April 15, 2013 at 18:2702 - Thin Client and Thick Server
Have a great day everyone! In the last issue of our blog, we talked about what virtualization is and how it works in general terms. Today we’ll talk about the iron on which all this splendor revolves.
The virtualized workstation does not yet work without hardware systems that allow a person to interact with him. We have not yet reached the neural interfaces, so we will consider more mundane technologies.
Many have heard the expression “thin client”? In 2007-2008, thin clients evoked more negative emotions than positive ones. Those that were really thin worked disgustingly, and those that worked well differed little in their device from a regular computer, respectively, and there was not much benefit from them.
Today, the concept of “thin client” meets your expectations. In most cases, this is a box the size of a home router, in which there are no moving parts, a complex and expensive filling that requires attention or maintenance. The main difference between thin clients (with the exception of the hardware features of some models) from the point of view of a company that decided to switch to virtualized workplaces is the presence of a mini-OS inside the “workstations”.
There are two main approaches in the industry. Install a special Windows Embedded (its license is included in the price of the device), based on Windows XP or Windows 7, or use the so-called Zero client, in which the entire OS takes several megabytes, and represents the Linux kernel and a set of micro-utilities and services that provide downloading, retrieving network settings via DHCP and connecting to a server that will “project” the guest OS onto the client.
There is another approach that is used in our thin clients: Smart Zero . Unlike ordinary Zero-clients - in the amount of “firmware”. It is still significantly smaller than Windows Embedded, and takes a couple of dozen megabytes, but contains more utility utilities and a network diagnostics and configuration manager. Smart Zero also supports some proprietary technologies to accelerate network deployment, install new jobs, but we’ll better talk about all the features of our software in a separate topic. :)
What users are working on is more or less clear. And what is the difference between a server and ordinary servers used to store data or work with internal resources?
There are practically no fundamental hardware differences. Of course, it should be understood that the better (more powerful) the server, the more simultaneously connected clients it can contain.
As an example of a server for “virtual machines”, consider the HP ProLiant Gen8 DL160. The classic “rack” server, which is often chosen in the SMB segment as a terminal server.
Inside, as usual, there is nothing supernatural: ordinary server Xeons, a large amount of RAM, and a disk system for the server OS. All features of the terminal server are in the software part. And, since it so happened that Windows has become de facto standard in most offices around the world, we will consider software based on a Windows server.
Windows Server installs the components for the terminal server and the license manager for terminal clients, and their initial configuration is performed. After activating licenses for terminal OSs and initial setup, the server lives its own life, and, in principle, it is not necessary to maintain it: once configured, it will ensure uninterrupted operation for all users using RPS or VDI for their needs.
Such a server can be supplemented with data storage systems with an increased reliability factor, powerful network equipment, and get a highly reliable centralized solution that will allow more than a dozen people to work without thinking that they have a “defective” computer.
There is another interesting case: using a powerful workstation as a VDI server.
Suppose you are an engineer or designer, and work in heavy packages that require really powerful equipment. A good multi-core processor, 32, 64, or even 128 gigabytes of RAM, Nvidia Quadro or Tesla, in general, is a complete mincemeat of high technology and you really know how to find it to use.
And it so happened that you are far from your workstation, and the project urgently needs to be edited, amended, shown or even put the next task to the calculation. In this case, you simply connect from any device that is convenient for you to your computer, which becomes a server, and work with a virtual desktop, as if you have not a kilogram with a small bitten apples on your lap, but your powerful workstation.
With the general theory, we have finished, all of the following posts - on specific practical examples. :) See you on the pages of Habr.
UPD from Apr 22, 2012: We launched a special project on Habré, which we talked about in the zero post of this season. It is located at: habrahabr.ru/special/hp/commercialand there is a lot of interesting information, the promised feedback form, as well as a test for knowledge in the field of virtualization. We are waiting for you and your questions!
The virtualized workstation does not yet work without hardware systems that allow a person to interact with him. We have not yet reached the neural interfaces, so we will consider more mundane technologies.
Many have heard the expression “thin client”? In 2007-2008, thin clients evoked more negative emotions than positive ones. Those that were really thin worked disgustingly, and those that worked well differed little in their device from a regular computer, respectively, and there was not much benefit from them.
Today, the concept of “thin client” meets your expectations. In most cases, this is a box the size of a home router, in which there are no moving parts, a complex and expensive filling that requires attention or maintenance. The main difference between thin clients (with the exception of the hardware features of some models) from the point of view of a company that decided to switch to virtualized workplaces is the presence of a mini-OS inside the “workstations”.
There are two main approaches in the industry. Install a special Windows Embedded (its license is included in the price of the device), based on Windows XP or Windows 7, or use the so-called Zero client, in which the entire OS takes several megabytes, and represents the Linux kernel and a set of micro-utilities and services that provide downloading, retrieving network settings via DHCP and connecting to a server that will “project” the guest OS onto the client.
There is another approach that is used in our thin clients: Smart Zero . Unlike ordinary Zero-clients - in the amount of “firmware”. It is still significantly smaller than Windows Embedded, and takes a couple of dozen megabytes, but contains more utility utilities and a network diagnostics and configuration manager. Smart Zero also supports some proprietary technologies to accelerate network deployment, install new jobs, but we’ll better talk about all the features of our software in a separate topic. :)
What users are working on is more or less clear. And what is the difference between a server and ordinary servers used to store data or work with internal resources?
There are practically no fundamental hardware differences. Of course, it should be understood that the better (more powerful) the server, the more simultaneously connected clients it can contain.
As an example of a server for “virtual machines”, consider the HP ProLiant Gen8 DL160. The classic “rack” server, which is often chosen in the SMB segment as a terminal server.
Inside, as usual, there is nothing supernatural: ordinary server Xeons, a large amount of RAM, and a disk system for the server OS. All features of the terminal server are in the software part. And, since it so happened that Windows has become de facto standard in most offices around the world, we will consider software based on a Windows server.
Windows Server installs the components for the terminal server and the license manager for terminal clients, and their initial configuration is performed. After activating licenses for terminal OSs and initial setup, the server lives its own life, and, in principle, it is not necessary to maintain it: once configured, it will ensure uninterrupted operation for all users using RPS or VDI for their needs.
Such a server can be supplemented with data storage systems with an increased reliability factor, powerful network equipment, and get a highly reliable centralized solution that will allow more than a dozen people to work without thinking that they have a “defective” computer.
There is another interesting case: using a powerful workstation as a VDI server.
Suppose you are an engineer or designer, and work in heavy packages that require really powerful equipment. A good multi-core processor, 32, 64, or even 128 gigabytes of RAM, Nvidia Quadro or Tesla, in general, is a complete mincemeat of high technology and you really know how to find it to use.
And it so happened that you are far from your workstation, and the project urgently needs to be edited, amended, shown or even put the next task to the calculation. In this case, you simply connect from any device that is convenient for you to your computer, which becomes a server, and work with a virtual desktop, as if you have not a kilogram with a small bitten apples on your lap, but your powerful workstation.
With the general theory, we have finished, all of the following posts - on specific practical examples. :) See you on the pages of Habr.
UPD from Apr 22, 2012: We launched a special project on Habré, which we talked about in the zero post of this season. It is located at: habrahabr.ru/special/hp/commercialand there is a lot of interesting information, the promised feedback form, as well as a test for knowledge in the field of virtualization. We are waiting for you and your questions!