View Single Post
Old 29-10-2015, 16:57   #12
General Maximus
cf.mega poster
 
Join Date: Nov 2008
Location: Lincoln
Services: phone + 1gbit BB + SkyQ
Posts: 11,021
General Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronze
General Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronzeGeneral Maximus is cast in bronze
Re: Cannot access themoviedb.org

Quote:
Originally Posted by Lee View Post
http://54.83.202.158 doesn't work either unfortunately.

I did wonder about routing, but others on here have said it works OK for them and I'd think routing would affect everyone?
nope, at worst it could affect a large proportion of VM's customers by blocking all traffic going out of one of the peering links onto another network but that would affect only the customers geographically situated near that transit point. Most routing issues go mostly unnoticed because they are very local to you and only affect a minority of customers (as in this case because none else has is experiencing the same symptoms). The last one I saw was earlier in the year in north Manchester (I think) where a router went down, everyone lost connectivity and all the traffic had to be rerouted through south Manchester which ended up screwing their router as well and everything in Manchester ended going at snail speed. Only the users served by those routers in Manchester were affected though.

The nice thing about traceroutes is that they give you an idea of the route your traffic take from your pc to a particular destination and more often my traffic will take a completely different route across VM's network to the destination than yours will and sometimes won't even go through the same exit point on VM's network so it is literally like two different paths. The best way to think of a network is like a spiders web and not a tree. We are not a tree where everything ends up going down the same route (the trunk), we are a spiders web where there are multiples routes from A to B and although their may be an optimal router (the most direct), there are lots of routes to get to the same destination. (Another analogy is driving; it might be quicker to go on the motorway to get somewhere but if the motorway is close between certain junctions then you take the A and B roads round it. You still get there but it just takes longer. If you and I are in different locations (north and south of England) and were going to Alton Towers, the chances are that we wouldn't go any of the same roads or motorways and the only common path in our journey would be the last few miles to Alton Towers which is the last hop(s) local to the destinations host network and routers).

I couldn't resist the temptation to do some screen dumps to show you the difference. They are 1-6 left to right:

1) Your initial trace using VM's dns
2) My trace using that ip VM resolved for you
3) My trace using the url (notice google's dns has resolved it to a different ip than VM and PIA)
4) My trace using your VM ip over PIA using a London server
5) Your trace using the ip PIA resolved
6) My trace over PIA using your PIA ip










You can analyse the routes at your leisure but the two things which stand out for me and highlight the differences big time is the ip VM gave you I get to the router in your hop 8 in 5 hops (using my VM connection) and in contrast it takes me 9 hops using your PIA resolved ip to get to what you do in 5 (using PIA). Even though we are both on the same network we take different routes!

(Ps: which ming mong said forum links don't work ?)
General Maximus is offline   Reply With Quote