Karlnet
[Top] [All Lists]

[Karlnet] Loosing our bandwidth when we lost the copper!

To: "Karlnet Mailing List" <karlnet@WISPNotes.com>
Subject: [Karlnet] Loosing our bandwidth when we lost the copper!
From: jna <jna@tcpbbs.net>
Reply-to: Karlnet Mailing List <karlnet@WISPNotes.com>
Date: Sun, 16 Nov 2003 01:50:37 -0500
List-post: <mailto:karlnet@WISPNotes.com>
PROBLEM: We are now only getting 400-800K on a 3MB connection out!?

Our main tower AP1000 was being fed from our uplink provider to our tower
via a T1 loop from there data center to our tower. Our wireless links from
that tower has always been able to achieve around a good 1.3 to 1.4 Internet
bandwidth obviously tapped by the limit of th 1.5 on the T1. Never the less
we needed to move from the T1 for the inability of the local telco to keep
it operations, etc... So we went completely wireless! We did not have line
of sight to our main tower from the datacenters building so to get around
this we ended placing a relay on the top of a tall building that did have
line of sight to both locations. So our basic setup now is we have an Rg1000
(Running Sattelite) on top of the data center with the ethernet plugged into
a a switch in the data center which allots us 3 MBPS to that drop. This is
pointed at the relay location using a 19db grid. The relay location has a
KN-200 installed with both interfaced installed in ISP base station mode.
One is attached to another 19db grid pointed back at the Rg1000 at the
datacenter which is feeding the relay unit the bandwidth. The other is
directed to our tower which now has another RG1000 (Running Satellite) on a
24db grid (Longer link) that RG is plugged into the switch that the AP1000
is plugged into and is what is feeding that location. So our setup now looks
like this

[3 MEG Switch Port Backbone] --> Ethernet Cable-->RG1000
(Satellite)-->KN-200 (ISP BASE RELAY)-->RG1000(Satellite)-->Switch-->AP1000
on tower (ISP BASE TO CUSTOMERS)

Now that we went to a faster backbone with twice the bandwidth our best
connection to the net is around 800K with 400K being the average!? We were
getting better tranfers on the T1. I plugged my laptop into the switch at
the datacenter that the RG that is acting as the gateway to the wireless
portion of our network is plugged in and get 2.8MBPS out on the net so
obviously it is past the switch and somewhere between the RG on the
datacenter out.

I used the ping fill utility and calculated that the RG on the data center
that is feeding the relay point has and average radio transfer at 2.85MBPS.
The relay to the tower link had a radio transfer average of 3.4MBPS so there
is plenty of over the air  bandwidth to each location to support faster net
output than what we are getting.

I suspect that it may be a routing issue? Our entire network was bridged and
the gatweway out was our cisco router. Since we went wireless backhaul to
the data center the cisco went out of the pic and we had to make the radio
on the roof of the datacenter use IP routing instead of bridging. Again can
a mistake in the routing loose us that much throughoutput? We have only two
routes setup in the datacenter RG:

Default router IP 216.223.169.70 (Our gateway out to the net as provided by
our provider)
Prefered router IP 216.223.169.65 (Our RG Satellite IP)
Default TTL 255

Direct Route 1 : 12.96.250.2 255.255.255.192 Interface 2 (This is our IP
block that all other units/customers are on)
Direct Route 2:  216.223.169.65 255.255.255.248 Interface 1 (This is the
small block assigned to our local equipment at the datacenter that the
interface on there router is assigned to i.e. our gateway 216.223.169.70 is
on there router.

Is this routing correct?? Is this were the problem is, if so what is wrong?
I attached a cap of our routing screen here I do not know if this list
supports caps?

Thanks in advance for any help!!

John
_______________________________________________
Karlnet mailing list
Karlnet@WISPNotes.com
http://lists.wispnotes.com/mailman/listinfo/karlnet
<Prev in Thread] Current Thread [Next in Thread>