Has anyone had a chance to do any performance testing on the RB333's?

I played around for a couple hours this afternoon but our warehouse isn't the best testbed..

test setup:
2 RB333's with ether1 bridged to nstream interface
4 R52's
4 dual band rubber duckie antenna's
desktop on one side, and laptop on other

nstream2 tests:
framer-policy=best fit
framer-limit=4000
Using 5ghz-turbo and 2ghz-turbo, I was able to get ~65 mbit tcp one way..
Using 5ghz and 2ghz-g, 42mbit tcp one way..

nstream tests:
polling on
framer policy = best fit
framer limit = 3200

5ghz-turbo = 37 megs tcp one direction(need to retest this one), 73.7 megs udp one direction..
5ghz = 34 megs tcp one direction, 42.6 megs udp one direction..

Whenever I would 5ghz with 5ghz or 2ghz with 2ghz I would get horrible speeds, but I'm pretty sure that was b/c I was testing in a metal warehouse with too large of antenna's.. Not to mention several wireless backhauls on a tower next to the building..

Sorry for the messy post.. it's late and the wife is bugging me to come to bed.. I'll try to test them some more tomorrow and post my results..

-Gerard
http://www.quicklinkwireless.com
Certified Mikrotik Trainer

Has anyone had a chance to do any performance testing on the RB333's?


Whenever I would 5ghz with 5ghz or 2ghz with 2ghz I would get horrible speeds, but I'm pretty sure that was b/c I was testing in a metal warehouse with too large of antenna's.. Not to mention several wireless backhauls on a tower next to the building..

Sorry for the messy post.. it's late and the wife is bugging me to come to bed.. I'll try to test them some more tomorrow and post my results..

-Gerard
http://www.quicklinkwireless.com
Certified Mikrotik Trainer

Hi

I am seeing the same stuff. Everything is horrible when using 2 wireless interfaces at the same time. Also only 2.4 b is working. G has serious issues. I have reached the same results as you regarding bandwith. I spend 6 hours yesterday trying to get 2 WDS bonded with EOIP but never got it working.. Can see traffic on the links but nothing gets past bridge so there's still a lot of issues with V3.. Would be nice to see if we could pull 100Mbit on 2 links ...

/Henrik

I think its the unshielded cards close to each other ( ie in the same board ) that does it. I suppose what you really need is this : Nstreme dual , involving 3 boards. One board/radio per antenna on the tower and one other board to combine the two.

I'm going to post this at the top level too, but I wanted to respond to the prior comment. I just moved two SR5's from a 532 to a 333 yesterday, hoping to pick up another slot on the same mountain. Same configuration all around. Signal levels on the clients we in the range of -64 to -81 across 25 clients of various types. After moving to the 333 signal levels fell to -75 to -88 on the 333 and from a low of -78 on the client side to -92. I put everything back to the 532 and all returned to what it was. I was worried about this in the card layout and this confirms it. Has anyone come up with a solution to this? Or is it a card per 333?