Gordonsbaai, it sync’s at 180mbps/180mbps with 99%/99% ccq and can do 130mbps of realtime traffic with an 1 to 2ms ping…
Please show me your short single pol grid link with real throughput?
To make this really fair you should look at short links vs short links ( Comparing apples to apples) but I don’t want to humiliate your statement to much
Changing OSPF costing on slower links should be done in any case… But imagine if you can do this plus having decent dual pol links all round. Seems like you like thinking inside of a box. We are not forcing anyone, if you feel you can’t save the R50pm for 2 years then you are welcome to not be on the wug. Anyone who thinks that you join the wug and Never have to buy any kit ever again got told a lie about what the wug is and how it works . We need to improve our infrastructure . We have much more active nodes than what we had 5 years ago.
If you really want to fight for the community then i’m afraid you are on the wrong side of the fence, as the whole reason for this thread is to improve the experience for all 1600 active members and not just do what the 20 members who voted no wants…
This should definetly be the longterm goal to aim for.
Agree at some we should start to move forward.
This is from ospf rules - We know that the TCP bandwidth test can max out the CPU, that is fine. All we want to see is that the combined speed is above 24mbps.
We should also look at maybe making the combined bandwidth test requirements 50mbs.
These rules should apply to all opsf routers carrying transit traffic.
Lets say we have a groove to groove link doing 65mbs traffic on a bandwidth test, will it be necessary to force the owner of this link to upgrade there link to MIMO? I think there will be some cases were upgrades will be not necessary.
I’m sure upgrading links to mimo can make a major difference, MIMO is so much easier to tune and to get good ccq with.
New mimo links will create new and shorter paths and upgrading old links that are are not performing as they should will in turn create better latency, from my point of view its a win win situation not just for me and you but for all of us as a community.
Yes to make that statement you need to run the equipment from and to the same locations. I can promise you in everycase the dual pol kit will win. Only time you might get better stability with single pol is going over a body of water like the ocean where radio waves can get reflected. But newer designs and advancements in radio technology has solved this issues and even grid links still struggle with water
At the end of the day the one BIG argument that stands for boh SP and DP though: FIX YOUR LINKS!
Once you can actually get people to take that interest, the battle is halfway won, be it to upgrade or not to upgrade. To many peeps a running link is good, even if the link does not conform to specs (ccq, signal, etc), and those are the guys to nudge awake in a hard way. Once that is achieved, the rest should follow & pick up speed as we go along…
on the matter…