Whats the real diference on the server performance, i dont see any real upgrade from 10 to 100, my server is the same, even with the 10 Mbps nic you dont get to that speed, since its uses only the conection it requires, why and when or for what would we have to upgrade? I mean how do you now you need a bigger nic on your server, my servers only hosts websites and doest not have software downloads or heavy traffic aplications.

If you know you don't need the speed, don't upgrade.

What G said.
Check your graphs in Orbit and if they flatline at 10mbit for long periods of time, then moving to 100 would make sense. But nothing else will change (just top speed).

10mbps (ex. 10 time 1mbps max)
100mbps ( ex. 10 time 10mbps max or 100 time 100mbps max)

Can compare this with a highway with 10 ways or 100 ways !


Check out the bandwidth utilisation !...

The graphs in Orbit should give you some idea of how close you are to maxing out your port. If it pegs out a lot at 10MBit for long periods of time, then you would probably want to upgrade.

Edit: Great...two other people posted in the time it took me to write that post.

Thanks guys thats a great idea, i really upgraded just for having a 100 Nic, i never really checked the Orbit stats before, but thats a great idea, i will see my stats and of course i will only use the Nic card for the speed i need.
thanks

Don't wait for flatlining before upgrading. If you're pushing a constant 8m of traffic you should upgrade.

Well, if you are on a metered Premium Network server then the 10mbps should be sufficient for most of your needs unless you expect constant peak at certain time frame. With the free 1200gb bandwidth, you only need a constant 4mbps or so to fully utlize the bandwidth. Anything more would mean that you will have to start paying for excess bandwidth usage at $0.50/gb.

Of course, if the peaks only occurs say during the weekend or for only several hours a day then it might justify upgrading to 100mbps to accommodate those peaks.

If you are hosting standard websites then the 10mbps would probably be sufficient since HTTP is stateless and usually web pages are sent rather quickly and get disconnected from that visitor while it serves others. If you however do a lot of file transfers either via downloads or streaming media then it is possible with sufficient visitors connection to hit the peak easily.

But in all situations, check the RTG charts in Bandwidth Utilization to determine if you need the exta port size.

If you do upgrade to the 100mbps, be sure to ensure that you do not get DoS as you might find your bandwidth usage soaring beyond control.

I have the 100Mbps link and my upload rarely even hits 6Mbps. I signed up with the 100 before I even had access to the server so I was not aware of this.

Anyway, I think I am going to switch to the 10Mbps connection. Can you guys think of any benifits since some/most of you have been with the planet a lot longer?

Advantage, you save $10/mon.
Disadvantage, you save $10/mon.

Unless you're a penny pincher or absolutly have to save some $$$, just stay where you're at. Incase you need the room, you got it.

Oh and props on digging up a year old thread.

Hehe. I did not notice the date; my bad.....

you'll notice it on the other threads you're digging up. Yup I pointed it out there as well. Currently you're up to 4

Might be more now

No, its 4. I counted.

I don't count the one's that are still 2006 - just the one's you replied to 2005 or older

Wow, who cares, they guy is new and trying to make a valid* contribution here!

Don't worry, ahampton, you'll soon learn that nForcer has nothing better to do

*Edit: after having read some of the other threads ahampton kindly revived, perhaps I should remove the word "valid"

Actually, 100Mbs will speed up your server even if you're only using a few hundred GB of bandwidth a month and almost all of it is outbound. In my case I have some inbound data as well, which makes 100 Mbps even more important. On a 100 Mbps link, it takes 1/10th the time to send a given packet of data.

Skip the next paragraph if you don't want to know about my inbound data.

I run a small text only Usenet server (no binaries, music, video, warez, pr0n, audiobooks, etc) for people in the book and crafts industry, so I have a continuous stream of inbound text type data from several peers, but they're mostly faster than me so they refuse my outbound feeds based on message ID without the text of the message having to be sent. Some small peers take a total of a few GB a month. Text Usenet uses maybe 500 Kbps bandwidth compared to the 200+ Mbps bandwidth for a binary Usenet feed. BTW, the app only uses about 1% of the CPU except for when several people are reading the groups and it peaks at about 9%. It's heavily spam filtered and refuses attachments of common music, video, executable, RAR, and a few more types. It also spam filters on some text phrases including some pretty disgusting things that appear in pr0n spam. For those who know what I'm talking about, it does NoCem on spool including the two Hippos.

Also, there's all that inbound junk mail uselessly tying up inbound bandwidth. Spirus is your friend! There's also bursts of outbound backup data. After Katrina, I started sending my CPanel backups from my secondary drive to some storage in the northern US, just to be paranoid. But, before I switched to 1000 mile offsite backups, I've seen DiskSync hit about 14 Mbps bandwidth (changes the scale and makes everything in your RTG graph just a little noise at the bottom).

If you have significant inbound usage as well, because of some app you're using, then the inbound data delays the inbound web server requests ten times as long if you're only on a 10 Mbps link. Going to a 100 Mbps link made my web server much more responsive to web requests even though I only use about 25% of my allotted bandwidth each month and that was before the recent bandwidth increases TP provided. I guess it would be well under 20% now since I've also recently optimized/reduced the inbound data.

Even if you don't have much inbound data, the time compression factor also applies to output. Your web server responses are being queued up and, given the same workload you have now, will wait 10 times as long for the outbound message packets in front of them to be sent so they can get their turn.

-- David