Dedicated hosting, unlike virtual hosting service, where set of various client's sites are placed on one server and are served by the provider, represents the allocated server located in data-center. To server in data-center is provided submission of an electric feed, the data link and a place in a rack.

Certainly, the allocated server gives to the client all resources without restrictions. The dedicated server's owner can establish on it any software, including developed by himself which any hosting service provider will not allow to establish on shared hosting.

Dedicated hosting service share on two categories: unmanaged (without administration) and managed (with administration). Support of unmanaged dedicated servers it is reduced to two actions - reboot the server on trouble ticket and reinstallation of operating system in case of its output out of operation. The reason of reinstallation can be the illiterate reference or breaking server with the subsequent capture superuser and substitution is critical important files.

Managed dedicated support it is rather indistinct and depends on a level of the provider's personnel, given working hours, the guaranteed network availability (SLA), guaranteed time of reaction and complexity of a problem.

The circle of problems support level 1 includes reboot, answers to frequently asked questions, and also procrastination of time by different silly specifications while ticket processes support with higher level. Sometimes in staff is not present such expert and the decision of a problem it can be tightened for a long time. In decent date the centers support level 1 transfers ticket, and itself fairly writes, that could not cope with a problem and has transferred its more qualified expert. It refers to as escalation ticket. Time of transition of escalation can sometimes borrow 24 hours. Take an interest in date the center, what at them the period of escalation.

The circle of problems support level 2 includes the decision of typical problems - installation and adjustment regular ON, updating of the knowledge base for support level 1, small troubleshooting. For example, if cgi-script on Perl demands module Image:: Magick, support level 2 in forces it to understand and put it. However it does not do troubleshooting, and in case of failure does escalation.

Support level 3 is engaged in the decision of global problems - troubleshooting. When the server slowly works or there is a frequent peak loading, any ON does not wish to be established or does not work after installation if there are suspicions on hardware problems - all this is included into the competence support level 3.

Support level 4 there is in staff very large and serious a data-centers or works on a contract hourly basis with fineer companies. Its competence includes everything, including work with unfamiliar to it ON, restoration of system after breaking, search of the account which is engaged spamming, installation of system limits, etc.

The following important characteristic is a quantity of given time. The typical practice data-centers in case of any problem is use of given time by experts of a low level as division on time of each level usually do not give. A typical situation for the unsophisticated owner dedicated server - "overload" time different specifications from support level 1. Therefore tickets should be as much as possible informative, giving everything, probably even, a surplus information. Or on the contrary - extremely simple, but with precise requirements, for example, to establish firewall, having opened 25,110 and 80-s' ports.

Usually, reboot of a server - free, and time for them is not written off. In some cases even access to the special equipment which is capable to execute reboot a server is given. But, nevertheless, not having found combination "Free Reboots", take an interest in date the center - what is the time it is required on reboot and whether it will be considered. On the average reboot borrows 5 - 15 minutes.

Quite probably, that to you will offer other operating mode - on number tickets. In this case any number of inquiries (20-100) which will answer free of charge is given, over a limit - it is necessary to pay. It is necessary to consider also, that it can be ticket for support level 1, and it is necessary to pay extra for the others separately.

SLA - a level of network availability. It means, that sent ticket will be read through even in the day off at midnight, and, accordingly, processed in the specified terms. It is necessary to consider, that despite of assurances and flaunting inscription SLA of 99 %, in practice can leave at all so.

Data-centers inform guaranteed time of reaction with the big reluctance, and it is quite natural. It is impossible to put to each server on the expert. However, if in the contract there is a mark, that reaction on urgent ticket - reboot, for example, 24 hours, it is necessary to become thoughtfull. It means, that the server can stand idle the whole day at the slightest problem.

But what such especial gives dedicated server? First, this all resources, and secondly, is non-standard opportunities. For example, for creation of high-speed access to sites on cursors of type phpnuke can be used 2 web-server, and it is not obligatory on the same dedicated server. The first server - back-end - gives usual Apache, with php, MySQL, etc. The Second - front-end, representing simple cashing web server (http-accelerator) which is capable to sustain loading in some hundreds visitings simultaneously.

On total volume of data links and presence of their reserve. The port by which the server is connected, also plays an important role. Whether data-center cuts off ports or protocols, whether disconnects for excessive network activity and what its threshold? The Opportunity to buy additional IP from another subnet. An opportunity in the further to put it is more than memory, additional HDD. Also it is necessary to pay attention to the list of panels, their price, for the list of possible operational systems.