Re: [opennms-devel] Address discovery versus internals discovery

--===============8891388861762170189==
Content-Type: multipart/alternative; boundary=000325574d7aa200fa04658f7cbd

--000325574d7aa200fa04658f7cbd
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit

On Fri, Mar 20, 2009 at 12:19 PM, AMT <[email protected]> wrote:

>  I read a post the other day which laid out some of the philosophy of
> OpenNMS with respect to discovery.
>
> It seemed to me that there was a clear distinction between address
> discovery, which is done via pinging, and node internals discovery which is
> done via SNMP and other protocols.  The little bit I have worked with the
> source code seems to confirm that.
>
> I was wondering if you have had any discussions about modifying this
> approach.  Ping sweeps are not going to fly when it comes to IPv6 networks.
> Even in IPv4 networks there are situations (sparse networks) where a crawl
> is much more efficient than a ping sweep.  By crawl I mean going into a node
> and poking around in various MIBs looking for other addresses to discover.
>
> Adopting this approach means that the initial address discovery phase would
> include a combination of ping, SNMP credential discovery (any site can have
> multiple versions of SNMP in operation with multiple valid credentials) and
> the the grouping of addresses by physical node/agent.
>

The real difference here is that 'configuration discovery' is a separate
processng from address/device discovery.  Any mechanism for device discovery
can be added to opennms and the configuration discovery phases doesn't
care.  If you want to scan SNMP arp tables to detect addresses that works
fine.  You just need to send a newSuspect event for each address that you
find and it will be scanned for snmp configuraiton and services.

Matt

--000325574d7aa200fa04658f7cbd
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<br><div class=3D"gmail_quote">On Fri, Mar 20, 2009 at 12:19 PM, AMT <span =
dir=3D"ltr">&lt;<a href=3D"mailto:[email protected]">rkostes@altmantec=
h.com</a>&gt;</span> wrote:<br><blockquote class=3D"gmail_quote" style=3D"b=
order-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; paddin=
g-left: 1ex;">






<div bgcolor=3D"#ffffff">
<div><font face=3D"Arial" size=3D"2">I read a post the other day which laid=
 out some of=20
the philosophy of OpenNMS with respect to discovery.</font></div>
<div><font face=3D"Arial" size=3D"2"></font>=A0</div>
<div><font face=3D"Arial" size=3D"2">It seemed to me that there was a clear=
 distinction=20
between address discovery, which is done via pinging, and node internals=20
discovery which is done via SNMP and other protocols.=A0 The little bit I=
=20
have worked with the source code seems to confirm that.</font></div>
<div><font face=3D"Arial" size=3D"2"></font>=A0</div>
<div><font face=3D"Arial" size=3D"2">I was wondering if you have had any di=
scussions=20
about modifying this approach.=A0 Ping sweeps are not going to fly when it=
=20
comes to IPv6 networks.=A0 Even in IPv4 networks there are situations (spar=
se=20
networks)=A0where a crawl is much more efficient than a ping sweep.=A0 By=
=20
crawl I mean going into a node and poking around in various MIBs looking fo=
r=20
other addresses to discover.</font></div>
<div><font face=3D"Arial" size=3D"2"></font>=A0</div>
<div><font face=3D"Arial" size=3D"2">Adopting this approach means that the =
initial=20
address discovery phase would include a combination of ping, SNMP credentia=
l=20
discovery (any site can have multiple versions of SNMP in operation with=20
multiple valid credentials) and the the grouping of addresses by physical=
=20
node/agent.</font></div>
<div><font face=3D"Arial" size=3D"2"></font></div></div></blockquote><div><=
br>The real difference here is that &#39;configuration discovery&#39; is a
separate processng from address/device discovery.=A0 Any mechanism for
device discovery can be added to opennms and the configuration
discovery phases doesn&#39;t care.=A0 If you want to scan SNMP arp tables t=
o
detect addresses that works fine.=A0 You just need to send a newSuspect
event for each address that you find and it will be scanned for snmp
configuraiton and services.=A0 <br>
<br>
Matt<br><br></div></div>

--000325574d7aa200fa04658f7cbd--


--===============8891388861762170189==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
easily build your RIAs with Flex Builder, the Eclipse(TM)based development
software that enables intelligent coding and step-through debugging.
Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
--===============8891388861762170189==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Please read the OpenNMS Mailing List FAQ:
http://www.opennms.org/index.php/Mailing_List_FAQ

opennms-devel mailing list

To *unsubscribe* or change your subscription options, see the bottom of this page:
https://lists.sourceforge.net/lists/listinfo/opennms-devel
--===============8891388861762170189==--