Enterprise V Consumer

 


Blog 1  

The past week I have researched the fundamentals of computer systems within the enterprise space the main issue is that of power and efficiency. The basic principles of enterprise and consumer computers is that they both share the same components such as same cores, architecture, ram, etc. The most notable different before comparing specs is the price of enterprise and consumer grade hardware. When taking a closer look server grade hardware tends to be much more compatible with a plentiful of different pieces of tech. 

Before I dive into a more in-depth overview of differences of each hardware piece a brief history of servers through the years is much needed. In 1989 Tim Berners-Lee a British scientist invented the World Wide Web while work at CERN. The original plan was for the WWW to be developed to meet the demand for automated information sharing between scientists within university's and institutes around the world. The first Web Server, running NeXT computer at CERN this being just one of the many servers that have started major leaps in technology as many companies such as AWS, Google, IBM, etc have massive data centers with the most modern and technologically advance servers reside. 

CPU's:

On the main topic of hardware, I intend to talk parts of both specifications of a hardware break down comparing functionality, appearance and its costs comparison. when comparing parts such as CPU a consumer cpu may be 3.x ghz and the server grade may also be that of the same value but hidden behind the server grade is support for some big advantages such as ECC memory support, features like this are crucial in servers where data protection is at the utmost important. Having a broad spectrum of compatibility allows for servers to very precisely select what specifications it uses. Consumer processors tend to stick around the range of 2,4,6,8,10,12,16 while server grade core counts 12,16,24,32,64 another trade off server grade processors tend to use is core speed for core count. This allows the usage of more PCIE lanes because of the processors design and intended use. The other main benefit for servers is that of generational compatibility with socket type many Intel and AMD sockets for the processor use the same socket as the previous generation this is to allow the trade off of upgrading and compatibility with newer and older components. 

Motherboards:

Further more motherboards are a clear distinction visually and functionality wise when compared from enterprise and consumer grade options. In the industry of motherboard manufacturing there are standard universal form factors that range from a broad spectrum of sizes the reason again for this is to allow as much a broad compatibility between hardware. The list of such said form factors start at Mini-ITX, Micro-ATX, ATX, E-ATX, XL-ATX, SSI EEB and lastly HPTX. Those listed respectively starting from smallest form factor to the biggest form factor, Consumer boards using from Mini-ITX to XL-ATX and servers using SSI EEB and HPTX primarily. Visually aside from the size the more drastic difference is the difference of available sockets for a processor to be installed on the motherboard. Servers once again primarily will run in a dual mode cpu configuration meanwhile consumer motherboards will only support one socket for one processor.

HDD:

Unlike most consumer grade computers size constraints enterprise grade size is modified and standardized into space allotments such as 1U, 2U and 4U in server racks to fit within the chassis that holds the components. Other added benefits of using enterprise fitted components and the differences are other parts such as Hard Disk Drives, for companies who make such storage solutions for example Seagate, Western Digital and Toshiba, after further research into looking at the products descriptions and spec’s the major differences are that of higher RPM and better sustained reads/writes as well as better longevity of its life span. One notable feature that is different between data center hard drives and consumer grade hard drives is extra precautions from unavoidable causes of damage such as vibration so within data center drives they are built more less susceptible to that sort of damage leading to a less likely broken drive to fault because of its tolerances within a confined space. Many components that are initially very good when considered in consumer desktops lack certain features and extra tolerances that are crucial in server grade setups. 

GPU:

Some of these features such as resistance to fault tolerances and difference of design are vital to a server type system hardware such as CPU coolers and graphic card cooler are built completely different to meet the needs of each system. Examples of this is that of regular gaming Nvidia and AMD Cards most of these cards have in mind to be placed within a consumer’s desktop so cooling on its own is vital as well as aesthetically pleasing. On the other hand, Data center graphics cards will not do any sort of cooling on its own and rather have the chassis intake air from the front of the server case and exhaust the hot air out the back primarily skipping plastic shrouds, fans and RGB lights. Most of these cards such as the A100 a NVIDIA Ampere Architecture can perform just as well as any consumer card but is designed around computational workloads and not videogames. Another key feature that have a difference between consumer cards and enterprise is power efficiency, on average consumer gaming computers may use up to 823 kw/h meanwhile datacenters pull an average of 90 billion kw/h.

References: 

https://home.cern/science/computing/birth-web/short-history-web#:~:text=By%20the%20end%20of%201990,This%20machine%20is%20a%20server.


Comments

Popular posts from this blog

Server Configuration Blog Post 2