erg.cs.waikato.ac.nz provides some historical info that can help with this type of question. I.E. it can tell you if the performance is the same today as it was last week and, if not, when it changed and if there was a path change associated with that. AMP supports http tests although we don't run then routinely to many sites. If you wanted more, we may be able to add them. Tony On 28/02/12 09:06, Glen Eustace wrote:
In my role at Massey University, I am often asked to investigate complaints from our Help Desk of a 'the Internet is slow' nature. We are finding this increasingly difficult to do in any meaningful way. Massey is peered at WIX, PNIX and APE we have two Internet peers and have multiple paths to them and they both have multiple upstream providers.
Simply ping testing for latency/connectivity doesn't really provide much of an insight, traceroute tells us where outbound packets went and which hops are 'slow' but doesn't indicate the return path. Most of our customers equate Website == Internet, so the responsiveness of a destination seems to also be an important consideration.
I am assuming that most ISPs have Help Desk calls of a similar nature. How does one substantiate or refute such complaints, pro-actively identify 'slowness/congestion' ? Evidence collected by 'tools' needs to be defensible when responding to our 'customers'.
Any pointers to how I can do this better would be appreciated.
Glen Massey University. _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog