On 30/12/13 01:11, Man_Without_Clue wrote: > Ok, > > Here are what I have done though I really have lost track of things
You might find it easier to follow if you used interleaved posting:- https://wiki.debian.org/DebianMailingLists#Posting_Rules.2C_Guidelines.2C_and_Tips http://en.wikipedia.org/wiki/Posting_style#Interleaved_style and plain text format:- https://wiki.debian.org/DebianMailingLists#HowTo_send_plain_text_emails_to_the_list I find it less difficult to follow threads that way. It's worth learning IMO :) > I've > done as I just follow what I could find... > <snipped> > > #netstat -tunlp |grep p6 |wc -l > > > This returns 0. Thanks. Just wanted to check you hadn't been following borked instructions that slowed your network. I agree with Greg Nowak, except in embedded and other resource limited situations the IPv6 stack isn't going to slow your network. On the other hand you might get a slower network if IPv6 is delivered and you're not using it.... (you didn't say). > > > > Now, I did following as you suggested, > > > ping -c 5 google.com <snipped> Those times aren't terrible. Though "speed" and "service satisfaction" are different things. You'd have to give some information about what your ISP has told you to expect before the service can be measured. > > > > also, > > cat /etc/resolv.conf > # Generated by NetworkManager > nameserver 192.168.1.254 > > ## 192.168.1.254 is the IP for the router. Unless your router is doing DNS caching it's just a continuous relay for another DNS. You'd have to check your "router" (modem/hub?) settings to see what that DNS is - by default that's normally set to your ISP's DNS. Generally their DNS is going to be the fastest (though not necessarily the most up-to-date). Choice of DNS server can make a significant difference to what's perceived as "internet speed*1". *1 Generally, a user's perception of "internet speed" (click) means "action time" (graphic result) - "response time" and can be a struggle to correlate "throughput time" to that. Those on-line "speed" tests give a poor indication of throughput time anyway. Loading a web page:- ;ask DNS for IP address (check cache first if one exists, then try DNS Nameserver entries in order) ;ask ISP for route to IP address ;ask server at IP address for page ;read page ;ask DNS for IP addresses for page components ;ask IP address for page components ;rinse and repeat as needed ;if not served ISP pre-cached, if Vodaphone may route page and components through "squasher" to compress pictures and code ;add in browser extension, malware detection and firewall processes, etags and other factors ;render results to screen So a "fast", low-latency internet connection can still appear slow due to poor page design, bad firewall and anti-malware systems, buggy extensions, video settings and hardware, DE settings and constraints (swappiness), and DNS settings. You don't say what it is that isn't as fast as you'd like and that's important because it may be something that only QOS can fix (i.e. VOIP). For your browser disable any unused extensions. Do install NoScript, FlashBlock and AdBlock if you use Iceweasel. Disable any networking desktop widgets/eyecandy. Disable any network applications. Install namebench:- # apt-get install namebench man namebench Optimise your DNS settings using namebench (it's got a simple to use GUI and good documentation). Then selectively re-enable any applications you use that use the network (weather widgets, rss readers, bittorrents, VOIP etc), testing each time with ping, curl and tcptraceroute to see what effects they have on your network. Note that your ISP's network latency and bandwidth may change considerably throughout the day. Ping won't test the speed of the route taken by pages. tcptraceroute will. # apt-get install tcptraceroute Test your internet connection. Ping, curl (or wget), and tcptraceroute will give you some useful metrics, but they won't tell you if your ISP is throttling certain types of traffic. And online "internet speed tests" are useless because they only test "total time to tranfer". For detailed analysis and measurement of your internet connection use:- http://mlab-live.appspot.com/tools/ndt http://code.google.com/p/ndt/source/checkout Let us know the results ;) If you find problems try the transparency and last mile tests elsewhere on the site. If you want to do longer term monitoring look at installing nubot. > > > > route > Kernel IP routing table > Destination Gateway Genmask Flags Metric Ref Use > Iface > default 192.168.1.254 0.0.0.0 UG 0 0 0 eth0 > localnet * 255.255.255.0 U 0 0 0 eth0 All good > > > > ifconfig $dev > eth0 Link encap:Ethernet HWaddr 00:1c:c0:19:ba:33 > inet addr:192.168.1.15 Bcast:192.168.1.255 Mask:255.255.255.0 > UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 > RX packets:325430 errors:0 dropped:23 overruns:0 frame:0 You've dropped a few packets there. Nothing much to worry about given the percentage. <snipped> > > > I could get up to here... > From "curl" thing, I get very long response tagged with <html>.... That's just the actual HTML from the page. What we're interested in is the very last line. e.g.:- 3.304:4.089:4.319:13401 Which, in my case, is acceptable. <snipped> Kind regards -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/52c25c5f.7050...@gmail.com
