We enhanced Mikrotik's implementation of Cisco's OSPF protocol with scripts linking signal quality to route cost; therefore insuring that signal fade on the active station would not disrupt internet traffic.
The scripts run every 6 seconds and change route cost according to tx and rx signal levels.
Thursday, May 17, 2007
The multipath solution
We erected multiple links in different locations and altitudes on the Lebanese side.
This we believe creates multiple physical paths to the Cyprus station therefore when signal fade is large on one of the stations (one of the paths) the other station would take over the traffic.
This we believe creates multiple physical paths to the Cyprus station therefore when signal fade is large on one of the stations (one of the paths) the other station would take over the traffic.
Big and small challenges.
Although the challenges at the beginning looked impressive, the truth was that everything done so far was the easy part, aligning the antennas was not easy, but did not prove hard.
Later on when this link was to be commercially tested, the real challenge surfaced. How to make it commercially stable?
The biggest problem it turned out is signal fade, immense unpredictable seasonal signal level variations that make the link very unstable especially in hot humid days.
Later on when this link was to be commercially tested, the real challenge surfaced. How to make it commercially stable?
The biggest problem it turned out is signal fade, immense unpredictable seasonal signal level variations that make the link very unstable especially in hot humid days.
The equipment, software, etc...
On both sides of the border we were most comfortable with Mikrotik, 2 routerboard rb532 with Ubiquity Networks SR2 cards were selected. Mikrotik's nstreeme wireless protocol solved the ack timeout problem.
The antenna on the Cypriot side is a 24dB grid centered at 2452 mhz.
Antenna on the Lebanon side, a 29dB grid (above).
The antenna on the Cypriot side is a 24dB grid centered at 2452 mhz.
Antenna on the Lebanon side, a 29dB grid (above).
The work
Lebanon August 2006, we assembled a station in Ehden 1600 meters above sea level.
We pointed to cyprus, and to our surprise we detected one Cypriot service set identifier.
After many improvements we detected more ssid's; signal levels in the range -90dB to -80dB
We connected to one of the ssid's and obtained a few pings, however the Atheros ack timeout
issue still needed to be resolved.
We pointed to cyprus, and to our surprise we detected one Cypriot service set identifier.
After many improvements we detected more ssid's; signal levels in the range -90dB to -80dB
We connected to one of the ssid's and obtained a few pings, however the Atheros ack timeout
issue still needed to be resolved.
Wednesday, January 17, 2007
How it all started (236 km wireless link)
2006 Lebanon War, known in Lebanon as the July War and in Israel as the Second Lebanon War.
During the war many telecom towers were hit, internet services in the north of Lebanon were disrupted.
We started contacting wireless ISP's in cyprus to check if any one was interested in attempting the longest commercial wireless shoot ever . Probably the longest wireless link over sea water.
To our surprise, Turkish Cypriots were very interested Kemal, co owner of Extendbroadband expressed his interest. Work was started some wireless calculations were done and proved that indeed the link was possible.
During the war many telecom towers were hit, internet services in the north of Lebanon were disrupted.
We started contacting wireless ISP's in cyprus to check if any one was interested in attempting the longest commercial wireless shoot ever . Probably the longest wireless link over sea water.
To our surprise, Turkish Cypriots were very interested Kemal, co owner of Extendbroadband expressed his interest. Work was started some wireless calculations were done and proved that indeed the link was possible.
Subscribe to:
Posts (Atom)