Hans-Christoph Steiner | 10 Feb 2012 20:52

olsr and the netmask


Maybe I'm overthinking this, but I can't seem to wrap my head around what each computer should have set for
its netmask. It seems to me that each node on the network should have a netmask of 255.255.255.255 since
each node has an entry individually.  But then I see that mesh networks use a defined set, like
10.10.0.0/16.  Is that just to define an IP for a given mesh?  Or should each computer also have a netmask of 255.255.0.0?

.hc

----------------------------------------------------------------------------

"It is convenient to imagine a power beyond us because that means we don't have to examine our own lives.",
from "The Idols of Environmentalism", by Curtis White

--

-- 
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users

Markus Kittenberger | 10 Feb 2012 21:09
Picon
Picon

Re: olsr and the netmask

in a mesh network, neither .. /16 or /17 or /18 ... /30 or /32 is correct, but usually anything (which is not too big) works..  (if the mesh routing software (e.g. olsrd) does work)


e.g. with a netmask of /32 you infact define that a node has no neighbours! (as its alone in its own network, and there is no network-route to anyone)
and u need a mesh routing software to add known neighbours,.. (and host routes to them)

while e.g. if u use a /24, the node has a network route to this /24, and thinks it has up to 255 neighbours (in its subnet) which are one hop away (which is usually wrong!)
so you need to add a mesh routing software to add routes (via correct direct neighbours) to the nodes that are multiple hops away (else this node will try and (most often) fail to communicate directly with all neighbours within the subnet)

main difference (regarding unwanted side effects), in the latter case you have unwanted packets (e.g. many/repeated stupid arp-requests), while in the first u have no change to communicate from one hop to its direct neighbour if the mesh routing software e.g. crashed

i hope this clarified things, if not:

yes u are overthinking,... (-;

Markus

(btw did u mange to ping between 2 nodes (without olsrd) via adhoc already?)

On Fri, Feb 10, 2012 at 8:52 PM, Hans-Christoph Steiner <hans <at> guardianproject.info> wrote:

Maybe I'm overthinking this, but I can't seem to wrap my head around what each computer should have set for its netmask. It seems to me that each node on the network should have a netmask of 255.255.255.255 since each node has an entry individually.  But then I see that mesh networks use a defined set, like 10.10.0.0/16.  Is that just to define an IP for a given mesh?  Or should each computer also have a netmask of 255.255.0.0?

.hc

----------------------------------------------------------------------------

"It is convenient to imagine a power beyond us because that means we don't have to examine our own lives.", from "The Idols of Environmentalism", by Curtis White





--
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users

--

-- 
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users
Hans-Christoph Steiner | 10 Feb 2012 23:54

Re: olsr and the netmask


It sounds like then the the IP addresses in OLSR networks must be
manually set on each computer on order to work.  Basically, you need to
know the ip range and netmask of the nearest neighbors, correct?

My goal is to make a "zeroconf" test network for the Debian package, so
newbies can just "apt-get install olsrd" then "/etc/init.d/olsrd start"
and have some kind of basic mesh working.  Therefore, it seems that
ahcpd is essential for that goal.

.hc

On 02/10/2012 03:09 PM, Markus Kittenberger wrote:
> in a mesh network, neither .. /16 or /17 or /18 ... /30 or /32 is
> correct, but usually anything (which is not too big) works..  (if the
> mesh routing software (e.g. olsrd) does work)
> 
> e.g. with a netmask of /32 you infact define that a node has no
> neighbours! (as its alone in its own network, and there is no
> network-route to anyone)
> and u need a mesh routing software to add known neighbours,.. (and host
> routes to them)
> 
> while e.g. if u use a /24, the node has a network route to this /24, and
> thinks it has up to 255 neighbours (in its subnet) which are one hop
> away (which is usually wrong!)
> so you need to add a mesh routing software to add routes (via correct
> direct neighbours) to the nodes that are multiple hops away (else this
> node will try and (most often) fail to communicate directly with all
> neighbours within the subnet)
> 
> main difference (regarding unwanted side effects), in the latter case
> you have unwanted packets (e.g. many/repeated stupid arp-requests),
> while in the first u have no change to communicate from one hop to its
> direct neighbour if the mesh routing software e.g. crashed
> 
> i hope this clarified things, if not:
> 
> yes u are overthinking,... (-;
> 
> Markus
> 
> (btw did u mange to ping between 2 nodes (without olsrd) via adhoc already?)
> 
> On Fri, Feb 10, 2012 at 8:52 PM, Hans-Christoph Steiner
> <hans <at> guardianproject.info <mailto:hans <at> guardianproject.info>> wrote:
> 
> 
>     Maybe I'm overthinking this, but I can't seem to wrap my head around
>     what each computer should have set for its netmask. It seems to me
>     that each node on the network should have a netmask of
>     255.255.255.255 since each node has an entry individually.  But then
>     I see that mesh networks use a defined set, like 10.10.0.0/16
>     <http://10.10.0.0/16>.  Is that just to define an IP for a given
>     mesh?  Or should each computer also have a netmask of 255.255.0.0?
> 
>     .hc
> 
>     ----------------------------------------------------------------------------
> 
>     "It is convenient to imagine a power beyond us because that means we
>     don't have to examine our own lives.", from "The Idols of
>     Environmentalism", by Curtis White
> 
> 
> 
> 
> 
>     --
>     Olsr-users mailing list
>     Olsr-users <at> lists.olsr.org <mailto:Olsr-users <at> lists.olsr.org>
>     https://lists.olsr.org/mailman/listinfo/olsr-users
> 
> 

--

-- 
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users

Henning Rogge | 11 Feb 2012 09:34

Re: olsr and the netmask

No, you can also just configure a complete subnet on the interface. It
will just cost you a bit performance because your nodes will do an
additional ARP for unknown addresses. But it doesn't matter that much.

You need to configure the address of each node (or determine it from
some 'autoconfiguration') for the routing daemon to work. If you have
this address, you can use it to set up the network interface too.

Henning Rogge

On Fri, Feb 10, 2012 at 23:54, Hans-Christoph Steiner
<hans <at> guardianproject.info> wrote:
>
> It sounds like then the the IP addresses in OLSR networks must be
> manually set on each computer on order to work.  Basically, you need to
> know the ip range and netmask of the nearest neighbors, correct?
>
> My goal is to make a "zeroconf" test network for the Debian package, so
> newbies can just "apt-get install olsrd" then "/etc/init.d/olsrd start"
> and have some kind of basic mesh working.  Therefore, it seems that
> ahcpd is essential for that goal.
>
> .hc
>
> On 02/10/2012 03:09 PM, Markus Kittenberger wrote:
>> in a mesh network, neither .. /16 or /17 or /18 ... /30 or /32 is
>> correct, but usually anything (which is not too big) works..  (if the
>> mesh routing software (e.g. olsrd) does work)
>>
>> e.g. with a netmask of /32 you infact define that a node has no
>> neighbours! (as its alone in its own network, and there is no
>> network-route to anyone)
>> and u need a mesh routing software to add known neighbours,.. (and host
>> routes to them)
>>
>> while e.g. if u use a /24, the node has a network route to this /24, and
>> thinks it has up to 255 neighbours (in its subnet) which are one hop
>> away (which is usually wrong!)
>> so you need to add a mesh routing software to add routes (via correct
>> direct neighbours) to the nodes that are multiple hops away (else this
>> node will try and (most often) fail to communicate directly with all
>> neighbours within the subnet)
>>
>> main difference (regarding unwanted side effects), in the latter case
>> you have unwanted packets (e.g. many/repeated stupid arp-requests),
>> while in the first u have no change to communicate from one hop to its
>> direct neighbour if the mesh routing software e.g. crashed
>>
>> i hope this clarified things, if not:
>>
>> yes u are overthinking,... (-;
>>
>> Markus
>>
>> (btw did u mange to ping between 2 nodes (without olsrd) via adhoc already?)
>>
>> On Fri, Feb 10, 2012 at 8:52 PM, Hans-Christoph Steiner
>> <hans <at> guardianproject.info <mailto:hans <at> guardianproject.info>> wrote:
>>
>>
>>     Maybe I'm overthinking this, but I can't seem to wrap my head around
>>     what each computer should have set for its netmask. It seems to me
>>     that each node on the network should have a netmask of
>>     255.255.255.255 since each node has an entry individually.  But then
>>     I see that mesh networks use a defined set, like 10.10.0.0/16
>>     <http://10.10.0.0/16>.  Is that just to define an IP for a given
>>     mesh?  Or should each computer also have a netmask of 255.255.0.0?
>>
>>     .hc
>>
>>     ----------------------------------------------------------------------------
>>
>>     "It is convenient to imagine a power beyond us because that means we
>>     don't have to examine our own lives.", from "The Idols of
>>     Environmentalism", by Curtis White
>>
>>
>>
>>
>>
>>     --
>>     Olsr-users mailing list
>>     Olsr-users <at> lists.olsr.org <mailto:Olsr-users <at> lists.olsr.org>
>>     https://lists.olsr.org/mailman/listinfo/olsr-users
>>
>>
>
> --
> Olsr-users mailing list
> Olsr-users <at> lists.olsr.org
> https://lists.olsr.org/mailman/listinfo/olsr-users

-- 
"Wo kämen wir hin, wenn alle sagten, wo kämem wir hin, und niemand
ginge, um einmal zu schauen, wohin man käme, wenn man ginge." (Kurt
Marti)

--

-- 
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users

Markus Kittenberger | 11 Feb 2012 09:36
Picon
Picon

Re: olsr and the netmask



On Fri, Feb 10, 2012 at 11:54 PM, Hans-Christoph Steiner <hans <at> guardianproject.info> wrote:

It sounds like then the the IP addresses in OLSR networks must be
manually set on each computer on order to work.  Basically, you need to
know the ip range and netmask of the nearest neighbors, correct?
if u manually set the olsrd broadcast address in configfile, you only have to know this,..
(and it can be 255.255.255.255) if you do this u can use a /32.

if not setting the address manually olsrd uses the one from the interface, and than yes u have to configure the interfaces correctly on the nodes

Markus

My goal is to make a "zeroconf" test network for the Debian package, so
newbies can just "apt-get install olsrd" then "/etc/init.d/olsrd start"
and have some kind of basic mesh working.  Therefore, it seems that
ahcpd is essential for that goal.

.hc

On 02/10/2012 03:09 PM, Markus Kittenberger wrote:
> in a mesh network, neither .. /16 or /17 or /18 ... /30 or /32 is
> correct, but usually anything (which is not too big) works..  (if the
> mesh routing software (e.g. olsrd) does work)
>
> e.g. with a netmask of /32 you infact define that a node has no
> neighbours! (as its alone in its own network, and there is no
> network-route to anyone)
> and u need a mesh routing software to add known neighbours,.. (and host
> routes to them)
>
> while e.g. if u use a /24, the node has a network route to this /24, and
> thinks it has up to 255 neighbours (in its subnet) which are one hop
> away (which is usually wrong!)
> so you need to add a mesh routing software to add routes (via correct
> direct neighbours) to the nodes that are multiple hops away (else this
> node will try and (most often) fail to communicate directly with all
> neighbours within the subnet)
>
> main difference (regarding unwanted side effects), in the latter case
> you have unwanted packets (e.g. many/repeated stupid arp-requests),
> while in the first u have no change to communicate from one hop to its
> direct neighbour if the mesh routing software e.g. crashed
>
> i hope this clarified things, if not:
>
> yes u are overthinking,... (-;
>
> Markus
>
> (btw did u mange to ping between 2 nodes (without olsrd) via adhoc already?)
>
> On Fri, Feb 10, 2012 at 8:52 PM, Hans-Christoph Steiner
> <hans <at> guardianproject.info <mailto:hans <at> guardianproject.info>> wrote:
>
>
>     Maybe I'm overthinking this, but I can't seem to wrap my head around
>     what each computer should have set for its netmask. It seems to me
>     that each node on the network should have a netmask of
>     255.255.255.255 since each node has an entry individually.  But then
>     I see that mesh networks use a defined set, like 10.10.0.0/16
>     <http://10.10.0.0/16>.  Is that just to define an IP for a given
>     mesh?  Or should each computer also have a netmask of 255.255.0.0?
>
>     .hc
>
>     ----------------------------------------------------------------------------
>
>     "It is convenient to imagine a power beyond us because that means we
>     don't have to examine our own lives.", from "The Idols of
>     Environmentalism", by Curtis White
>
>
>
>
>
>     --
>     Olsr-users mailing list
>     Olsr-users <at> lists.olsr.org <mailto:Olsr-users <at> lists.olsr.org>
>     https://lists.olsr.org/mailman/listinfo/olsr-users
>
>

--

-- 
Olsr-users mailing list
Olsr-users <at> lists.olsr.org
https://lists.olsr.org/mailman/listinfo/olsr-users

Gmane