¡°We had a lost package issue, and it was so incredibly well handled. Another snag surrounding the used VISA card unusable, and Newegg worked around it famously. We are astounded and so happy with the service and speed by which our issues were resolved. A thousand times thank you to an amazing company. We love you. I posted your amazing feats on facebook, and I'd write any more positive reviews you'd wish. Thanks for listening to the pleas and troubles of your loyal customers!¡±— Lyndsey 9/5/2016 3:34PM
Different server computers are used for different tasks. Whether a server is storing media files accessible over a home network, or used for enterprise database management, knowing which type of server system best suits your home or business network saves money when it comes time to buy a server.
Why would you need a server in the first place? As the number of endpoints grows beyond five or ten users, the peer-to-peer network model wherein computers communicate through the router eventually becomes inefficient and bottlenecked. A few tell-tale signs that you need a dedicated server include:
So-called cloud servers offer leased data storage and computing resources that companies might buy in lieu of server hardware. This is ideal for a very small business or a casual home user without any server management know-how. As a company scales up it usually makes financial sense to deploy physical servers, hiring in-house (or outsourced) server management personnel, and perhaps utilizing cloud servers for redundancy or part of a backup plan.
A businesses server stores all the files, applications, operating systems, programs, and peripherals accessible to computers on an office network. IT administrators use server software to manage user access by way of credentials or authentication information. A Web server faces the Internet, and makes available data, content and applications beyond the physical local area network (LAN).
Just a few years ago, one server typically handled a one task for office IT. One provided user credentials for accessing a database; another dedicated server managed the sending and receiving of e-mails; a different server was used for distributing Windows, another server for hosting a Web domain, and so forth.
This might still be standard practice at a small business, but server management has changed drastically for larger companies within the past decade. Now, medium and large organizations can save on server hardware costs by consolidating infrastructure onto one or a few enterprise servers.
Modern enterprise servers can tackle several tasks with virtualization technology. Server virtualization uses a software application called a hypervisor to divide the computing power of a physical server into several virtual machines (VMs)—commonly called instances, or containers in IT semantics. When looking at system specifications for servers for virtualization, look for high memory (RAM) capacity and number of CPU cores and CPU threads—all of which factor heavily in workload consolidation.
Knowing a few specifics about server hardware will help tie server specifications to system needs when buying a server or building one from components.
The basic component structure of a server is similar to that of a PC. Since most servers are meant for 24/7 operation; and servers for virtualization have powerful, multi-core processors and expansive disk space to handle a heavier computing load and data storage, server components are more robust.
Servers for virtualization have motherboards with LGA 2011-3 sockets that support larger CPUs with the X99 chipset (Intel Xeon E5 and Intel Core i7 X-series processors, for example) that have multiple cores and threads for running virtual machines. AMD server CPUs have their own socket types and are not compatible with Intel. Some motherboards have dual CPU sockets to add processor redundancy for 24/7 uptime.
When you talk about scalability in servers, one huge consideration is the number slots in the motherboard. The more slots there are, the more room there is for expansion. Different types of slots include DIMM slots for RAM, and PCI-e slots for graphics cards and next-generation solid-state storage drives.
The number of cores and threads dictates performance for server CPUs. Clock speed, measured in GHz, plays a role in how fast the processor executes instructions, but in terms of functionality for modern servers, frequency does not gauge performance as much. A larger amount of cache memory boosts performance: the cache stores frequently accessed data for faster recall, which speeds up server performance in many situations.
Budget-level dedicated servers and NAS typically have a low-power dual- or quad-core processor (Intel Core i3) typically found in PCs; more expensive servers have CPUs capable of greater processing power are more common in large tower servers and rack models.
Every server is meant for network use. Bandwidth considerations are important; inadequate networking components will sap performance for the system. For virtualized servers, a dual-port or quad-port network interface card (NIC) is recommended for greater connectivity, and a 10 gigabit (Gbe) NIC boosts performance for environments with an increased number of end users.
File servers—sometimes called storage servers or Network Attached Storage (NAS) —have room for multiple hard drives that protect against data loss in the event of drive failure, a configuration called a RAID array. They are meant to make available data for a pool of end users. A group of file servers configured together in as Storage Area Network (SAN) improves data availability across larger networks, and can improve application performance over multiple satellite office networks.
Enterprise hard drives have built in fault tolerance and a faster interface (SAS) than those in consumer drives (SATA) SAS drives have a premium price point. Make sure to note compatibility when you buy storage drives, as SATA and SAS drives are not interchangeable. Other interfaces for server storage include USB 3.0 for external drive connections, and Thunderbolt, the high-speed Apple interface.
Server cases have different form factors. A MicroServer is a square-ish server box, typically used as a file server or NAS, or a low-power single use server. It supports basic RAID and usually has a low power server processor, ideal for a small home office. Tower systems resemble full-size desktop PCs and vary in computing power and specifications; low-end models offer good value and are easily expandable—ideal for back office and retail settings.
Medium and large companies usually store IT infrastructure in server racks. Rack servers have chasses designed for stacking vertically in server racks. Size measurements for standard rack servers are designated by units starting at 1U, which has a width of 19 inches and 1.75 inches in height. Servers range from 1U-4U. The standard height for a server enclosure is 42U tall.
Server software manages user access to different parts of the data infrastructure using credentials or authentication information, and makes available data (content) beyond the physical LAN. Windows Server is the most commonly known administration system among IT professionals, which makes support relatively easy to procure. The value proposition to buy Windows Server licensing hinges on readily available support technicians.
Linux systems, which are popular especially for larger businesses, are managed with open source software that is free of licensing costs. The tradeoff is that Linux systems have more custom-built features, which places a higher demand on specialized administration staff that can command a higher fee for support services.
In most cases a home server is used for storing digital media and streaming files to devices connected on a home network, and offers a so-called private cloud solution for home users to store files accessible from beyond the network. This is typically done with a NAS or single use server that sits on a home network. NAS servers have simple graphical user interface (GUI) operating systems proprietary to the manufacturer; there are open source options like FreeNAS which are popular.
Windows Storage Server offers value for a SMB user that uses a file server in a professional capacity for the use-as-needed Microsoft support that comes with the management software. It also supports streaming media to Xbox 360 and other digital display devices that can access a home network.
A user may host web pages or an e-mail domain on an inexpensive home server. Apache server software is commonly deployed for web hosting management.
What’s the difference between a server and a workstation?
A server and a workstation are not that different on a fundamental hardware level. Server hardware and PC hardware have many have similar specifications; the main difference is that server hardware is meant for 24/7 operation. Uptime is absolutely critical for a server, thus making the hardware architecture of servers and server components different—there is built in hardware redundancy, and is optimized for different workloads. These features make server components (mostly) incompatible with standard PC components.
A workstation is a high-power desktop PC that is meant to be used directly by a user. A server computer interfaces with other computers over a network—it does not have a GUI, and is not meant for direct interaction with users in the conventional understanding. Server administrators configure server machines using another computer using the textual command line, or an object-oriented programming language like PowerShell for Microsoft Windows.