Ping

[boron]

PING time is a time in which a (hosting) server needs to answer on a package of data sent from some computer, like your PC. It is used for checking how fast will a hosting work for you, that is how fast your site will load. PING time depends on server responsiveness and distance between a PC and a hosting server.
All above is a theory, which would mean, if you access a hosting based in Australia from Europe, your PING time will be long. I tested this a little and it's not necessary true...

Whatever, acceptable PING times are below 150ms, good times are around 30ms. So before buy a hosting, check PING (http://dnsstuff.com) << it's in the right column of this multitool site.

[Submerge]

It's always a great idea to buy hosting near your target market. That way the ping will usually be the fastest for them, making them happy. Shouldn't go about making a middle eastern oriented website (for visitors from the middle east) and buy hosting in the U.S. The ping will be a lot slower for them because of the distance.

[boron]

I personaly am not quite sure if distance affects PING so much, or in other words: PING itself doesn't tell enough about hosting server speed.

Instead of PING, Traceroute checking (http://dnsstuff.com) << (in the right column) was recomended to me. But I'm not sure how to read and explain it correctly.

So, what exact information regarding host server speed can a Traceroute give you? Also: Which URL should be put in: Hosting Home page or...?

[Submerge]

Distance would be a big factor in PING, but it does rely on the network tubes first and foremost.

[boron]

Which Hosting URL has to be actually tested for PING - their Home page or their nameservers?

[Reckless]

I think there Home page, because they probably host it on there own host (and that will also be the one your site will be on)

[Submerge]

Well, sometimes hosts have multiple servers. I'd try their name servers and then their homepage.

[Mystic]

It's always a great idea to buy hosting near your target market. That way the ping will usually be the fastest for them, making them happy. Shouldn't go about making a middle eastern oriented website (for visitors from the middle east) and buy hosting in the U.S. The ping will be a lot slower for them because of the distance.
Distance does affect ping speeds. Back in the days of dialup, it used to matter a lot. However, nowadays majority have access to broadband connections. So, shouldnt be any noticable difference in load times.
Hence, pingtime doesnt really matter now. Whats more important is to go for a fast and reliable host - even if its not located in your target country.

[boron]

Whats more important is to go for a fast and reliable host - even if its not located in your target country.

I guess I will end up with hosting reviews on forums and with the stop-watch checking loading speed of the hosting company Home page.
...
Until then: I use dnsstuff.com "DNS lookup" and "DNS timing" tools for checking hosting servers.

[rjp]

A ping shouldn't be the deciding factor honestly. It's more about server speed and bandwidth available. Just because the ping is slightly higher to you doesn't mean the server / hosting isn't faster than someone else.

[boron]

On the end it's all about site loading speed.
I use these two tools:
http://www.websiteoptimization.com/services/analyze/
http://netmehanic.com/toolbox/power_user.htm

which show loading speeds on various connections and point out reasons for slow speed.

[Mystic]

On the end it's all about site loading speed.
I use these two tools:
http://www.websiteoptimization.com/services/analyze/
http://netmehanic.com/toolbox/power_user.htm

which show loading speeds on various connections and point out reasons for slow speed.
Thanks boron for the links. They were usefull.

[boron]

PING again:
I've just found out that I was doing the same mistake all the time - I was asking people around the world to PING hostings from Dnsstuff.com, which actually means we all tested them from the same place, lol.

Now, here's one free cpanel hosting (subdomain only, I think) with servers in Australia: http://zoomcities.com . I invite few people from different places to test PING with Dos command from home PC: ping zoomcities.com.

[Mystic]

Here you go (from India):

Pinging zoomcities.com [69.65.98.157] with 32 bytes of data:

Reply from 69.65.98.157: bytes=32 time=444ms TTL=252
Reply from 69.65.98.157: bytes=32 time=406ms TTL=252
Reply from 69.65.98.157: bytes=32 time=380ms TTL=252
Reply from 69.65.98.157: bytes=32 time=426ms TTL=252

Ping statistics for 69.65.98.157:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 380ms, Maximum = 444ms, Average = 414ms


Frihost (http://www.frihost.com)is a good free host (it now provides Direct Admin instead of cpanel, but DA isnt bad).
Ping times for that


Pinging frihost.com [216.32.68.254] with 32 bytes of data:

Reply from 216.32.68.254: bytes=32 time=344ms TTL=252
Reply from 216.32.68.254: bytes=32 time=338ms TTL=252
Reply from 216.32.68.254: bytes=32 time=346ms TTL=252
Reply from 216.32.68.254: bytes=32 time=337ms TTL=252

Ping statistics for 216.32.68.254:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 337ms, Maximum = 346ms, Average = 341ms

[boron]

Oh, bad results. Recommended PING is under 150ms. How fast does the site actually opens from your side?

[Mystic]

Not slow enough to make it stand out from other websites. Offcourse, the homepage itself is also very light.

[rjp]

I still perplexed as to why you're insisting on using PING statistics to see how good a web host is.

[boron]

I still perplexed as to why you're insisting on using PING statistics to see how good a web host is.

Because one Canadian hosting disadviced me to go with them (!), since, like they said, my site will load slow on my end (Europe) because of high PING. Now I want to test if PING has any useful value or not - it's more a question of principle.

[rjp]

Well the answer is simple: It has no value to you.

Anyone that suggests you use a PING as a solid bandwidth metric is kidding themselves.

[jhnrang]

Well ---according to boron--I did the ping test and here is the result--( I am from India and I have a dial-up connection with 56kbps spped.

http://img69.imageshack.us/img69/8067/pinguw2.jpg

[boron]

Asked a mate from U.K. and he got PING results from zoomcities about 103ms on 1/2 Mb broadband connection.

[jacob]

"Ping is used diagnostically to ensure that a host computer you are trying to reach is actually operating." It's not meant as a speed, let alone, web hosting test. It is meant as "Ping: Hi are you there?", "Pong: Yes I am!" which is partly where the name came from - yes, the game of ping pong.

"traceroute is a TCP/IP utility which allows the user to determine the route packets take to reach a particular host. traceroute works by increasing the "time to live" value of each successive packet sent. The first packet has a time to live (TTL) value of one, the second two, and so on."

OSPF (Open Shortest Path First): "An interior gateway routing protocol developed for IP networks based on the shortest path first or link-state algorithm."

Data is broken down into packets for transport. When those packets are being routed, they don't necessarily travel together, let alone one after the other in sequential order. Packet A might leave Vancouver and go thru Toronto, Ottawa, Seattle, Moscow, and Timbucktwo before it's final destination somewhere in India. Packet B on the other hand, might hit Vancouver, Seattle, India. And guess what? Packet B might not necessarily arrive at the destination host before Packet A.

Why is that? Because it does it's best to send the packet thru the least traffic with the most bandwidth available at that given time using OSPF. You will get different results based on time of day you do this as well! Internet is global which means Internet traffic will be "heavier" given a timezone and general working hours for that zone throughout different times of the day, kinda like rush hour.

you only have control of two factors - your host and your ISP. You can change either but the "hops" in between the two (not to mention between your users and your host) you have no control over. take users and hosting
being within the same country out of this equation... There will be a hop or two to get from a requesting machine to get off the continent and then subsequently "landing" on the other continent and reaching the host (or atleast there will be in North America) (or in technical teams, your LXC to your IXC which in Canada would be going from shaw/telus to bell and off the continent). so, look at those two areas to see if it's lagging or not paying more attention to the hops at your hosting providers end and really ignoring the crap in the middle unless you're actually diagnosing something. The reason ping/traceroute are NOT good ways to determine a webhost is really because they're not meant to be used in this way - they are networking diagnostic tools used to troubleshoot when there's a problem.

to answer the question on distance, the answer is really "it can". To put TCP/IP into a real world context, just think of yourself as being that packet having to travel around the world only there's a LOT more highways/roads/airplanes for travelling on. No matter what, the distance from work to home does not change. how I might travel between the two does. in Vancouver's rush hour traffic, it will take me longer, but at 3am, provided crackers aren't walking in front of my car on hastings, it will take much shorter - which can be a difference of 15mins at 3am or 45mins at 3pm. Knowing I can take alternate routes home to get there faster, (OSPF) I more then likely will. And of course again, there may be something along the way (such as a cracker on hastings) that will hold me up along the way. In college, I was told that it needs to be >10ms. i went to school in canada and didn't think until looking at your results that this could be different around the globe but i think that >10ms between hops on the same continent would be ideal.

If you are looking for a web host, you're better off spending your time and energy looking at uptime reports, how fast does their support get back to you, and how willing they are to help you out when you have problems.

Oh, and always, always, ALWAYS ensure their tech support is 24/7/365!

ears smokin yet? how about this, i can pull up zoomcities.com in a browser, but it times out when i ping it. what's that do to your ping theory ;)

[rjp]

The moral to the story is you don't want a host with a terrible PING time (unless you're living in Antartica--everyone has a terrible PING time to you), but that a PING time is really meaningless when there are plenty of other factors to consider.

:D

[jacob]

An excellent summary ;)

I'd worry more about ensuring you're optimizing graphics, code, and your content aka doing your own part to speed up your website. It's somewhat simple to move hosts - it can be a bitch refactoring and optimizing code.