Community discussions

MUM Europe 2020
 
mickeymouse690
Member Candidate
Member Candidate
Topic Author
Posts: 105
Joined: Sun May 20, 2007 3:08 am

Not bad for 106 KM :)

Sat Nov 28, 2015 6:25 am

I must say im impressed. I was hoping to get around 30 Mbps. Was also not expecting this signal strength event though RadioMobile was spot on.! -58 both chains.!
notbad.png
You do not have the required permissions to view the files attached to this post.
 
mickeymouse690
Member Candidate
Member Candidate
Topic Author
Posts: 105
Joined: Sun May 20, 2007 3:08 am

Re: Not bad for 106 KM :)

Sat Nov 28, 2015 6:27 am

I have to confess that i get half these results with same gear in semi urban environment at a third of the distance.!
 
User avatar
chechito
Forum Guru
Forum Guru
Posts: 1747
Joined: Sun Aug 24, 2014 3:14 am
Location: Bogota Colombia
Contact:

Re: Not bad for 106 KM :)

Sat Nov 28, 2015 7:25 am

I have to confess that i get half these results with same gear in semi urban environment at a third of the distance.!

whats your equipment??
 
h3ll1
just joined
Posts: 15
Joined: Thu Mar 31, 2005 11:34 am

Re: Not bad for 106 KM :)

Sat Nov 28, 2015 1:07 pm

Test with TCP and post results
 
User avatar
macgaiver
Forum Guru
Forum Guru
Posts: 1723
Joined: Wed May 18, 2005 5:57 pm
Location: Sol III, Sol system, Sector 001, Alpha Quadrant

Re: Not bad for 106 KM :)

Sat Nov 28, 2015 1:28 pm

Test with TCP and post results
Why? what benefit will that give?
With great knowledge comes great responsibility, because of ability to recognize id... incompetent people much faster.
 
mickeymouse690
Member Candidate
Member Candidate
Topic Author
Posts: 105
Joined: Sun May 20, 2007 3:08 am

Re: Not bad for 106 KM :)

Sat Nov 28, 2015 9:22 pm

Link already in use with client so I can't easily get a computer behind both radio to test real tcp throughout. TCP tests with just the radios, I get around 180 Mbps.

Equipment
Netmetals
RocketDish 30 db

Client side is at a large tv and radio telecom site on top of a mountain so I didn't dare use a Mikrotik dish antenna with all the ice flying around off the towers in the winter, too damn thin the dish material. I can dent the dish with one finger.. The Rocket Dishs have already flown of buildings into brick walls with fair damage but not that bad considering what it just went through and have been put back in service. Sorry MT but make your Dishs thicker and I would be more than happy to use them :p. Love the Netmetals :)

Edit: Also using slant 45 on dish
 
h3ll1
just joined
Posts: 15
Joined: Thu Mar 31, 2005 11:34 am

Re: Not bad for 106 KM :)

Sun Nov 29, 2015 7:09 pm

Test with TCP and post results
Why? what benefit will that give?
TCP test is closer to real life data
 
User avatar
macgaiver
Forum Guru
Forum Guru
Posts: 1723
Joined: Wed May 18, 2005 5:57 pm
Location: Sol III, Sol system, Sector 001, Alpha Quadrant

Re: Not bad for 106 KM :)

Mon Nov 30, 2015 7:45 am

Test with TCP and post results
Why? what benefit will that give?
TCP test is closer to real life data
OMG not again!
Please, name me this "real life" traffic"???

1) speedtest?? - there is nothing "real life" here - it is sintetic single connection test,
2) Browsing?? - any modern browser using 8+ connections at the same, and connections are relatively short, so TCP bandwidth test will not represent anything
BTW not all browsing is not using TCP now - Google already moved to UDP - read this:
https://en.wikipedia.org/wiki/QUIC
So all your youtube videos actually use UDP now.
3) torrent download?? - it works with both protocols, but better with UDP.
4) gaming?? - uses both protocols, and in case of TCP - it is low speed connection so so TCP bandwidth test will not represent anything

bottom line:
TCP test is closer to real life data
it was true 10+ years ago, when i was in university Professors repeated this over and over again...
but that was 10+ years ago -in modern internet real traffic is "UDP".

while i was writing this post on my main gateway - 73,6% of client's traffic was UDP, 21.2% was TCP, rest was GRE, ICMP.
With great knowledge comes great responsibility, because of ability to recognize id... incompetent people much faster.
 
User avatar
normis
MikroTik Support
MikroTik Support
Posts: 24317
Joined: Fri May 28, 2004 11:04 am
Location: Riga, Latvia

Re: Not bad for 106 KM :)

Mon Nov 30, 2015 9:19 am

TCP test is closer to real life data
could you please give an example of "real life TCP" ?
No answer to your question? How to write posts
 
InoX
Forum Guru
Forum Guru
Posts: 1964
Joined: Tue Jan 09, 2007 6:44 pm

Re: Not bad for 106 KM :)

Mon Nov 30, 2015 10:42 pm

while i was writing this post on my main gateway - 73,6% of client's traffic was UDP, 21.2% was TCP, rest was GRE, ICMP.
How can you tell? :shock:
 
User avatar
macgaiver
Forum Guru
Forum Guru
Posts: 1723
Joined: Wed May 18, 2005 5:57 pm
Location: Sol III, Sol system, Sector 001, Alpha Quadrant

Re: Not bad for 106 KM :)

Tue Dec 01, 2015 7:52 am

How can you tell? :shock:
I created passthrough rules for TCP, UDP and ICMP and one rule for all traffic, that quickly found out that GRE was also present, it is easy to set up.
With great knowledge comes great responsibility, because of ability to recognize id... incompetent people much faster.
 
h3ll1
just joined
Posts: 15
Joined: Thu Mar 31, 2005 11:34 am

Re: Not bad for 106 KM :)

Tue Dec 01, 2015 1:44 pm

Synthetic test TCP on mikrotik gives more credible results for passing data over wireless links. This is my opinion based on my personal experience. I have seen speeds over 500Mbps when using UDP, but TCP is near 200Mbps - that is big difference. So i expect on this particular link TCP speed test should be around 110Mbps, which is actualy OK.
 
wispwest
Member
Member
Posts: 477
Joined: Tue May 19, 2009 3:48 am

Re: Not bad for 106 KM :)

Tue Dec 01, 2015 8:08 pm

Synthetic test TCP on mikrotik gives more credible results for passing data over wireless links. This is my opinion based on my personal experience. I have seen speeds over 500Mbps when using UDP, but TCP is near 200Mbps - that is big difference. So i expect on this particular link TCP speed test should be around 110Mbps, which is actualy OK.
Also check the CPU on the routerboard when running TCP test, it usually maxes out, so your TCP test is ONLY limited to what the CPU can handle. I've done TCP tests over PtP links on the radios themselves, and seen horrible results, but then test with RB1100AHx2 boards behind each end, and seen 300mb/s, whereas the ptp radios can only handle 80-120
 
Zorro
Long time Member
Long time Member
Posts: 676
Joined: Wed Apr 16, 2014 2:43 pm

Re: Not bad for 106 KM :)

Thu Dec 10, 2015 6:24 am

if you(or anyone else) obessed with "real usage/traffic"-alike stress/testing - then (commercial or free)software packages like NetChariot - had several relevant patterns for such testing.
or you can simply replay Exactly captured traffic - back again and again ;)

Who is online

Users browsing this forum: No registered users and 37 guests