RE: BGP sync - Why? Why not?

Thank you all. I appreciate the clarification.

--Hammer

"I was a normal American nerd."
-Jack Herer

From: karim jamali [mailto:karim.jamali_at_gmail.com]
Sent: Monday, November 22, 2010 9:38 AM
To: --Hammer--
Cc: Mirco Orlandi; miken miken; Cisco certification
Subject: Re: BGP sync - Why? Why not?

Hi Hammer,

In some cases, you do not need synchronization. If you do not pass traffic
from a different AS through your AS, you can disable synchronization. You
can also disable synchronization if all routers in your AS run BGP. I have
attached a good link about the subject. Other than those reasons i don't
find any reason to disable synchronization.

http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a00800c95bb
.shtml#synch

Regards,

On Mon, Nov 22, 2010 at 5:21 PM, --Hammer-- <bhmccie_at_gmail.com> wrote:

OK. I don't see where redistribution plays into this. "no synch" simply
allows the EGP to advertise a route to a peer w/o it being in the route
table yet from a static or IGP. My point is that it is simply a gut check.
If I peer up and check my routes (with synchronization on) and I don't see a
route I'm looking for, I can very easily surmise that I have failed to
introduce it via my IGP to the proper routing table. If I have
synchronization off and I see the lack of an expected route, I have to
troubleshoot differently. I'm trying to understand why it seems the vendors
encourage it to be disabled. Is it a standard policy in the real lab? So we
are just following suite? Or am I missing some adverse side affect.

Again, in the big world, I understand. But this isn't the big world. This is
the lab.

--Hammer

"I was a normal American nerd."
-Jack Herer

From: Mirco Orlandi [mailto:mirco.orlandi_at_gmail.com]
Sent: Friday, November 19, 2010 4:14 PM
To: --Hammer--
Cc: miken miken; Cisco certification

Subject: Re: BGP sync - Why? Why not?

Bgp synchronization tipically requires redistribution of bgp into igp.
Probably, (not sure) in your lab scenario you will have more than one igp
and a redistribution plan.
L3 loops detection/prevention is a brain-intensive task. Taking care of
extra external routes may result in wasting time (bgp->igp1->igp2->igp1 =
loop).

-mirco.

On Fri, Nov 19, 2010 at 5:47 PM, --Hammer-- <bhmccie_at_gmail.com> wrote:

Absolutely. For real world (Internet, big WAN, etc.) I can see it as
suboptimal. I'm just struggling with it in the context of practice labs and
the impending real lab.

--Hammer

"I was a normal American nerd."
-Jack Herer

From: miken miken [mailto:miken_at_sisna.com]
Sent: Friday, November 19, 2010 10:13 AM
To: --Hammer--
Cc: Cisco certification
Subject: Re: BGP sync - Why? Why not?

The simplest explanation I have found for real world scenarios is given by
Philip Smith from Cisco in his presentation that can be found on the NANOG
website. He states that it is not feasible to wait for the ~222,000 prefixes
to synchronize in IGP (OSPF/ISIS) before BGP will select the path. You can
hear this explanation about 1:11 into the presentation. However, I do not
know how to apply this to the lab, other than best practice??

Thank you
MikeN

On Fri, Nov 19, 2010 at 8:39 AM, --Hammer-- <bhmccie_at_gmail.com> wrote:

OK, I understand what synchronization is and how it works. So, all the
videos/audios/labs/etc. always say to explicitly disable it (I know it's
disabled by default). What I'm struggling with is why? If the end all result
of a lab is "full" reachibility and the lab doesn't disallow
synchronization, then isn't it helping us? I know I'm missing the big point
here but I don't see why it's a bad thing. Can someone enlighten me? The
byproduct would be that your troubleshooting might be a little different..

--Hammer

"I was a normal American nerd."
-Jack Herer

Blogs and organic groups at http://www.ccie.net
Received on Mon Nov 22 2010 - 10:21:21 ART