You’ve probably reached the point where you need to upgrade from your shared host to the real world.
Those byte-crunching beasts known as dedicated servers are the topic at hand. Confusion reigns while trying to make sense of all the choices and offerings.
But Have No Fear!
I’ll do my best to help you find the best possible server by offering some advice.
Now, let’s not dally and get on with the first part.
To begin, let’s define a “dedicated server.”
A dedicated server is a networked computer used exclusively for server tasks (internet). This machine’s resources, including CPU time, are set aside for a particular user or group. You won’t have to worry about sharing your computer with anyone else. You control the system and may install and use any programs you like.
Dedicated servers are typically used to fulfill requests for website content (such as HTML, photos, videos, etc.), but they also have many other potential applications.
How many servers will I need?
For the sake of clarity, we’ll classify dedicated server usage into two categories:
Complete Web Hosting
Hosting a Particular Service (Database, email, HTTP…etc.)
Web hosting for a single site Web hosting for a single website. In this configuration, all services and applications (HTTP software, database software, email software, etc.) are hosted on a single server. These services and software applications are sometimes called “servers.”
Since all of the site’s essential software can be found on a single machine, this configuration is ideal for low- to medium-traffic websites.
The cost savings of this configuration come at the expense of all the software and processes having to share the same access to the machine’s resources.
Hosting of a Particular Service
Use of a physical computer for the exclusive purpose of hosting one program or service. The machine’s hardware, software, and networking are all optimized to provide a single service or function. For example, to ensure that the server’s resources are used exclusively for processing Database Queries, one could install only Database Software on the server. As a result, you can process more requests, which means you can serve more pages.
Websites use the majority of the time, this forms of arrangement with moderate to heavy traffic.
In a typical architecture, one or more servers are dedicated to processing just web (HTTP) requests, another set of servers is responsible for processing only database requests, and another set of servers is responsible for processing only emails. There is no upper bound on the number of servers that can be used. Instead, the servers work in tandem to handle the tens of thousands of requests each second.
Which configuration do you favor?
How many requests and visitors your website will get will determine this. Therefore, to answer this question, we must first understand the services required for user access to a website.
Modern dynamic websites typically rely on two main types of services to provide a fully functional website to the user’s browser.
TELECOMMUNICATIONS SERVERS / INTERNET SCRIPTS ( Apache, IIS, PHP, Perl, java )
SERVER OF DATABASES (MYSQL, MSSQL, ORACLE)
The term “servers” used above refers not to the hardware but to the underlying software. The term “application” is commonly used to refer to this field.
The Web server processes all of the browser’s initial queries and determines how and where information is retrieved from the Database.
The database server processes query sent from the web server. These requests do the heavy lifting of retrieving data from the storage (hard disk) and sending it back to the web server.
The final product (the website) results from the collaboration between the web server and the database server.
A faster CPU and more RAM are always more important than a quicker hard drive on a web server because the server handles most of the logic.
More data retrieval is done by the database server. Hence it requires a faster hard drive and more memory than the typical server.
Regardless, extra RAM is always a good thing:)
The server software will always utilize some of the computer’s processing power, memory, and storage space. There is a point at which competition for few supplies escalates into open warfare.
Separating them onto their dedicated server is one solution to ending resource competition.
The fact that it’s simpler to have a computer running at peak performance on a single task than it is to do so across various tasks is another argument in favor of separating different sorts of services.
I’ll give you an example:
Because of this newfound understanding, we may install more expensive, quicker hard drives in DB computers and leave the cheaper hard drives in WEB machines, where HD performance could be more critical.
However, WEB machines’ processors should be upgraded because they handle much of the logic.
We can achieve peak efficiency without squandering time, energy, or money by approaching it this way.
These are merely recommendations; the actual settings will vary from site to site based on their particular needs.
Components and Features of a Server.
Now that you know what resources each type of server (web, database) needs, what hardware should you order?
There are so many variations on each component design that it would take multiple books to cover them.
Rather than bore you with specifics, I’ll provide some guidelines for each category of a server component that might be helpful when upgrading.
Processor speed is crucial. The faster (in GHz) the processor, the quicker the calculations can be completed, and the pages can be provided to the user.
I suggest reading into which CPU architectures are known to perform well in server environments.
Cores or Processors in the Central Processing Unit:
And here’s another obvious one. The greater your server’s number of cores or processors, the more tasks it can perform in parallel. Since most server activities are in response to requests, and since each request often initiates a new process, it is easy to understand how adding CPU Cores improves the server’s ability to handle several requests simultaneously.
Maximum Random Access Memory: “MORE” seems to be the recurring theme here. The more RAM your server has, the more requests it can handle simultaneously. That’s because a fixed amount of RAM is available for processing requests, and those requests can quickly add up. Of course, numerous variables, such as software and operating system versions, installed modules, and so on, affect this cost.
For instance, the widely used APACHE web server software can consume 7 – 15 MB of RAM per request or process. It doesn’t include any ancillary processes that may already be running or be generated to support that request.
Increases in memory for a dedicated server are among the best upgrades you can make.
Rate of rotational movement of hard drives
Similarly, the faster the disk spins, the sooner you may access the information stored there. Seventeen hundred and two hundred and fifty revolutions per minute (7200RPM), ten thousand (15,000RPM), and fifteen thousand (15,000RPM) are all typical Har Disk speeds.
Applications that make frequent disk reads and writes will benefit most from upgrading to faster hard drives. Database servers, file servers, and video streaming servers are a few examples.
Choosing Between SATA, SCSI, and SAS
These abbreviations refer to the three most popular types of Hard Drive interfaces found in dedicated servers.
SATA is the cheapest option but is typically the slowest and least reliable. However, SCSI offers superior performance compared to SATA, but at a significantly greater price and with much less storage space. Compared to SCSI, SAS provides outstanding performance and increased storage capacity.
Of course, there are many other distinctions between these user interfaces. Thankfully, many publications are available online that go into detail. So please do some investigating; it will be worth your time.
SATA is great unless you’re operating a database server with a very high number of transactions per second. And if you worry only about performance and not price, SAS is your best bet.
Should You Use RAID?
So, unfortunately, we are still using hard drives. When referring to a server’s storage configuration, “RAID” indicates that two or more hard drives (HDs) have been arrayed together to function as a single HD. In most cases, the redundancy offered by Raid improves storage dependability. You’ve likely seen some of the many possible RAID configurations: RAID 0, RAID 1, RAID 5, RAID 10…etc.
The two most typical ones you may see in server settings are RAID 1 and RAID 5
RAID 0 offers some performance and total disk capacity, but no redundancy exists. The data is lost if even one of the drives fails.
You won’t receive more storage capacity with RAID 1, but you gain redundancy through mirroring and, in some situations, improved performance. In addition, a drive failure won’t stop operations; you can swap out the bad one without losing any data. Two hard drives are needed for this configuration.
RAID 5 can enhance storage capacity and ensure data integrity. At least three hard drives are needed for this. Maximum disk space allowed is (Size of smallest drive) * (Maximum number of drives) (Number of Drives – 1). You can keep working even if one of your hard drives fails.
Both software and hardware approaches are often used to implement RAID.
Hardware RAID relies on a specific resource on the card, while software RAID uses the computer’s resources.
The performance gap between hardware and software RAID may be small if your machine is powerful enough.
In most cases, it is recommended to use Hardware RAID if available.
I recommend implementing RAID 1 on your server if possible. Website downtime can be devastating if a server, which typically works uninterrupted, 24 hours a day, seven days a week, has a single point of failure, such as a failed Hard Drive.
Bandwidth
The quantity of data a server can send and receive in a given time is measured in terms of bandwidth. Bandwidth fees can be charged by hosting providers in several different ways.
However, here are a few of the most well-known:
Traffic metering for Internet access
Zero-Bandwidth Usage Limits
Metered bandwidth plans often offer a set monthly data transfer allowance and charge a predetermined sum for additional GB (GigaByte) of data.
With an unmetered plan, you can upload and download as much data as possible to the plan’s maximum speed. In bits per second, this is the most effective possible data transport rate ( megabits per second ).
There are benefits and drawbacks to every available pricing model.
With metered bandwidth, your connection speed is not limited by the hosting provider’s upper limit but can spike as high as you need it to. This will speed up the download process if the user can do so. Since you pay only for the data you send, your only constraint is the throughput of your uplink port.
You are restricted to a maximum transfer rate (cap) with unmetered bandwidth (10Mbps, 50Mbps, 100Mbps…etc.). Transferred data volume is not monitored. In short, the hosting firm sets a maximum speed and charges you a flat rate regardless of whether you use the full potential of your connection or not.
When using metered bandwidth, check that the monthly allotment can be sent across the uplink port without overloading it.
Suppose your server’s uplink port speed is just 10 Mbps, but it allows 6,000 GB of outbound transfer per month.
If you were to consistently transfer at the maximum speed of 10 Mbps for one month, you would only be able to move 3,240 GB of data with this uplink speed. This means you will need to upgrade to a faster port speed if you want to use your transfer quota fully.
Program That Runs Computers
Hosting providers typically provide access to various OSes, albeit most fall into two broad categories: Linux/Unix-based and Windows Servers.
As far as server operating systems go, OS X Server is a rare bird.
The needs of the programs, libraries, and applications you intend to use on the server should be your first consideration when settling on an OS.
Specific programs can be run on all three major operating systems (Linux, Unix, and Windows), while others are OS-specific.
Linux-based OSes are widely used because of their low resource requirements and purchase prices.
Management Console
Dedicated servers with control panels are simple to administer. They make it easy to create new virtual servers, domains, email accounts, and other actions that would be time-consuming or impossible without them.
Some are open-source and free to use, while most require payment. One benefit of paying for a service is access to additional features and customer assistance. Tech Blog
Cpanel, Plesk, and DirectAdmin are the three most well-known commercial alternatives.
A few of the most well-known open-source options include Virtualmin GPL, W.
If you’re just starting with dedicated servers and are not using a managed hosting service, I recommend setting up a control panel. It’ll simplify server administration, allowing you to spend your time sleeping instead of attempting to figure out why Apache is giving you a “500 Internal Server Error.”
It’s important to remember that certain Control Panels are so intrusive that they can prevent you from installing or configuring software that the Control Panel doesn’t endorse.
What Are the Advantages of Managed vs. Unmanaged Web Hosting Support and Management?
The two primary categories of Dedicated Server plans are “Managed” and “Unmanaged.”
When working with an unmanaged server, it is your job to keep it running smoothly. This means that you are responsible for managing the server’s software. While unmanaged servers are typically less expensive, they might be difficult to administer without prior “System Administration” experience.
To simplify server administration, one alternative is to acquire a control panel.
The hosting provider takes care of your server administration tasks with a managed server. This is your best bet if you are completely green to this or don’t have the time to manage it yourself. However, managed servers typically cost significantly more. There is a spectrum of management available for these managed services. Find out in advance what services are included in the maintenance of your server, as some may require additional fees or software. In contrast, others may only offer support on an as-needed basis.
Remote Power Cycle
Describes a feature that lets you perform a hard reboot immediately. This is typically done through a web interface, and it’s helpful because it offers you the ability to restart the server whenever you choose. Occasionally, the server will crash or freeze due to an issue with the operating system or some software running on it.
This calls for a complete reboot of the device (cutting the power off to the machine and starting again).
Even if you can’t reboot remotely, most hosting providers will do it for you if you submit a support ticket or give them a call. The problem is that the support personnel may take a while to respond to your issue; a remote reboot is a great help in these cases.
Help Desk Technician
Things go wrong, and generally, when you least expect them to. What do you do if your server crashes, your hard disk fails, or your processor overheats? Since you are probably thousands of kilometers from the server, there is little you can do about it. Technical assistance is handling the situation right now.
Companies that provide hosting services can range from being owned by a single person to having hundreds of highly trained employees. Not all services offer the same level of support via email; others only respond to emails during set business hours. When deciding on a hosting provider, consider the quality of customer service they offer and the services they cover. Some hosting providers charge as much as $150 an hour for help outside of what’s included in your package. Also, check their Service Level Agreement (SLA) to check if they cover downtime with a guarantee.
