Charles as I am not in the biz of selling cheap 4.00 bandwidth I am however in the biz of knowing how to setup, run, program and administer High Quality Hosting and have earned the reputation of such. OK from the small amount of information you have posted here I can say (with confidence) your setup is all wrong. Take a good look at what my buddy Phill21 wrote and Listen to him. Phill is one of the best admins around and knows what he is doing. From what he wrote here he is on the right path. I suggest contacting him and keep an open mind to what he says...

Good Luck

We add movies since a couple of days and just added another 750gb as well as an extra server for the load balance.
250gb wont last longer then a week on my end.
Hit me on icq and ill shoot you some details :winkwink:

I just wanted to say thank you, to all that contributed to this tread. I have found it very informative, and I am a bit humbled hearing about the T packages too :)

I will of course be in contact with many of you soon, and do look forward to taking with you more.

Thax
SEan

Hey guys - if you can shoot me an email at [email protected] - i would be happy to get each of you an individual quote.

Charles - also, we would certainly be happy to do a call with our Sales Engineers and go over our recommendation for what you're trying to do. Bandwidth is cheap - it's a matter of making sure you have the right setup from the beginning. Also - we can talk about how CDN might fit into the equation as well - to offload alot of work from the server to the CDN.

We have a few specials running this month if anyone is interested in discussing them.

A sample would be,

Quad Core Intel Xeon
4GB Ram
500Gb Sata
100Mbit Bw
$599/Month

If you need assistance with your setup to spread/decrease your loads our techs can assist and show you pointers to do this. Once you are big enough you will of course want to invest in a more robust drive setup.

www.ReliableServers.Com
icq: 3835797 aim: DavidHindman
david at reliableservers.com

Arent you a reseller of Choopa?

I plan on writing Phil just after I'm done with this thread.
The setup I've been on has been tweaked for nearly a year and I'd love some input from a tech as I've had 2 people thoroughly go through it.
WG

WG, If you have the source code for Java look at the option of getting a static compiler to compile the Java code or you can allways go and get http://www.azulsystems.com/ , you should see few of them on ebay as Lehman Brothers was using them a lot

Do you know any static compilers and more importantly, do you know any with a track record of improving the CPU management/performance?
WG

http://schmidt.devlib.org/java/compilers.html

Interesting to see that TowerJ which i used in the past has gone out of biz. May be their is no market.

wow great thread

What's the overage price per mbit if someone were to take the gige port with 100mb included?

I'd also like to say.....so much for Brad Mitchell's assertion that the days of $6/mbit and lower bandwidth are over now that Alphared is gone. :glugglug

$400 per 100mbit is the overage rate. It's basically 100mbit increments for this special.

Re: the Brad comment.

He's actually more right than you realize. He is talking about something that goes a bit deeper than what appears on the surface, and those in the business know precisely what he is getting at.

Will $6/meg (or less) flat-rate pricing become viable in the future? My take is probably. For this to happen, bandwidth use must increase to offset the per-unit cost decrease. My crystal ball is malfunctioning today, but I do not currently see a stop to the slow decline of bandwidth prices. I do believe there is a floor on the wholesale rates, but I'm not going to peg it at a number. As long as revenues can continue to increase overall (if I can go to Level3 and say "hey, we're paying $100,000 for 10,000mbit right now.. we'd like to increase our spend with you to $200,000 for 40,000mbit" that's in general viewed as a win) you will continue to see the slow downward trend in bandwidth pricing, just as you have over the last 10 years. At some point though, usage is going to start to flatten out, and you will then see a stop to wholesale IP price erosion as there will be no more "winning additional revenue" on the table. I think we're "close" to that floor, but hey - I've been proved wrong before!

Basically, someone coming to me asking to pay $6/mbit @ 95% and 15 servers simply does not interest me. Some hosts may take that business, but it would not be profitable for me to do so. Carefully selected specials can certainly get the per-mbit rate down for select customers, and those customers should absolutely take advantage of that fact. I've probably talked to 25 or so folks since Friday night, and I've actually only offered competitive quotes to perhaps 4 of them. The rest, we simply were not willing to compete on pricing they were seeing from other providers.

For us, the specials listed are carefully crafted promotional specials. They cultivate good word of mouth, as I'd put our support up against anyone out there. Once someone is in the door with us, we generally have them as a client for over 3.5 years (with a few exceptions, where I fully admit we have dropped the ball). Getting people in the door is hard - the sheer number of hosts posting just on GFY should give everyone an idea of that fact. We also, could not survive on *actually* selling $4/mbit like Brad is getting at. For some this is a great deal, for others we won't line up well with their goals.

Brad is a marketing genius in my book, I wish I had 1/10th of his PR and sales skills. While he may be a bit blunt in his comments, and perhaps to an outsider not completely explain what he's really getting at, he is absolutely correct. As I've said many times to many potential clients - folks like him, and a few other hosts here, I absolutely do not mind competing against since I know I'm going up against honest bids. It's the hosts Brad is getting at that are a cancer on this industry offering things they simply cannot sustain. To put blunt words on it - many hosts out there are lying to you about what they provide.

So, in my limited opinion - jump on the specials you are seeing offered soon. Brad isn't completely off his rocker :)

Back to football for me! Go Vikes!

One of the things we are missing is the improvments in optics and DWDM technology which allows you to push more and more BW on a single strand of fiber. The capx cost of pushing 40gb is same as what it cost 10gb 3 years ago and in 3 years we will see 160gb on the same cost. http://www.lightreading.com/ has soem very good reading on this topic.

Ok so basically 101mbps would incur a $400 overage charge?
That's still not bad really, even at that it's $8 per meg with 99 more to go without spending anymore....I just want to make sure I understood you correctly.

This is the first special like this I've seen where a gige connection is available, which means you could use the whole 100mbps. With a 100mbps capped line you can really only use about 70 before performance becomes an issue (based on personal experience)....so pricing on specials like these is, as you've said before, something of a marketing gimmick, it's not usually really $4 per mbit.

This is something I've thought about as well, although I'm not well versed in the specifics. With level3 laying down 10gbps lines, they basically decupled their network capacity for a very small capital outlay (relatively speaking), so it was inevitable that prices were going to drop significantly.

I've always considered myself someone who's always got along with numbers and math, but for some reason I cannot fully comprehend what "95th percentile" actually calculates too. I've even looked it up at Wikipedia but its just still unclear.

Can someone please explain the terminology to me? In the context of how web hosts use it.

Hey everyone,

Sorry, if I hadn't had such a busy day, I would have posted in this thread earlier. Since the ever popular 100mbps server is what all the other companies are offering up, I'll put mine in the mix, for consideration.

Quadcore Xeon e5405 (2.0ghz / 12mb cache / 1333mhz fsb)
8GB DDR2 FB-DIMM (667mhz)
2x500GB 7200rpm SATA II hdd's (Raid-1)
1000mbps NIC
100mbps Dedicated Bandwidth (capped or uncapped)
Transit providers: Nlayer, DTAG, Level3 (NO Cogent garbage)

Full Management, which includes: OS installation (custom kernel modules OK, CentOS w/PAE module is standard), software installation (again, custom apache/lighttpd modules OK), as well as server optimization and hardening

$699.00/month

Feel free to hit me up on ICQ or email at any time, and I will respond promptly. That is, unless you've caught me during one my few hours of daily sleep we keyboard commandos get these days.

I hope you all are well, and I can help you with your next server ASAP. :thumbsup

what do you find so great about it?

The cost of putting fiber is fixed and now days paid thanks to all the chapter 11 we had after the dot com crash. So every time say L3 want add more usage all they have to do is lit more fiber or replace the optics (start location, end location and every 40km or 160km telco huts) and they go from 10gb to 40gb to 160gb, you get the idea.

Jay

how much is over charge if I go over 100 meg (I see that the port speed is 1g)

Since this thread is of great interest to me (and is a great thread in and of itself...a lot of good input from some very knowledgeable people, esp. k0nr4d), here is my $.02


I feel compelled to point out the inaccuracy [imho] of this [alleged] claim. Brad, I respect both your tenure in the industry, as it is greater than mine, and Mojo's excellent reputation. However, I disagree with this [alleged] statement on the grounds that I believe it is more of an appeal to your favored price point for a sale, than the actual truth. I don't think that needing $4-6 bandwidth is a sign of being in dire straits at all. Different sites monetize at different rates, and some cut it pretty slim. When you're pushing tens of gigabits, and bandwidth cost comprises >%90 of your total operational expense, lowering that bandwidth cost is directly proportional to money-in-pocket.

Thank you, DH. You rock!

I agree %100.

Here too. I've personally setup port-channeled servers that pushed 2gbps.

Feel free to ICQ me to see how my offering compares, too. :)

95th percentile is basically how your providers are billed, almost exclusively. The other option for providers is if course a full link (e.g. 10 gigabit) for X. Most providers usually will opt for the 95th percentile as it lets you cheaply have additional capacity, without paying for it sitting there unused. For example, I could have 2x10GE uplinks to level3, each pushing 5gbit and pay for the actual usage of 10Gbit/sec vs. having to have 2 10gbit links, one maxed out, and the other essentially unused while paying for the full 20gbit.

This also works for you as well, it's how a provider can provider you a gige or 100mbit line, and bill you for a fraction of it. Your host having the extra capacity for you to use above your committed data rate, does not come free in terms of internal infrastructure and transit/peering links to other providers. Thus, average billing (otherwise known as per-GB billing, the math is identical) incurs substantial risk for a host - what happens when a user maxes out their gige for a single hour during each day, but has zero usage otherwise? Via average billing they would be billed for nearly nothing, but you still had to have a full gigabit of capacity for them - obviously taking a rather substantial loss. Again, a numbers game :)

That explains *why* 95th percentile is used. Hopefully I can explain the math behind it concisely. My favorite way to try to describe it's intent to folks, is it is "average peak utilization" of a given link. The number was found to largely capture the actual rate used on a day to day basis, during a given customer's peak times - while allowing for extraneous bursting to not be billed (so if you hit a full gigabit for 4 hours one day, and you otherwise are at around 200mbit during your daily peaks, you will be billed for that 200mbit, not the full gige).

The math works like so. Imagine you have 30 days in a month. 10% of this figure is 3 days, so we have some nice round numbers to work with. Lets say I take an average usage rate for each day.

So, I have 30 "samples" of your average daily usage. I then look at this data, and throw out the 3 highest days of usage. The next highest sample (day) is what determines your billing rate. This lets you have 3 days of "free" bursting, and you pay for the next highest daily average after those 3.

95th percentile for billing works exactly the same way. Simply swap out the 30 samples (days) with 3600 samples (5 minute averages), and the top 10% of those with 5%. In a provider billing case, we are throwing away the top 5% of those 5 minute averages, and then billing you on the next highest 5 minute average usage sample. This equates to roughly 1.5 days of "free" usage. So, if you get slashdotted one day out of the month, you will not be billed for your quadruple usage. If you get slashdotted for 5 days of the month, you will.

It sounds somewhat complicated at first, but once you become familiar with it, it's not so bad. In fact, it's pretty amazing how accurate it really is at getting to the "average daily peak usage" number I mentioned at first.

Hope that helps :)

SnakeDoctor - yep, you understand it fully. Let me know if you have further questions.

Peace,

-Phil

$8/mbps

If this is a worry, we have no problem implementing a cap, at any level. For example, if you want to commit to 100mbps, want to be able to realize growth, but don't want to be liable for a 600bmps 12-hour spike, we could just commit you to 100mps, and cap you at 150mbps.

Just some food for thought. :2 cents:

Phil, my hats off to you. My fingers hurt just thinking about typing all that. :P

I emailed Andy over the weekend and got a pricing quote at around $14/megabit, definitely not as competitive as I had hoped.
WG

Sean, please contact us on ICQ or via email. I'm sure we can work something out in your favor.

thanks for the explanation Phil. I think I'm a bit closer to understanding 95th Percentile now. Your right it will be a case of actually getting the bill and working with it more to feel at home with it.

Phil seems to offer a good price ($6/meg).

i have unmanaged single xeon quad core 4gb ram, 2x500gb sata 7200rpm and 100mbps unmetered at $500 per month server. if you can get me managed server for the same or better price, please contact me on icq.

WG,

I'm certain there was a misunderstanding at some point. Contact Andy again, or myself with your needs, and we'll develop a solution in the $6-7 range, most likely.

I suggest Choopa - they have excellent 24/7 service and their billing is a dream!

If you want very cheap but non-managed, check out LeaseWeb.

Good luck with that

From Andy's email:

>> 3x
>> Dual Quad Core
>> 4 GB RAM
>> 2x 120 GB SATA
>> Raid 1
>>
>> 100 Mbps Dedicated, Tier 1 BW
>> $1399/ Mo

Unless I misunderstood that to be 3 * 100 megabit packages, but from the looks of it, it's 100 megabit total.

WG

Phil's been extremely helpful via email too to improve the setup I've been using. Just waiting back to hear on a tech of his but he's been very gratious. Thanks Phil.
WG

WG,

This is THREE SERVERS, with an aggregate bandwidth commitment of 100mbps.

:)

Wow! im still paying $825 for one of my managed servers with 50mbps, and that's Cogent.

what's the etiquette for contacting my host and telling them if they cant lower my monthy hosting cost, I will move? lol

Pretty much exactly that :)

"hey, I signed up X years ago for Y price. Since then it looks like the market changed quite a bit, I can get offers A B and C from these competing comparable hosts. I'd love to stick with you guys since you've provided excellent service, but since you are 40% higher priced I do need to consider by bottom line. Is there any way you can match this pricing I'm getting offered elsewhere?"

Your host should actually very much appreciate such an e-mail. It always sucks to lose a customer over something like a 20% price difference. I do appreciate it when I get a chance to save a long-term relationship.

I'd say about 50% of the time, I can straight up match or beat the price asked for, the other 25% of the time I can make a counter-offer that is accepted. The remaining 25% we do lose and the customer ends up leaving - but usually to a non-comparable host.

We also (and a lot of other good hosts here) try to be somewhat proactive about contacting customers and offering price discounts when appropriate. This is a bit harder (for me, at least) than it sounds due to the ridiculous number of customized plans have - it does take some considerable time.

-Phil