• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 1172
  • Last Modified:

Ping and Tracert - How important is latency?

Hi experts,

  I am currently evaluating which hosting company I'm going to use to host two servers. I'm located in Brazil and most of my audience is here as well.

  I'm currently using ping and tracert to assess how is the speedy, number of hops, etc, to some hosting companies. However, I'm not sure how to analyse the results. I'll explain it better:

  I'm thinking of hosting with datacenter in USA, since their price and level of service seem far better than Brazilian companies. However, the times I get when I use tracert are around 170 ms, and 16 hops. If I tracert to the other company I'm evaluating, in Brazil, the tracert times are around 40 ms, with 12 hops.

  My question is: how exactly does this affects my final user experience? How exactly should I read those number that tracert and ping provide?

  Please keep in mind that I'm dealing with critical and realtime content (I'm doing live video streaming), so it's crucial that my audience receives the content as fast and as reliably as possible. I'm just trying to balance cost, service and speed.

  I'd really appreciate your help on this. Since I have urgency in this, I'm assigning 500 points. Thanks in advance.

Helder
0
HelderConde
Asked:
HelderConde
  • 8
  • 3
  • 2
  • +3
3 Solutions
 
humeniukCommented:
Naturally, your ping & tracert times to a server in the USA will be longer than your times to a server in Brazil if you are pinging from a location in Brazil.  So, the question is, where are most of your website users located?  If geographic proximity is important to you, you want your server located near your users rather than near yourself.  You can also do a tracert from another part of the world at http://www.tracert.com/cgi-bin/trace.pl or http://www.traceroute.org.

Generally, a ping time under 200ms is pretty good.  However, that is only one way to measure speed and isn't the best way.

From http://whreviews.com/testing-speed.htm:

"Traceroute is a good way to determine path but, just as ping, it should not be used as the last word when it comes to speed evaluation.

The ultimate test is the download speed test. To do that you should ask the host for a test download file. That's about the best way to judge speed. The goal is to find a host that has the capability of sending at least a few hundred kBytes/second. If you're on a dial-up connection you'll not be able to test this yourself. Read on and you'll find a solution to this particular problem.

Because the host might try to fool you by giving you a test file located on a fast, almost empty server, it might be even better if you'd contact a current customer and ask him/her to post a test file on his account.

The file should be big enough to allow you to see the speed stability over time. A 10-15 Mb file should be enough. Another important aspect is the time of the test. The best times are rush hours actually, when the server are busy. These are are in the morning when most people read their emails (8am on the east coast of the US) and dinner time. Take care to compensate for time differences as not all servers host American websites. "

If this makes sense to you, follow the link for ideas on how to run these tests.
0
 
coreybryantCommented:
Another thing to consider as well - where will your consumers be located?  Since most are all in Brazil - you might consider staying there.  if they were in the UK, you might consider over there.  Afterall, it is not necessarily acceptable for you to get the fastest speeds (although preferable :) )

-Corey
0
 
humeniukCommented:
Something else worth mentioning is that streaming video typically includes buffering, etc., to compensate for expected variations in transfer rates, etc., but not necessarily for a host with a download speed that's just too slow.  Again, for streaming video, I think a download test would be a must.
0
Restore individual SQL databases with ease

Veeam Explorer for Microsoft SQL Server delivers an easy-to-use, wizard-driven interface for restoring your databases from a backup. No expert SQL background required. Web interface provides a complete view of all available SQL databases to simplify the recovery of lost database

 
kohashiCommented:
I would say figure out where your target audience is.  If you are aiming for local traffic, go with a local company since it will be faster for everyone in your local area.  If you are targetting a US market, then go for a US server.  It will be slower for you, but faster for them.  

-kohashi
0
 
frugleCommented:
> I'm located in Brazil and most of my audience is here as well.

Glad to see the "experts" are skilled readers!

ask the US datacentre what other domains it hosts (or use whois.sc to find out) and get a group of friends, relatives or associates to surf around these sites and some of the sites hosted by your brazillian webhosts at various times of the day. This should tell you how responsive the sites are.

If you're hosting large files, as previously suggested, try downloading something large - if not, normal pages will give you a better indication.

If bandwidth is to be a consideration, assume that all brazillian customers on a dialup/dsl connected to the same ISP aren't going to be generating external bandwidth - reducing your costs somewhat.

and lastly, for information only, the figures in ms in your pings should be as low as possible. 90-200 for gaming, up to 400 for web surfing, and over 500 - don't bother.  The packet loss should be pretty much <10% - try PING -t www.host.com for about 30 seconds (no more else they could interpret it as an attack) and CTRL+C to end it. If you're getting less than 90% back it's not too good a connection.

Hope this helps,

Mike
0
 
humeniukCommented:
"> I'm located in Brazil and most of my audience is here as well.
 Glad to see the "experts" are skilled readers!"

rhetorical - adj. concerned with effect or style of writing and speaking; "a rhetorical question is one asked solely to produce an effect (especially to make an assertion) rather than to elicit a reply"
0
 
frugleCommented:
paranoid - adj. Exhibiting or characterized by extreme and irrational fear or distrust of others.

My comment wasn't aimed at you humeniuk, it was related directly to the previous comment.

> I would say figure out where your target audience is.

Sorry if it caused you offence ;-)

Mike
0
 
humeniukCommented:
smartass - n. SLANG - humeniuk

No offense taken and no apology necessary.  On the contrary, I am sorry for not including a strategic 'JK' or ' :-)' to more properly communicate the tone of my smartass comment.

Andrew


(ps. is it paranoid if the fear or distrust is extreme and RATIONAL rather than irrational? . . . is there a TA for this question?)
0
 
coolestdesignzCommented:
What is the connection spped of the target audience?
0
 
frugleCommented:
humeniuk,

I would ask the question in the lounge but last time I was in there everyone was looking at me funny and talking about me behind my back!

HelderConde,

Have you any test results yet that you are willing to publish for analysis?

Mike
0
 
humeniukCommented:
Hi Helder,

Now that frugle & I are finished hijacking your thread, do you need any more help?
0
 
HelderCondeAuthor Commented:
Hi experts,

Thanks for your answer and I'm sorry for taking a while to re-visit this thread. Busy days... Still urgent subject, though.

Ok. First let me answer a few questions:

1. 100% of my target audience is here in Brazil.
2. Although most of them will be using at least 128 kbps DSL, some of them will be using dial-up. Obviously, for those on the dial-up there will be no video streaming. I'd say that 90% will be on DSL, and 10% on dial-ups.

I forgot to mention, but I already have a couple of servers on a datacenter in San Francisco. I'm willing to move to another datacenter in Texas, or to one in Brazil. In order to have a better understanding of the numbers, I did the following test. I'd really appreciate if you could read it and give me your opinions.

1. I had a couple of files, at a total of 14.4 MB. Those files are exactly the kind of data that I'm going to load on the servers, when they are up and running in the real world.

2. I uploaded those files via FTP to two different servers: one in San Francisco, and one in Sao Paulo (my area). I compared the results and will show them in a minute.

3. I'm using a pure, unshared 512 kbps line.

Here a the results for San Francisco:
Total: 14.3 MB
Transfer time: 12 minutes 09 seconds
Transfer rate: 157 Kbps (Total MB/Time) -- Remember I'm using a 512 kbps line! Why is this number so low?
average ping time (just as a reference): 211 ms

Here are the results for Sao Paulo (my area)
Total: 14.3 MB (same files)
Transfer time: 4 minutes 4 seconds
Transfer rate: 403 Kbps (Total MB/Time) -- Compare this number! This is much closer to the max 512 Kbps I can have.
average ping time (just as a reference): 20 ms -- This number is significantly lower than the other!

Now, you'll notice that the difference is huge! Have a look at the transfer rates! Have a look at the transfer times. My question to you now is: why does that happen? People say that a 200 ms ping is ok. Even 400 is said to be ok. But have a look at the differences in terms of ping and transfer rate values...

Please help me understand this and make a decision. If prices in Brazil weren't so amazingly high, I'd definitely stay here. But prices in US are way, I mean way lower. That's why I'm trying to find a cost-effective solutions, without having to compromise too much on the speed.

Again, thanks for your answers.

Regards,

Helder
0
 
humeniukCommented:
Ultimately, you have to base your decision on transfer rates rather than on ping, because your users will be downloading, not pinging.  Based on those download tests, you're much better off locating your server in Brazil.  

However, there are a few other to consider, such as do you anticipate having multiple simultaneous connections?, what sort of quality of video are you hoping to broadcast?, ie. pic size, frame rate, compression, etc.  These are relevant questions, but regardless of your answers, I think you will probably want (and need) to do better than the results you got from the SF server.

To that end, I would suggest that you peform download speed tests on a few other hosts in the U.S. (and Brazil or other S.A. countries if possible) before making your decision.  If you look around a little more, you may find a better combination of speed and price.
0
 
humeniukCommented:
After 21 days with no comments this question has been classified as abandoned.  I will make a recommendation to the moderators on its resolution in the next few days.

EXPERTS: Please make a recommendation on how you feel the question should be closed.  If you feel you deserve points here, but don't care to respond, the question might very well get DELETED.

ASKER:  Please return and finalize this question.  Abandoning a question is in violation of the member agreement.  If no response, my recommendation will very likely be NO REFUND.  If you do not know how to close the question, the options are here - http://www.experts-exchange.com/help.jsp#hs5


Humeniuk
EE Cleanup Volunteer
0
 
humeniukCommented:
No comment has been added to this question in more than 21 days, so it is now classified as abandoned.

I will leave the following recommendation for this question in the Cleanup topic area:
Split: humeniuk {11763886} & coreybryant {11764783} & frugle {11799323}

Any objections should be posted here in the next 4 days. After that time, the question will be closed.

Humeniuk
EE Cleanup Volunteer
0
 
HelderCondeAuthor Commented:
Thanks, everyone.

I've split the point among those who helped me the most.

I appreciate your attention and expertise.

Regards,

Helder
0

Featured Post

Independent Software Vendors: We Want Your Opinion

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

  • 8
  • 3
  • 2
  • +3
Tackle projects and never again get stuck behind a technical roadblock.
Join Now