Slashdot Log In
What Should One Look For in Colocation Services?
Posted by
Cliff
on Mon Jun 19, 2000 02:23 PM
from the finding-your-servers-a-good-home dept.
from the finding-your-servers-a-good-home dept.
willeg asks: "I am charged with the task of recommending a company for co-locating our auction site. The problem is that on paper, they all look about the same. What other information should be used in order to make the decision on a good hosting pick service?" This will probably be a question that many will ask in the future. With the Internet being the 'next big thing', everyone will be looking to stake a claim on their own portion the digital real-estate. Problem is, not many people can afford the bandwidth costs of getting wired for high-speed access, especially for businesses in out-of-the-way areas. Is colocation the easy answer? Or is just as expensive?
This discussion has been archived.
No new comments can be posted.
What Should you Look for in Colocation Services?
|
Log In/Create an Account
| Top
| 174 comments
(Spill at 50!) | Index Only
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
This question is awfully general. (Score:3)
Features you want (Score:5)
You want the following features:
* electronic security (good to find out who was in when)
* multiple 100 meg pipes...
* full BGP4 routing so all those pipes are used at all times, not just when one fails.
* good connections... who are they peering with?
* indivdual locked cages, video security is very good.
* switched ethernet (ok, if you're talking 100 meg colos... but if you find you're on a hub, thats a major sign they are incomentent)
* obscure locations - you don't want a sign saying "very expensive computers inside"
* 24/7 on call support if necessary... what if you need to have that critical machine rebooted at night? sure it'll cost you, but it'll cost you more to have it down.
I work at a colo company, and our fast facility meets these requirements... Actually, I note that Vancouver is one of the best wired cities in North America, we got fat fat pipes to Seattle which is of course set up beyond belief..
Anyways...
Secure, physically and digitally (Score:4)
Don't ask US... (Score:5)
Ask all of your bidders: "What can you give me that the others can't?" You are sure to get a bunch of crap like "a sense of well-being that your blah blah blah". Take all non-nonsensical (would that be "sensical"?) responses and redistribute to the vendors for response. Let them respond again.
At the very least, this will weed out the people who don't know what they're talking about.
--
Compaq dropping MAILWorks?
Co-locating queries. (Score:3)
Ask them "How many hops to the backbone?". Ask them how much bandwidth they have to their nearest peering point. Most importantly, ask them if you can put a test server in one of their racks, FTP and TTCP to it (Test TCP), really drown the link - to see what it's capable of, and then insist on comparing the results to similar results after your contract is processed. (Don't put your Auction Box in there until you've re-tested with an identical FTP set up).
Re:Features you want (Score:5)
-what kind of heating/cooling the facility has?
-is your equipment going to be off the floor in case of a flood?
-what kind of fire supression equipment do they have?
-are the NOC monkeys knowledgable?
-do they monitor your equipment for you?
stay away from level3 (Score:3)
at level3 the people in the gateway are pretty good on average, but the people up the chain from there are pretty much a waste of o2.
my experience with abovenet in san jose has been pretty good. they are helpful and fast. they are well connected and offer lots of service levels.
Go read Kuro5hin's recent thread (Score:4)
--
Choosing for bandwidth or full service (Score:5)
Once you've decided what you want, find out who else uses a given service. This applies both if you are looking for bandwith only or are looking for full service. You want access to be fast. You want it to be reliable. Do traceroutes from different locations to determine response times to the other clients of the service you are looking at. You can figure out a lot about their peering arangements and see if there are problems with a given site.
Ask them directly about their peering arangements. Find out about failover strategies. Ask them about service guarantees and make sure that partial refunds on service are available if those marks aren't met. This is key. If they don't meet their obligations, they don't get paid.
If you are also looking for 24/7 admin, find out how familiar they are with the apps you'll be using. Find out how they monitor the apps in question. Do they have people on site 24/7 or is most of the work done remotely? Ask them about their backup strategy and how they go about recovery of a fully lost system. Ask them to recommend a configuration for your site. You don't have to use it, but it should give you a good idea of how well they understand the networking issues.
Especially if you are doing the admin yourself, you need to know where the physical facility is and what kind of access you can have to it. If you have a server at an ok prompt, you'll need to get to it physically (unless you've set up a portmaster with remote access).
That should help you get down to a few possibilities.
Its all about service (Score:5)
Where things are different is the type and level of service you'll get. Some co-lo's are just that- straight co-location. Others provide more managed services. Some claim to have really good managed services, but really really suck at it. Others specialize at managed services and suck at plain co-lo.
Another thing to worry about is the fish/pond issue. Are you a big or small fish? What sized pond do you want to swim in? Small fishes in big ponds tend to get the shaft, but their co-lo's tend to have more negotiating power with other carriers. A small fish in a small pond will get better service, but their co-lo may not have the power to get better connectivity when they need it.
Your most important thing is references from other companies with the same needs as your own. Talk to them at length and get the real scoop. When something broke, how quickly did it get fixed? Did it happen more than once? If they were down for a period of time, did the co-lo's SLA cover their lost revenues/good-will with their customers? How well were they kept informed with regards to the situation and scheduled maintenance.
My .02:
Exodus- good for straight co-lo. Lowsy managed services. Has a great or horrible backbone depending on who you ask.
NaviSite- good for high-end managed services. Lowsy at straight co-lo. Very different network design which may be a good or bad thing depending on your point of view.
AboveNet- Seems to be a nice pond to swim in, but they seem to have more than their fair share of BGP issues than most.
GlobalCenter/Frontier- Seems to have a lot of peering problems. No idea on their level of service.
sorry (Score:3)
Connectivity (Score:3)
Granted, 100% uptime is an impossible goal. But, assuming that you can keep your server up and running, you are dependent on your colocation service for keeping your server connected to the net. The main factor here is connectivity. Make sure they are connected to multiple backbones on different carriers. Look at the network maps of their carriers and make sure the physical location of their colocation service is located near those backbones. Don't get into a situation where you depend on a single trunk line. Try testing their connectivity with a service like netmechanic.com as well. The numbers you get from the test may not be that accurate, but they'll let you know if there's a problem.
Other issues: bulding security, rack security, multiple power feeds and backup generators, dry fire suppression system. Ask them what their disaster plan is. And, of course, talk to some of their existing clients to get some feedback. Also, if you get the feeling they offer "sales support" rather than real support, forget about it. (a good sign of this is that a call to the sales number always gets through but a call to the support number more often gets a machine.)
Re:Next big thing? (Score:3)
Who Invented It?
It is commonly believed that the Internet came into existence as part of the United State's government's nuclear war fighting strategy. Originally conceived as a robust and fast communications network called the ARPAnet (Advanced Research Project Association Network), it was designed to help scientists and technology researchers communicate, in fact, it was created in direct response to the Soviet Unions launch of Sputnik, the first man made satellite. It was part of a whole series of initiatives taken by the US government to enhance science and technology development. Later, the multi-routed and redundant telecommunications lines, switches, and computers were discovered by the military, to be an ideal network to prevent Command Control 'de-capitation,' in the event of a Soviet 'first strike' against the continental US. source [hostfind.com]
Co-Lo's (Score:4)
Ask yourself this question would you have your server housed with 100 Network Operation people running around with their heads cut off, do not know your name, and would not give you the time of day unless you flashed $100 in front of their noses, because the are so big.
Or a small company, were you know they will know your name, and are able to take the time to hear your concerns and not pass them up the food chain to middle management. I personally know every client by name and their box even better. I know each of the boxes personal quirks and their needs as well as the box owner's quirks and needs. Because it is a small company I'm in better position to do the job I love and able to pass that along to the customer.
So my suggestion is to meet the network operation personnel the real people taking care of your investment, not just the sales manager. Ask to take a look at the server room. Make sure it is clean and free from obstruction, environmentally controlled. Do not let some yahoo tell you that you need dual oc-48's to host your site. I can host a site just as well as anybody from my dsl-line. Get a feel on how well they take pride in their job, and you should be fine.
Which one is better?
"Kill -9 needs no justification" BOFH
Different Types, Topologies Of Colo Service (Score:5)
(Disclaimer: I work for AT&T, but not in the hosting group., and this is my personal commentary, not a company statement, in spite of the occasional shameless plug in the content.)
Take a step back... (Score:5)
Did anyone see Jurassic Park? OK, dumb question, everyone saw Jurassic Park.
All these ultra-complicated solutions sound very Jurassic-Park-ish to me. There is, it seems, a turning point where you have such a complex setup that it is guaranteed to fail; and when it fails, it will fail in unpredictable but spectacular ways.
Like, as someone else pointed out, the colo that was so secure that he couldn't get in with a valid ID card. Doesn't it defeat the purpose of security if the people who should be able to get in cannot get in? Isn't that just as much of a danger?
And what's faster -- multiple T3s that are saturated, or single T1s that aren't?
And exactly how long does it take Cisco's biggest router to boot? And how many of them are between you and the rest of the world? And why would you want that?
The biggest difference between ALL co-los, in my opinion, is customer service. THAT is the value added that really makes a difference. Your system is unavailable at 3 AM. DOES ANYBODY CARE? This means so much more than all the geeky stats, hops off the 'bone, etc. If they CARE, they will put you right, no matter what their situation; if they DON'T care, you will be unhappy even if they have all the gear in the world.
So, evaluate your potential providers the old-fashioned way: ask other customers if they are satisfied.
--
What should co-location services promise? (Score:3)
I run a web hosting firm that currently does only virtual domains, but is looking at starting to offer co-location services. We're trying to figure out how much headache/cost is associated with each "level" of service: UPS backups, 24/7 support, multiple net connections, sysadminning, etc.
For those of you who already offer co-location services, what's reasonable to offer your customers who want an "average" (i.e. not ridiculously expensive) co-location experience?
What's good, what's not (Score:4)
1) Exodus (NJ location) looks obscure, but lacks any decent security persence to the extent that I'd like. You just sign in to gain access, and all that they check that on is a contact list.
2) GlobalCenter (NY) uses the same system, but there, you can sign in as "Samuel L. Jackson," "Tommy Lee Jones," etc., and they let you in anyway. They don't even check.
These two also share another aspect that I dislike: Communal rack space. You can have your servers right next to those of another company, and there's no barriers at all. Makes it incredibly easy for gear to get inexplicably damaged, although it's rare that it happens.
GlobalCenter's got a decent amount of private cages available, though, and those are pretty good.
3) Level 3 (NY) has lockdown cabinets in the datacenter, decent security, and biometric identification for datacenter access. If you don't belong, or aren't escorted by someone who *does* belong, you're not getting in very easily.
However, they take forever to actually get your line installed sometimes, which certainly puts a damper on your critical business setups.
4) Globix (NY) has much of the same setup as Level 3. Lockdown cabinets, biometric identification, etc., but they also have beefier physical security, and even if you *do* belong there, you still have to sign a key out if you want access. For some systems, you'd need to authenticate across no less than 4 keycard locks, 3 of which are biometric, two security guards, and a room full of NOC techs. The drawback? Their turnaround time can be dog slow as well, and they have been known to oversell well beyond their frazzled tech staff's capabilities. To make matters worse, the datacenter has windows, and you can see their gear from across the street. That's extremely discomforting.
As for key questions, I'll bypass most of the redundancy, and just recommend a tour, possibly inviting your own network engineer or a consultant to help out if you don't know all the BS from the actual details. Key things to ask about are:
- Raised floors / Drop ceilings? (Key for them wiring you quickly and flood avoidance)
- Conditioned power/Backup batteries/TESTED generators?
- Glycol or Halon fire suppression? (Glycol is less likely to kill you if you're in the room.) To that end, ask if there's a way to halt the fire suppression if you're still in the room. I can't stress the importance of being able to get your people out before the FM-200 kicks in.
- Security and surveillance? You need ID checks, sign-ins, the whole nine yards, since nothing is more critical then your information and the gear it runs on.
- Network redundancy and quality. I'd focus on the redundancy more than the quality myself, since most providers do have pretty much equal bandwidth. However, if they only have one way out of the datacenter, that's going to hurt you at some point.
Those are the key points that I'd worry about. NOC competence isn't something you can always rely on with any company (I should know, I *was* one
Raptor
Re:Features you want (Score:3)
Monitoring is always useful, and also check to see what kind of access you will have (better be 24x7).
People have already stated about power and cooling, but having been on the wrong end of this before - make sure they are redundant!!!! One generator dying and taking down a datacenter is unexcusable.
One thing I didn't see mentioned, is when are the maintenance windows for the infrastructure that you are sitting on. If they have to do a router firmware upgrade - they better do it in the middle of the night.
OK, thats my $.02
Answered my own question... (Score:3)
OC3 = 155.52 Mbps line
OC48 = 2.48832 Gbps line
Some stuff learned along the way (Re: Features) (Score:3)
I once had a friend at a start-up ask me whether they should colo or do it themselves. I have an edited reply here: http://www.nspf.net/colo.txt [nspf.net]. In it, I talk about several things that they or a colo provider would have to think about and plan for if they were building and managing a data center. As you look at a data center, you might want to think of some of these issues as well.
One of the older reasons that people started using colocation was that local loop charges from an office into the ISPs was expensive. By putting a server at an ISP, you avoided those charges. The data centers at the colo were better than one's typical office. Once you get past DS3 speeds, the LEC charges don't look as expensive as they used to. Smaller sites (<=10Mbps) are still better served at a colo facility.
Recently, the reasons for picking a colo provider have morphed into: "What am I willing to outsource?" When you pick a colo site, you're trusting your machines, networking, facilities management and physical security to another company. Are they better at doing it than you?
Some colo providers distinguish themselves in various ways. One might be better-connected into AOL or broadband networks. Some focus on connectivity. Some might be better aquainted with NT than others. Some might have a great daytime Sun or Linux staff. One might offer database outsourcing/management. One might offer backups. One might rent you EMC/SRDF disk space. One might offer managed servers instead of using your servers. These features could good reasons to select one over another, but only if they complement your operation. Make sure, though, that you don't get attached too much to any colo provider's value-adds. For example, if they have network problems every week and your data is trapped in their managed database, you'll have to live with it until you can duplicate the functionality that you'll have to leave behind.
Initially, in an R&D and trial/rollout phase, many companies can get away with outsourcing as much as they can to focus on their site's development, but as the site becomes more popular and as users depend on it more, the uptime and reliability of the site becomes much more important. You can't just let your colo or network provider screw up anymore. It's unacceptable. Decrease your dependence on any colo provider as your site becomes more important.
A very annoying feature of running a colocation site is working remotely. If you need to do anything more than hit a power button on a machine, you need to figure out a way to not have to do that. Don't change tapes, buy an autoloader with lots of tape capacity. Don't use machines that need keyboards/monitors - use serial console servers to access the serial consoles of your devices or use something like Citrix to manage your NT boxes. Buy reliable machines that don't crash as often. Inside a server, a 10000 RPM or 15000 RPM disk tends to fail more than a 7200 RPM disk. A beefier power supply running at half capacity runs longer than a cheaper one running at capacity. Buy more servers than you need for everything so that you can migrate your service from failed servers to standby ones. Don't run the latest version of an operating system. Run the most stable/patched version of the OS. Eliminate all single points of failure from the networking side (including having more than one upstream ISP if possible). Routers and layer4 gear need to reboot sometimes. Always buy more than one of each.
The best way to avoid failure is to have multiple data centers. You can care less about the reliability/availability of one data center because you can always direct traffic to another one. Many web sites make a mistake early on of building a single dependency into their site, whether it's a database or a filesystem, something keeps the site from running in parallel with a similarly-configured. Plan from the start by running your web site from 3 locations, and you'll be able to scale your site very well. You'll also be able to pick and choose from cheaper colo providers that don't do N+1 redundancy to help reduce your costs.
At some point, every colo provider will let you down. It's inevitable that something bad will happen. Picking a colo provider that learns from its mistakes can be better than one that strives to make none. We would expect the same of our employees if we ran a data center ourselves.
If you're good at managing servers remotely, the location of your site(s) becomes less important:
Some thoughts on networking for colo providers:
Random musings:
Just stuff to think about.
--
Eric Ziegast
(PS: I used to work at ISPs and colo providers just like synx. I currently help run a very popular web site at several different colo facilities.)