Anybody using Mikrotik full duplex (x4) in PTP links between 7 to 14Kms?
What throughput can be expected ?

no mikrotik products come in full duplex modes as of today.

you need 4 radios cards 2 on each side for full duplex to work.

no mikrotik products come in full duplex modes as of today.

you need 4 radios cards 2 on each side for full duplex to work.

Or use OSPF and 4 QRT's for example and base setup on http://wiki.mikrotik.com/wiki/Dual_Setup_with_OSPF.

thank you.

that looks interesting but has anyone deployed it in actual environment?

thank you.

that looks interesting but has anyone deployed it in actual environment?

I have a few sites doing that - for example we use an RB850GX2 at each end and 4 x QRT's.

Set up two unique /29 subnets for each wireless link and the radio's as simple bridge (bridge/station-bridge).

Apply the different costs to the RB850 interfaces facing the radios so traffic flows in a clockwise (or CCW) direction and you will see uplink on one segnment and downlink on the other.

Disable one link to prove recovery. This gives full duplex and fail-over should one link fail.

Another way (not personally tried) would be to use bonding and two layer two bridges. This is not strictly full duplex but gives good aggregated bandwidth and failover http://forum.mikrotik.com/viewtopic.php?t=70808

I had for 2 years a link with 4xSXT bonded. It worked very well.

90% of our towers uses this method.

Usually we use LigoWave because they are more reliable in high noise environments.

For the other links where there is no interference we use rb912uag-5hpnd and bridge them.

Gets 100Mbps + both ways.

If one link dies ospf takes care of the rest.

I would say it works so well that if you don't have network monitoring you wont notice when one link goes down

I had for 2 years a link with 4xSXT bonded. It worked very well.

What througput have you there?

Hi,

guys you think interface bonding is better than OSPF full duplex ?? any real experience ?

thanks

In my opinion: ospf is better.

90% of our towers uses this method.

Usually we use LigoWave because they are more reliable in high noise environments.

For the other links where there is no interference we use rb912uag-5hpnd and bridge them.

Gets 100Mbps + both ways.

If one link dies ospf takes care of the rest.

I would say it works so well that if you don't have network monitoring you wont notice when one link goes down

Can I ask the max distance between towers using this method?

I had for 2 years a link with 4xSXT bonded. It worked very well.

could you share your code used for bonding the 4 x SXT please? which device was used to bond them?

I've just found my first bonding result. Sadly the tutorial no longer exist. I've used 4xSXT and 2xRB750GL.
http://community.ubnt.com/t5/Business-T ... d-p/194143

Found it : http://wiki.specialistoff.net/index.php ... ti:Bonding

It was a better choice for me than full duplex.

I've just found my first bonding result. Sadly the tutorial no longer exist. I've used 4xSXT and 2xRB750GL.
http://community.ubnt.com/t5/Business-T ... d-p/194143

Found it : http://wiki.specialistoff.net/index.php ... ti:Bonding

It was a better choice for me than full duplex.

the last link is in Russian... can it be seen in English?

I've just found my first bonding result. Sadly the tutorial no longer exist. I've used 4xSXT and 2xRB750GL.
http://community.ubnt.com/t5/Business-T ... d-p/194143

Found it : http://wiki.specialistoff.net/index.php ... ti:Bonding

It was a better choice for me than full duplex.

the last link is in Russian... can it be seen in English?

Google Translate

90% of our towers uses this method.

Usually we use LigoWave because they are more reliable in high noise environments.

For the other links where there is no interference we use rb912uag-5hpnd and bridge them.

Gets 100Mbps + both ways.

If one link dies ospf takes care of the rest.

I would say it works so well that if you don't have network monitoring you wont notice when one link goes down

Can I ask the max distance between towers using this method?

Hi,

The longest link we have is about 150km.

Using RBNetmetal 5's and the biggest dishes ubnt makes.

Signal is -57/-55 getting about 40-60Mbps

The longest link we have is about 150km.
Using RBNetmetal 5's and the biggest dishes ubnt makes.
Signal is -57/-55 getting about 40-60Mbps

Respect!
Do you have any photos?

The longest link we have is about 150km.
Using RBNetmetal 5's and the biggest dishes ubnt makes.
Signal is -57/-55 getting about 40-60Mbps

Respect!
Do you have any photos?

Not of the physical equipment, but I can share point A and B

I can share a screenshots of the signal and distance

Not of the physical equipment, but I can share point A and B
I can share a screenshots of the signal and distance

It is something to be proud
Show us

Hopefully wierd stuff don't happen to the layout

Link 1 Signal/Uptime/Distance



Link 2 Signal/Uptime/Distance



Link1 Signal/CCQ



Link2 Signal/CCQ



Link1 tx



Link2 rx



Google earth distance and elevation profile

Hopefully wierd stuff don't happen to the layout

can i keep your photo around all my installations as an inspiration??

are you having a clear LOS for 150kms? because i thought the earth's curvature spoils RF signals when you work in the GHz frequencies.

please correct me if i am wrong.

Hopefully wierd stuff don't happen to the layout

can i keep your photo around all my installations as an inspiration??

are you having a clear LOS for 150kms? because i thought the earth's curvature spoils RF signals when you work in the GHz frequencies.

please correct me if i am wrong.

That is true.

I haven't checked LOS when curvature is calculated.

For this link I only want stability, the tower at B are providing internet to mostly farmers with voip, since there is no cellphone reception.

I think it is near perfect because point A has a elevation of 1.9Km and B 1.1Km

Any good tool to check the fresnel clerance?

Hopefully wierd stuff don't happen to the layout

can i keep your photo around all my installations as an inspiration??

are you having a clear LOS for 150kms? because i thought the earth's curvature spoils RF signals when you work in the GHz frequencies.

please correct me if i am wrong.

That is true.

I haven't checked LOS when curvature is calculated.

For this link I only want stability, the tower at B are providing internet to mostly farmers with voip, since there is no cellphone reception.

I think it is near perfect because point A has a elevation of 1.9Km and B 1.1Km

Any good tool to check the fresnel clerance?

if you can PM or Email me the Co-ordinates i can send you the complete terrain details. i use Radio Mobile.

if you can PM or Email me the Co-ordinates i can send you the complete terrain details. i use Radio Mobile.

Co-Ordinates sent

I dint understand. Can ospf double bandwitdh? Or only layer2 bondiing?