4 ms·
15 Tools to Help You Develop Faster Web Pages
- sant0sk1 18y agoI read the title wrong and thought it was 15 tools to help me develop web pages faster, not faster web pages :( Oh well, yslow + firebug ftw!
- natch 18y agoDon't forget getting a hosting provider with low latency. Actual hostnames have been changed to protect the guilty and to avoid spamming on behalf of the innocent. If you really want to know, the IP address is there. BAD: $ ping suckyhosting.com PING suckyhosting.com (66.39.3.7): 56 data bytes 64 bytes from 66.39.3.7: icmp_seq=0 ttl=56 time=85.547 ms 64 bytes from 66.39.3.7: icmp_seq=1 ttl=56 time=85.164 ms 64 bytes from 66.39.3.7: icmp_seq=2 ttl=56 time=84.244 ms 64 bytes from 66.39.3.7: icmp_seq=3 ttl=56 time=89.111 ms ^C --- suckyhosting.com ping statistics --- 4 packets transmitted, 4 packets received, 0% packet loss round-trip min/avg/max/stddev = 84.244/86.017/89.111/1.848 ms GOOD: $ ping myprovider.net PING myprovider.net (216.218.186.2): 56 data bytes 64 bytes from 216.218.186.2: icmp_seq=0 ttl=57 time=12.773 ms 64 bytes from 216.218.186.2: icmp_seq=1 ttl=57 time=11.947 ms 64 bytes from 216.218.186.2: icmp_seq=2 ttl=57 time=12.747 ms 64 bytes from 216.218.186.2: icmp_seq=3 ttl=57 time=15.037 ms Geography matters too, but even for locale-specific sites, it may be better to have a US-based provider with super-great network connections (with good access to a MAE, for example), hopefully also redundant connections, than to have a poorly-connected provider closer to the geographic region you are serving.
- boucher 18y agoLow latency is great, but I would take your concerns with a grain of salt. 80ms isn't bad. boucher$ ping google.com PING google.com (64.233.187.99): 56 data bytes 64 bytes from 64.233.187.99: icmp_seq=0 ttl=241 time=83.783 ms 64 bytes from 64.233.187.99: icmp_seq=1 ttl=241 time=83.536 ms 64 bytes from 64.233.187.99: icmp_seq=2 ttl=241 time=81.096 ms 64 bytes from 64.233.187.99: icmp_seq=3 ttl=241 time=87.137 ms 64 bytes from 64.233.187.99: icmp_seq=4 ttl=241 time=84.813 ms
- senthil_rajasek 18y agonot if your browser caches pages...
- rufo 18y agoI'm not sure that you can make these kinds of judgments of hosting providers that easily: $ ping 66.39.3.7 PING 66.39.3.7 (66.39.3.7): 56 data bytes 64 bytes from 66.39.3.7: icmp_seq=0 ttl=54 time=43.215 ms 64 bytes from 66.39.3.7: icmp_seq=1 ttl=54 time=53.490 ms 64 bytes from 66.39.3.7: icmp_seq=2 ttl=54 time=44.041 ms 64 bytes from 66.39.3.7: icmp_seq=3 ttl=54 time=42.953 ms 64 bytes from 66.39.3.7: icmp_seq=4 ttl=54 time=45.672 ms 64 bytes from 66.39.3.7: icmp_seq=5 ttl=54 time=50.263 ms ^C --- 66.39.3.7 ping statistics --- 6 packets transmitted, 6 packets received, 0% packet loss round-trip min/avg/max/stddev = 42.953/46.606/53.490/3.938 ms $ ping 216.218.186.2 PING 216.218.186.2 (216.218.186.2): 56 data bytes 64 bytes from 216.218.186.2: icmp_seq=0 ttl=55 time=99.194 ms 64 bytes from 216.218.186.2: icmp_seq=1 ttl=55 time=105.039 ms 64 bytes from 216.218.186.2: icmp_seq=2 ttl=55 time=97.285 ms 64 bytes from 216.218.186.2: icmp_seq=3 ttl=55 time=99.197 ms 64 bytes from 216.218.186.2: icmp_seq=4 ttl=55 time=98.486 ms 64 bytes from 216.218.186.2: icmp_seq=5 ttl=55 time=98.926 ms ^C --- 216.218.186.2 ping statistics --- 6 packets transmitted, 6 packets received, 0% packet loss round-trip min/avg/max/stddev = 97.285/99.688/105.039/2.481 ms Similar results on two of my hosted boxes, which may or may not be in crappy hosting environments.
- sabat 18y agoThank god, a posting that isn't by DHH (wink) or about some doomsaying economist! THIS kind of thing is why we hang out here. Thx for posting it. Useful!
- DaniFong 18y agoExcellent, this is right up my alley. Thank you.
- brandonkm 18y agoThanks for this. Theres some really useful ones I hadn't heard about before.
- tlrobinson 18y agoCharles (http://www.charlesproxy.com/ http://www.charlesproxy.com/) is also great for HTTP debugging and performance testing. It's got a cool "throttle" feature that lets you simulate slower connections.