Slide 1: measurement needs

measurement needs


nms breakout session 
19 apr 02

www.caida.org

Slide 2: measurement needs

measurement needs

workload
longer traces (requires more funding)
tcp/ip headers, w options, other headers if possible
at least 24 hours 
concurrent at several places
several across time for trend analysis
traffic matrices
same constraints as above
locations
disa (rqeuires sri intervention)
enterprise (billc will ask lucent)
university (k will ask ucsd, u.auckland)
backbone isp
peering point
cdn's (akamai)
wireless
requires sri/DISA to provide it
headers, location, signal strength
needs further requirements analysis

Slide 3: measurement needs (2)

measurement needs (2)


active measurements
correlated with passive (planned experiments)
bandwidth estimation (not funded by NMS)
single source vs two end point
available bw vs capacity
routing
BGP updates (requires more funding for equipment/maintenance)
IGP updates (requires contact/release from ISPs)
parameters from ISPs, e.g MRAI
implementation details from vendors

Slide 4: causes of Internet perf. problems/outages

causes of Internet perf. problems/outages 

In roughly the order (courtesy sean donelan nanog post)

1. network engineers (what's this command do?)
2. power failures (what's this switch do?)
3. cable cuts (backhoes, enough said)
4. hardware failures (what's that smell?)             
5. congestion (more bandwidth! Captain, I'm giving you all she's got!)
6. attacks (malicious, you know who you are)
7. software bugs (your call is very important to us....)

        //    ``I prefer the wicked rather than the foolish.
        //    the wicked sometimes rest.'' - alexandre dumas

Slide 5: Problems that remain persistently insolvable

Problems that remain persistently insolvable 
should always be suspected as 
questions asked in the wrong way.     
                      -- Alan Watts

Slide 6: www.caida.org/publications/presentations/

www.caida.org/publications/presentations/


kc claffy
UCSD/SDSC/CAIDA
kc@caida.org
www.caida.org