What is the most improtant factor in chossing your dedicated server - Price - Performance of the data center and network - Service on the server - or reliability of the company (solid - world class data center etc.)

I see some choosing price and going with providers that have servers sitting on the floor and no backup systems - and i see others only going for top notch service - but what do most people buy on??

People generally cut their teeth on something bad, get screwed, then go with a company like yours if they're still serious. :)

No, really. Sad, but, true. It's like starting with a Packard Bell before moving up to a Dell or homebuilt (or heck, even an e-machine -- those ain't too bad compared to most of the 'em ;).

Also, people don't realize the true costs... so they don't compare them. Nor do they recognize everything that goes into a server. There's needs to be some education here before people will realize the value of services like yours. Whatever people don't realize, show. Show them.

For instance... you can add up all the extras for a decent machine... an XP1700 w/1 gig ram, extra hard drive, etc., and show that, all added up, your startup fees are comparable. (for instance, on that kind of machine you're about $100-$150 more than the absolute lowest cost and often worst providers on here)I think too many people see the $99 or whatever for a machine that really doesn't have enough memory or backup facilities... and that's it. That's all they see. Make them see more. Show 'em. Graphs, charts, tables.

You need to educate people what things like BGP4 are for. What route tuning is. How your traffic takes the best route, the shortest route, the fastest routes... so data arrives faster and more reliabily... whereas many other low-cost solutions are solely on a sole a sole provider (usually Cogent). 1 connection means no connection when that connection goes down. Use analogies to roads. If there's only way to get there, and that road is blocked, your f*cked in a bad way, plain and simple (unless you're Go Go Gadget and your car can convert to a helicopter! but that's just a fantasy, of course; the same kind of unreal fantasies people have about most of the bottom-rock server deals, curiously enough!).

Tell them that's why we have multiple roads for cars... so when one route is down, we can take another. And when one is congested, we can take a less congested route. And when one road is faster, we take that. And sometimes the faster roads have tolls (Florida is a good example of this! When in Orlando, for instance, all the locals told me that they LOVE to pay the $1-$2 of tolls a day... so they can live their lives ON TIME... while the tourists, not really thinking the situation through, mostly try to AVOID the toll roads, and thus, waste the precious and limited time in this life. stupid, stupid, stupid. mr magoo short sighted, short circuited thinking)

Not even RackShack does this. Takes the best routes available to them, that is. They're a good provider, but they go cheapest first ("low cost routing"), and that's just not as good as what you and other higher end providers give. It's like those tourists wasting their vacations and lives on the slowest, most congested roads in Orlando, trying to avoid the buck or two that will get them to the joy in life a little faster. They just don't know any better (until they're EDUCATED by someone who cares enough to educate them, that is... and that's what you can do).

And on the worldwide internet, sometimes they don't remembe that a connection that's fast for them might be not worth the time to visit from people internationally. Analogies to airports can help here. There are nonstop routes and there are flight segments with lots on unpleasant, time wasting changovers. And while you might be able to stand a few changeovers for a cheaper faire in the U.S., you might be at your wits end doing this internationally. Getting the least stopovers you can without raising the price too much is the most reasonable route, in both air and net travel.

Pictures can help a lot to compare quite a few things. Show your data center in bright, real looking pictures (not weird polished looking crap that nobody will believe are real, or really from your data center). Take them yourself. The pictures should be honest looking, and perhaps should be even what you could call amateurish; the data center should not be. (amateurish, that is!)

You should point out that if the data center looks like something you could have in your home, that it's not going to be reliable. Homes are setup for very little power comparatively. And the AC in homes has to fight only ambient temperatures, not hundreds or thousands of little heaters in the form of computers. Point out that if they think their AC has a hard time cooling their home on a particularly hot day, then imagine what it'd be like if they set up hundreds of space heaters, turned on their AC, and tried to get the AC to fight THAT kind of heat.

Let them know that if a toaster can blow a fuse in a home, then imagine what 200 250-450 watt servers can do. Use metaphors and show equivolencies. For instance, 200 servers take up as much energy as 100-200 older style microwaves, or about 50 of the really modern 1100 watt ones.

Mention how often one computer these days has heat dissipation problems. For hundreds or thousands of computers in one area, there'll be a meltdown. Show a photo of a standard home or office air conditioner vs. a photo of one of your multitown unit. Try to get the scale to be the same in the photos. These are significant differences. Let them know that servers in an area not built to host them is more like having a funny little candle in the SHAPE of a computer. heh! true, though.

Most people aren't good investors. They usually go through a process of getting their feet wet, and if that wetness turns out to be napalm, then and only then, do they start learning and trying different things. To bypass this tendency in people, you really need to educate.

You need to get them worried about things they should be worried about (this is not just ethical -- it's an honest and valuable service to them), and get them happy about and attracted to things they should be happy about and attracted to. You gotta build a propulsion system... you need to push them AWAY from bad things (the specific negatives offered by the danger zone providers), and get them moving TOWARDS the specific, good things offered by sane, reasonable people at better facilities.

Let them know that all the good things you have, all the good things you do, are done for a reason. Be generous with your knowedge, and your thinking. Share those reasons liberally. Increase their knowledge about these things, talk about them, and you'll increase their want from them. That's your TOWARDS.

But, remember, your first gotta build your AWAY... AWAY from the pain and potential pain of having the power go out... of having your bandwidth cut off... of having your data lost (a very real possibility on even RS' ext2 filesystem servers)... of having your server on the floor in an area that looks like a child's messy room...

It's just basic marketing stuff. I say basic, but I guess some of it's not. Heh (a lot of marketing that's taught is mindless, timewasting BS). It is reality, though. And reality is very rewarding when you work in and within it. Reality works. I can give you some more specific tips via pm if you'd like. I really like high value companies like yours that offer some of the good stuff at near the cost of some of the not-so-good or bad stuff, and would like to see folks like you succeed more.

My heart cringes with all of the responses I see to some of the worst quality, but lowest price offers here. Some of these folks really don't know the pain they're in for. They're good people, too. That's why my it hurts me to see the criteria that they're evaluating servers on. And if they intend to resell any space on those servers, then that pains gonna trickle down to yet more innocent people, too.

A lot of the folks here gotta realize that if they're sometimes willing to go to Outback steakhouse, which is reasonably priced and tends to overdeliver for that price, instead of Burger, which is cheaper, yes, but by comparison, sh*te (or at least a completely and totally different class), they really need to know what here is the Burger King and what is the Outback. Know what I mean? In fact, drawing analogies like that can really help. It puts things into terms people know.

In fact, if you think about it, it's not that higher quality providers like that are competing for the low-cost market are really charging more than the lower quality providers. It's more like those lower quality providers are overcharing. If you went to Burger and paid almost as much as you pay at the Outback or another high value restaurant, you'd be getting ripped off and you should be pissed off. From this perspective, some of the deals I've seen recently should be $40-$50 bucks a month, 'cuz when you add up the lack of real facilities, connections, and crew, they're giving you at least 50-75% less than what they're charging you for.

Also, you can draw up a table of all major features, prices, options, connections, facilitiy details, etc., and, just for the folks here, put up a link to that table (I'm saying just for WHT folks, cuz I don't think you'd want to alert prospects found elsewhere to competition they'd actually be better off not even knowing about in the first place -- although a RackShack & RackSpace comparison might be fair for casual browsers... ). They can then see that you offer pro level services, high end Toyota or Lexus type quality and speed of service, for about the price of the Pintos. It's a Hyundai like move, really. Bringing fully outfitted cars to the market for $14-$17k, that would be twice that price from any other brand... and within a full 5 year/10 year warranty, with free roadside assistance, to boot (which is roughly equivolent to having a 100% SLA).

Well, best of luck. I understand what this poll is about... I've been wondering the same things, too. And again, if you need any free help or tips, feel free to pm me. I'm more than glad to help people out when I can. Makes me feel good. :)

Wow, thats almost the longest post I've ever seen.

Anyways, I look for a combination of some of the things you mentioned.

Price - Pretty important. I'm willing to pay for what I require, but I'm going to compare services and find the lowest price for what I need.

Performance of the data center and network - This is everything to me. A dedicated server is worthless unless it's connected to a fast, reliable network.

Service on the server - Un-important. I just need an unmanaged box. Prompt hardware fixes would be nice, but not crucial since I maintain a hot failover server, so even if it took a day, I don't mind.

Reliability of the company (solid - world class data center etc.) - Unimportant to me. If it's a fast, reliable network, I'm set. You can't put your business in the hands of one company's survival anyways, so I'm ready to move at anytime. Plus, the hot failover servers are a nice insurance policy.

Holly smacks batman! :eek2:
That's the longest thread i have ever seen!

:eek:

jmars...wow!

Here's how i would rank it for me.

1. Performance of the data center and network
-You are pretty much dead in the water if your network does not function along side the competition.

2. Service on the server
- If the server goes down....who's going to fix it. What if it requires hands-on twicks? I sure would hope someone is there to get the job done. Or else...you are ..well..screwed! for lack of a better term.

3. Price
- If the price is a bit higher than usual, i will still signup. Ever heard you get what you pay for?

4. Reliability of the company
- Hey...as long as they keep the network up and running, then it's all good. :cool:

Originally posted by pcsteve
Holly smacks batman! :eek2:
That's the longest thread i have ever seen!

:eek:

jmars...wow!

Now if I could only get the chicks sayin' stuff like this. :pimp:

Or, well, close, y'know... like umm... for starts, not calling it a THREAD would be good. :look:

Ahem.

But, seriously, folks...

I felt that stuff had to be said by someone. There's a bunch of super low price deals out there right now that have the potential of burning people, and those people's own clients, very badly. Just like you wouldn't drive a car without insurance, you shouldn't depend too much on some of the ultra-cheap deals out there. I'd recommend testing them only, as a load balancing server you can lose, or as a mirror .. you know, something you can quickly pull... but never as a first choice, especially for an e-commerce or resale box. Not until they're more proven, and the weaker providers die off or get stronger.

I'm very glad there's guys like sailor and a couple of others emerging who are offering truly usable, if not awesome, mid to pro level server deals for *near* the rock-bottom costs of some of the unproven deals out there. For that, I'm willing to write the longest posts WHT has ever seen. HAHA! :nuts: Erm... Uh. Nevermind. I get too enthused about this stuff, I suppose. heh

Its not as simple as that. From a customers point of view, I start by listing all offers that match my needs. Then, I start excluding offers in terms of features only, without looking at the price. For example, a host with a single connection to the net is excluded here, or a host that doesn't offer dual Athlon MP server (assuming that this is what I need). Now, I have a list of all offers that are usable and I have full confidance that whatever I choose will meet my needs. The equation now is like this:

value = price / (cons - pros)

Originally posted by driverdave
Service on the server - Un-important. I just need an unmanaged box. Prompt hardware fixes would be nice, but not crucial since I maintain a hot failover server, so even if it took a day, I don't mind.

Hot failovers are a hot idea. It's a good way to use some of the inexpensive deds appearing these days. As a very quick aside, may I ask what method(s) you personally are using to accomplish this?

Hot Failovers

Pretty low tech, but it works excellent for us.

rsync nightly to our backup server. Then, we just ping our main server from our backup every minute of the day. If we can't see it, we turn on DNS that points to our backup. We also traceroute to our main server and email the admins with the results. When we can see our main server again, we turn off the DNS that points to the backup and email the admins.

Works great for us. We have gone down in the middle of the day on Monday and not recieved one phone call. It's a pretty low traffic server with quite a few clients on it. Also, we keep our TTLs real low, like 60 seconds, so DNS changes quickly. It increases our DNS hits, but it's nothing we can't handle right now.

I'm sure a higher traffic server would need a better solutuion than this, but we'll cross that bridge when we come to it.

Also, we keep our TTLs real low, like 60 seconds, so DNS changes quicklyJust be careful because some ISP override this and can make the TTL hours.

I always thought 3600 (1 hour) was the generally accepted minimum. In the case of Dynamic DNS, it's fair to break this rule. 60 Seconds works for me!

Drew

jmar excellent post and long too!

i feel like we've met before ? :)

I beg to differ, ive seen longer posts by tim_greer back in the days where post lengths were more restricted ive seen him post 3 posts one after another because it was too long to fit into one or even two posts!!

But on topic, what I read by jmars I would agree with almost all aspects of it. the most important being the choice of the best route and not the cheapest route, you can always see a cheap route in a traceroute as when you pass it you see your ping jump up an additional 50-100ms - choosing a good provider you can potentially avoid that situation.

-XT

On that note, I saw a fun traceroute recently from AT&T Atlanta to Cogent Los Angeles (ASAPCOLO). It screamed cheap route.

Atlanta -> San Fran (AT&T)
San Fran -> AboveNet -> SprintLink
SprintLink SF -> SprintLink Philadelphia
Cogent Phil -> Cogent Miami -> Cogent Houston -> ... Los Angeles

Something like 28 hops and 160+ ms from the OC-3 here.

Brings to mind the song "I'm bad, I'm nationwide, times 3."

Drew

This was not on the dv2 network.