Skip to contentSkip to conversation

Will advanced linux routing help me

LARTC Sushil Suresh 3 participants 8 messages Apr 25, 2001 – Apr 27, 2001

lore
Share & tools
lore
#0 lore raw

Hi,

    I have a private network which is connected to 3 diffrent ISPs. I also have a linux box acting as a gateway and internet firewall for the internal network. The network is as shown below


          Linux Gateway/Firewall
                      |                
                      |
                      |
              Ethernet Hub ----------------------------------------------------------------------Ethernet Hub--------|--
             /        |           \                                                                                                |--   Local Network
          /           |             \                                                                                              |--
       /              |                \

ISP-I ISP-II ISP-III

My problem is i have a couple of sites which are to be divided into 3 groups and each to be routed through a particular isp ( this is generally choosen by the requirements of the site. Some of the sites need very fast response time, some need large volumes of data to be transfered, and some need a very stable connection ).

Appart from these few sites the rest of the traffic is routed through one of the isp ( the one with the highest bandwidth )

And when a particular ISP fails the routes currently being used by this ISP is to be shifted to the other 2 ISPs.

I also require the routes to be restord back to as decided earlier once the isp which was down comes back

Is there any means to automise the whole process. The effective aim is to utilize the most of all the isps that are available at a moment of time.

( At the moment we attain this via static routing and we have one personal monitoring this round the clock, in case a rerouting is necessary )

with regards,
Sushil Suresh

#1 lore raw

On Wed, 25 Apr 2001, Sushil Suresh wrote:

[snip, automatic rerouting when interfaces are downed]

Quoted 6 lines

Is there any means to automise the whole process. The effective aim is to utilize the most of all the isps that are available at a moment of time.

( At the moment we attain this via static routing and we have one personal monitoring this round the clock, in case a rerouting is necessary )

Okay, the thing to do is to figure out what this person is doing to check if the links are up. Once you know this, you put that knowledge into a script that checks if the link is up and run that script from cron. Depending on the status of the uplinks you then let the script select one of the possible static routings and enable it. (Rebuilding routing tables barely hurts reliability, *especially* if you do not flush the route cache).

Hope this helps.

Doei, Arthur. (The tricky part is the link-state monitoring... at the TCP/IP
level this is theoretically impossible...)

Signature · 8 lines
  /\    / |      [email protected]      | Work like you don't need the money
 /__\  /  | A friend is someone with whom | Love like you have never been hurt
/    \/__ | you can dare to be yourself   | Dance like there's nobody watching


_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/
#2 lore raw
In reply to Sushil Suresh
Quoted 2 lines · Arthur van Leeuwen

Doei, Arthur. (The tricky part is the link-state monitoring... at the TCP/IP
level this is theoretically impossible...)

Yes, indeed. There is no special protocol for this. That's why OSPF and his younger and older brothers have those HELO packets. That's why TCP has the optional keepalive segments (although, in case of TCP it's an end-to-end behaviour).

If you know the IP of the next hop router and if that router has not been instructed to ignore ICMP requests, ping might help. If you don't know the IP, pinging to the subnet broadcast will also do.

Ramin

Quoted 10 lines
Signature · 9 lines
-- 
  /\    / |      [email protected]      | Work like you don't need the money
 /__\  /  | A friend is someone with whom | Love like you have never been hurt
/    \/__ | you can dare to be yourself   | Dance like there's nobody watching


_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

#3 lore raw
In reply to Sushil Suresh

Hi,

OSFP is a dynamic routing protocol, isn't it. So is dynamic routing
going to solve my problem.

The problem is i do not want to use the best possible path ( cost vise ) to route a packet. I have to make use of all the availabe bandwitdh of the 3 ISP's in the most efficient manner.

I have read about TEQ traffic equalizer. But it has been mentioned that you cannot combine 3 different isps to act as one.
As if you are using TEQ at one end, you are bound to have it at the other end.

with regards,
Sushil Suresh

----- Original Message -----
From: Ramin Alidousti <[email protected]>
To: Arthur van Leeuwen <[email protected]>
Cc: Sushil Suresh <[email protected]>; <[email protected]>
Sent: Wednesday, April 25, 2001 6:20 PM
Subject: Re: [LARTC] Will advanced linux routing help me

Quoted 3 lines · Arthur van Leeuwen
Quoted 2 lines

Doei, Arthur. (The tricky part is the link-state monitoring... at the

TCP/IP

Quoted 7 lines
Quoted 1 line

level this is theoretically impossible...)

Yes, indeed. There is no special protocol for this. That's why OSPF and his younger and older brothers have those HELO packets. That's why TCP has the optional keepalive segments (although, in case of TCP it's an end-to-end behaviour).

Quoted 9 lines

If you know the IP of the next hop router and if that router has not been instructed to ignore ICMP requests, ping might help. If you don't know the IP, pinging to the subnet broadcast will also do.

Ramin

Quoted 3 lines

--
/\ / | [email protected] | Work like you don't need the

money

Quoted 1 line
Quoted 1 line

/__\ / | A friend is someone with whom | Love like you have never

been hurt

Quoted 1 line
Quoted 1 line

/ \/__ | you can dare to be yourself | Dance like there's nobody

watching

Quoted 5 lines
Quoted 5 lines

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO:

http://ds9a.nl/2.4Routing/

Quoted 1 line

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

#4 lore raw
In reply to Sushil Suresh

Hi,

At the moment we are using trace and ping to check the availability of a
particular site, and the status of a particular ISP

The problem faced in making use of these commands in a script is that, they may fail in case of congestion and may be misunderstood as failure of a particular ISP.

Also how and why do you flush the route cache?

with regards,
Sushil Suresh

----- Original Message -----
From: Arthur van Leeuwen <[email protected]>
To: Sushil Suresh <[email protected]>
Cc: <[email protected]>
Sent: Wednesday, April 25, 2001 1:09 PM
Subject: Re: [LARTC] Will advanced linux routing help me

Quoted 10 lines · Sushil Suresh

[snip, automatic rerouting when interfaces are downed]

Quoted 6 lines

Is there any means to automise the whole process. The effective aim is to utilize the most of all the isps that are available at a moment of time.

( At the moment we attain this via static routing and we have one personal monitoring this round the clock, in case a rerouting is

necessary )

Quoted 2 lines

Okay, the thing to do is to figure out what this person is doing to check

if

Quoted 3 lines

the links are up. Once you know this, you put that knowledge into a script that checks if the link is up and run that script from cron. Depending on the status of the uplinks you then let the script select one of the

possible

Quoted 6 lines

static routings and enable it. (Rebuilding routing tables barely hurts reliability, *especially* if you do not flush the route cache).

Hope this helps.

Doei, Arthur. (The tricky part is the link-state monitoring... at the

TCP/IP

Quoted 4 lines

level this is theoretically impossible...)

--
/\ / | [email protected] | Work like you don't need the

money

Quoted 1 line

/__\ / | A friend is someone with whom | Love like you have never been

hurt

Quoted 1 line

/ \/__ | you can dare to be yourself | Dance like there's nobody

watching

Quoted 1 line

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

#5 lore raw
In reply to Sushil Suresh
Quoted 2 lines · Sushil Suresh

At the moment we are using trace and ping to check the availability of a
particular site, and the status of a particular ISP

Yes, that is the only thing you can do.

Quoted 3 lines

The problem faced in making use of these commands in a script is that, they may fail in case of congestion and may be misunderstood as failure of a particular ISP.

Unfortunately, this cannot be helped, unless you have some layer-2 indication of the link still being up (something like carrier detect on modems). This is not usually the case, however.

Quoted 1 line

Also how and why do you flush the route cache?

The why is trivial: to make sure the routing code starts off with a clean slate and you're not surprised by routes you think you have deleted.

The how is equally simple: ip route flush table cache

Doei, Arthur.

Signature · 8 lines
  /\    / |      [email protected]      | Work like you don't need the money
 /__\  /  | A friend is someone with whom | Love like you have never been hurt
/    \/__ | you can dare to be yourself   | Dance like there's nobody watching


_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/
#6 lore raw
In reply to Sushil Suresh
Quoted 8 lines · Sushil Suresh

Hi,

At the moment we are using trace and ping to check the availability of a
particular site, and the status of a particular ISP

The problem faced in making use of these commands in a script is that, they may fail in case of congestion and may be misunderstood as failure of a particular ISP.

Your script should not rely on only one missed ping. You should add some intelligence to the script.

BTW, what do you do with trace?

Quoted 2 lines

Also how and why do you flush the route cache?

Arthur covered this just fine ;-)

Ramin

Quoted 3 lines

with regards,
Sushil Suresh

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/

#7 lore raw
In reply to Sushil Suresh
Quoted 4 lines · Sushil Suresh

Hi,

OSFP is a dynamic routing protocol, isn't it. So is dynamic routing
going to solve my problem.

No. You cannot run a routing protocol on your own. It needs the ISP's co-operation.

Quoted 4 lines

The problem is i do not want to use the best possible path ( cost vise ) to route a packet. I have to make use of all the availabe bandwitdh of the 3 ISP's in the most efficient manner.

TC is the way to go !!

Ramin

Quoted 8 lines

I have read about TEQ traffic equalizer. But it has been mentioned that you cannot combine 3 different isps to act as one.
As if you are using TEQ at one end, you are bound to have it at the other end.

with regards,
Sushil Suresh

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/