Why Use Server Hardware for Servers?
- by Antoniy Yushkevych
- in Server
- View 171
There are desktop computers and then there are server computers but how different are they exactly. Well, any computer actually has the same basic components processor, working memory, storage etc. and you may be surprised to hear that fundamentally they're not that different with some of those components even coming off the same assembly line.
But that doesn't mean that you can just grab off the shelf desktop components in like fashion yourself a server. I mean sure core i7 runs at 3.8 gigahertz and Xeon runs at 3.8 gigahertz but it costs way more. Well, sometimes there's more to it than gigahertz and gigabytes.
The average customer is often indifferent toward the type of hardware and technology products available in today’s market. However, opinions would be definitely swayed if the average customer knew the significant differences between consumer-grade technology and enterprise hardware. Here’s a look at how consumer technology and enterprise hardware differ:
When it comes to mainstream and high-end equivalents, the customer sticks with the big names with the assumption that the product is as good (if not better) than enterprise technology. Of course, there’s a bit more validation behind the name and some extra customer support. However, consumer technology couldn’t be further apart from enterprise hardware in terms of frills and features.
One of the major differences is that enterprise hardware is built from the ground up, meaning that the issues with mass production rarely happen, as more care is involved when the product is being built. Another would be the fact that the hardware is less flashy-looking than what is provided in the consumer-grade materials.
It wasn’t long ago when many poorly made capacitors hit the motherboard market. Since much of the problem was due to water corrosion and improper electrolytes, it shows that not all capacitors are created the same. Enterprise hardware is known to use higher grade caps, with many consumer-grade capacitors being known for charring or leaking when put under the pressure that enterprise hardware is built to sustain.
Consumer-grade systems rarely need more than 8GB of RAM, but enterprise hardware usually needs at least 16GB (or higher) to simply operate. This helps explain why consumer-grade only comes with four memory slots and enterprise hardware comes with eight slots for memory. In addition to this, enterprise-grade materials feature RAM support with error-correcting code (and consumer-grade products omit this kind of support completely).
Consumer-grade platforms are known to be in a hurry whenever dispensing with their legacy interfaces. The reason for this is due to the design complexity, cost, and the fact that only a few users actually operate with older interfaces. Since consumer-grade materials are known to dispose of older technologies as they evolve, enterprise hardware develops software so that it can work around their older formats. The amount of legacy support on enterprise products has increased over recent years, due to the service contracts and consumer warranties associated with them.
Enterprise technology has its fair share of heat generation, so the technology has to fit accordingly. Enterprise software is set up so that it can operate in high-demand environments, so the exhaust paths do no overlap and the chips are set up to receive proper cooling. In addition to impacting performance, the enterprise layout is set up for longevity.
The consumer-grade platforms of today are being developed to compete with enterprise hardware’s management systems, but enterprise hardware still has an edge. Where enterprise hardware really shines is in the Intelligent Platform Management Interface (IPMI) department, as most systems can still operate if they have crashed or shut down.
Consumer-grade and enterprise-grade technology both have a wide range of casing options to select from. One advantage that enterprise casing hardware offers is that it can convert into a 4U chassis. This type of hardware is especially handy when hardware needs to be transported or permanently placed in a data center.
Considering the information above, enterprise hardware and consumer technology have significant differences. Although enterprise hardware seems like the route to follow in terms of overall performance, the decision whether to turn to consumer-grade materials all comes down to how you’ll be using the technology.
Designed to last longer and more reliable
Server hardware is designed to have a longer time span that consumer hardware, because servers in general stay in production for a longer time than consumer hardware. Moreover, servers are designed to be on 24/7, unlike personal machines that most people turn them off for hours daily.
Mean Time Between Failure or MTBF is a unit that measures the average time a component will take to failure under optimal operation conditions. As an example, according to western digital WD Gold MTBF is 2.5 million hours. On the other hand, consumer hard drives do not come with MTBF, however, they tend to fail after 500 thousand hours.
Any enterprise will have more profit if they used their equipment for a longer time and less downtime, because their ROI will be higher.
The other reason IT pro choose servers grade hardware over consumer-grade hardware is its high fault tolerant and redundancy. As an example, servers come with two PSUs, to prevent downtime because of a dead PSU.
Fault tolerance allows devices to run even if one or more of its components is faulty. As a result, no service will be interrupted and sysadmins will be happy.
The Software Differs as Well
Now that you have seen the differences between server and PC hardware, we will discuss the differences between the Operating Systems they use.
One component in particular needs to be available so that application software can access a computer’s hardware resources: the operating system. No alternative exists that allows users to operate physical machines, as these serve as intermediaries operating between material resources, users, and their respective applications.
In order to carry out this task, every operating system is composed of a series of programs that enable users to interact with and manage individual hardware components. These programs are made up of three fundamental elements:
- The kernel,
- the command-line interpreter,
- and the file system
The kernel, also known as the system kernel, constitutes the system’s basic functions and is responsible for tasks like managing storage as well as other individual processes. The command-line interpreter allows the computer user to communicate with the operating system with the help of commands, and the file system makes it possible to store files within a file tree.
It’s clear that operating systems aim to achieve a diverse array of goals, and these have gone through some radical changes over the years. Protective functions, like distributing working memory or limiting user rights didn’t play any role whatsoever in the early days of developing operating systems. However, these features are key in helping to thwart both program and system crashes and preventing outside attacks.
In the early days, when computer systems started to become increasingly interconnected with one another, operating systems were needed to administer servers, like file, database, or web servers.
Due to the requirements that such servers put on operating systems, some systems may be more suitable for these tasks than others, which is why these set-ups are often referred to as server operating systems.
In the following paragraphs, you’ll find out more about the origins of server systems, the first systems used in server environments, and which server operating systems dominate today’s scene.
From text-based user interfaces to complex program systems
In 1974 Gary Kildall released the operating system CP/M in floppy disc form. Given that Unix, which appeared a few years earlier, could only be used within specific environments (e.g. universities and research institutions), it’s widely considered to be the first PC operating system for home use. Originally an acronym for Control Program/Monitor, CP/M’s basic input/output system (BIOS) was a standardized hardware interface and could thus be used for different computers.
With the help of the command line interpreter CCP (Console Command Processor), the system was able to receive and execute user commands. What’s more, CP/M is the only noteworthy disc operating system (DOS) that was available for microcomputers prior to 1980, making it the model upon which a concept that was to become the dominant administration solution was based.
It should be mentioned, however, that DOS didn’t present a unified operating system for quite some time; instead, it was more a concept that many different providers adopted and implemented in varying ways. In the end, the well-known MS-DOS by Microsoft, which played a fundamental role in further Windows versions (until Windows 98), was the model that caught on.
DOS was to IBM computers as ‘System’ — which was renamed to Mac OS in 1996 — was to Apple computers. With the release of Apple Macintosh in 1984, users became acquainted with the company’s initial graphical user interface, which could be operated with the aid of a mouse.
Apple had already released a similar operating system, called Lisa OS, a year prior to this point; due to its $10,000 price tag, however, this model proved to be more-or-less a flop.
Microsoft dominates the home PC territory while Linux becomes the go-to for servers
It wasn’t until 1993, with the introduction of the Windows Server operating system, Windows NT 3.1, that Microsoft was finally able to deliver its answer to Apple’s success.
Up until this point, the software company had only been successful with the release of its Windows 3.0 graphical interface for MS DOS. Even if the NT system, which came in workstation as well as server varieties, never managed to quite catch on, Microsoft nonetheless used the kernel as the basis for most of its subsequent versions (2000, XP, Vista, 7, 8, 10); the exception here being the Windows 9x series, which was also based on MS DOS and was the starting point for Window’s unparalleled success within the private sector. Further variants from this line include names like Windows 98 and the Millennium Edition, which simultaneously signaled the end to this range of products.
In terms of server operating systems, another alternative began to meddle with things: Linux, a solution based on the pioneering operating system, Unix. Initially offered as a proprietary option, the applied Linux kernel was made available under the free GLP license in 1992, a move that generated considerable buzz among developers and sowed the seeds for the operating systems broad dissemination.
When combined with utility software, various distributions, like Yggdrasil Linux or Slackware, emerged — these acted as complete software systems and, like the kernel itself, didn’t involve any additional costs. And given that they can be expanded and adjusted according to user demands, Linux distributions weren’t only free-of-charge: they were also highly flexible.
The fact that Linux offered excellent options for system and rights management led to the operating system quickly becoming the preferred server solution for many administrators, especially for web development, a status that the open-source options enjoy to this day.
Today’s server operating systems market
In the age of web 2.0, searching for the right server operating system is a public affair. While the web was once a place in which the majority of users only consumed available content, today’s internet landscape looks a bit different: it’s increasingly transforming into a reservoir of many different types of web projects; often these are actively generated and co-designed by the internet community.
Whether this means running a blog, maintaining a standard web presence for a business or organization, or offering a web application, all of these projects rely on web servers, which are set up and managed by operating systems.
When setting up a web presence, a decision needs to be made as to whether putting together a hosting environment is a task that’s best delegated to others, i.e. a provider, or taken into one’s own hands. Whatever you decide to do, many choices await you once you’ve made a decision.
Even today, Linux plays a leading role: over half of all Unix-like systems, which according to W3Techs are used for roughly two thirds of all business projects, are Linux distributions (often referred to as distros).
Debian and Ubuntu are especially popular distributions, as is CentOS; less frequently used models, but nonetheless equally noteworthy, are Red Hat, Gentoo, Fedora, and SUSE.
Microsoft is the biggest competitor of software operating systems running on Linux and/or Unix. As a part of its NT series, the software giant has had the Windows Server Edition on offer since 2003.
The saga of Linux vs. Windows has mostly stagnated over the past few years, as there are hardly any more deciding qualitative differences between the two when it comes to operating server systems; most often a decision is made based on cost factors or personal preferences.
The Linux Distributions
It doesn’t always come down to a simple choice between Windows and Linux. If you opt for the latter, there are many different distributions to choose from. For example, the Debian derivative, Ubuntu, may not preside over the enormous range of functions that its predecessor does, but it’s precisely this trimmed-down approach that makes it easy to set up.
What’s more, users making the switch from Windows or other competing operating systems will be greeted with lots of useful help settings. Debian itself, on the other hand, is much more flexible in its use and supports a wide range of architecture. As one of the pioneering distributions, it also features top support in the form of an active community.
Red Hat Enterprise Linux (RHEL), the commercial system of the Red Hat developers, has been especially designed with the needs of businesses in mind; the system is also stable and enjoys the support of a diverse range of independent software developers.
Unlike RHEL, Fedora and its derivative, CentOS, are both free of charge; however, they understandably don’t come with the same support found in other options. One of Fedora’s defining features is that it’s constantly being updated, a trait that makes it particularly popular among developers.
Those aiming to put together their own individual Linux systems will appreciate Gentoo, another system known for pumping out update after update. Like Fedora, however, this server OS lacks long-term support, making it less suitable for projects that are to span over a longer period of time.
Due to the high prices of server hardware and software, for most starting out businesses or users that would like to have a server, it is simply not viable to spend such large amounts of cash. This is exactly why dedicated server providers exist.
For a mere fraction of the price, you can get an extremely powerful machine that will satisfy all your dedicated server needs. Not only will you get the hardware, but also the appropriate infrastructure for the server to run smoothly and a team of data center technicians ready to support you if anything goes wrong.
We at MonoVM provide excellent dedicated servers with a wide variety of plans to choose from. Pssst, they are also customizable to your preferences.