[Sysops-list] Re: [npix-wg] intermittent problems at npix

[Ritesh Raj Joshi]

Thanks for your input, Gaurab.

I just wrote to the sysops-list as the problem still remains unsolved.

Btw, as a well-wisher of npix, I would like to suggest something to all 
npix stakeholders.

I have seen in the past that there has been routing issues on several 
occasions among some of the ISPs. We are so used to having fast access 
of resources (thanks to npix) that I get dissappointed when I get a 
sluggish access to resources at another ISP. I have seen that such 
problems sometimes takes weeks to resolved...and everybody suffers poor 
access speed for the duration of the problem. NPIX is only going to grow 
and become a very mission-critical infrastructure for the Internet 
access within Nepal. Hence it is important that there be some better 
management of npix at all levels.

I propose that there be some kind of a coordination/response team that 
periodically monitors and does pre-emptive health check of the routing 
and other performance issues of npix as a whole. This could be a group 
of guys who are well versed with ISP routing stuff. They will be 
responsible to take corrective action of any problems they sense on the 
npix facility with cooperation from all concerned members....and resolve 
the issues at the earliest. The group can comprise of technical reps 
from the current npix member ISPs as I think it will be difficult to 
find experienced people, who are fit for the job, outside of ISPs.
Later on this can also be developed into a NOC when the demand arises.

These are just my thoughts...pls feel free to comment.

Ritesh

Gaurab Raj Upadhaya wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hi,
> 
> I don't think this is a problem with npix. :-). the switch is working  
> just fine.
> 
> Also, for these kind of stuff, i would suggest using the new mailing  
> list setup for this purpose..
> http://lists.npix.net.np/mailman/listinfo/sysops-list
> 
> back to point, I would say that this is a problem with routing inside  
> Mercantile/Vianet. From what it looks like it's asynchronous routing  
> issue. Maybe MOS are not seeing vianet's prefixes. A trace from  infocom 
> looks good to me
> 
> [gaurab at ns gaurab]$ /usr/sbin/traceroute www.mos.com.np
> traceroute to mos.com.np (202.52.255.35), 30 hops max, 38 byte packets
> 1  gw.info.com.np (203.78.160.1)  133.178 ms  125.532 ms  109.423 ms
> 2  203.78.160.205 (203.78.160.205)  95.391 ms  9.631 ms  7.256 ms
> 3  202.52.228.185 (202.52.228.185)  6.399 ms  6.781 ms  7.971 ms
> 4  serial-0-1-ktp-rt1.mos.com.np (202.52.228.13)  9.307 ms  9.412 ms   
> 9.761 ms
> 5  fast-ether-1-1-dmg-3640.mos.com.np (202.52.233.254)  16.837 ms   
> 24.711 ms  20.425 ms
> 6  * * *
> 
> Better for Vianet and Mos ops people to work together to resolve  this. 
> There is nothing others can do.
> 
> 
> thanks
> 
> 
> On Dec 6, 2005, at 11:04 AM, Ritesh Raj Joshi wrote:
> 
>> Hi friends,
>> I have been observing intermittent problems among some peers at  npix. 
>> I am not sure where the problem lies. It could be routing  issues or 
>> bandwidth bottlenecks on the way. I get terrible speeds  when I try to 
>> POP my mails at office from Mercantile server. Btw, I  use a broadband 
>> link from Vianet. I have informed about the issue  to the concerned 
>> ISPs at many occasions but I am not sure if they  are yet able to find 
>> out the cause and a permanent solution as the  problem still remains 
>> the same. I am writing this mail as I am  POPing 73 mails from MC 
>> server and I am struggling to on the 27th  mail. When the connection 
>> used to be normal I could POP the mails  in no time.
>>
> 
>            -- gaurab
> 
> 
> /////////////////////////////////////////////////////+9779851038080
> gaurab at lahai dot com
> 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1 (Darwin)
> 
> iD8DBQFDlXdYSo7fU26F3X0RAulUAKDJ/JlR6oSF2rQRf1EdpYccrezzpACg0aYo
> ku4uAxfOO5Cxk/tQ5xqeiy0=
> =onti
> -----END PGP SIGNATURE-----
> 
>