Friday, April 4, 2014

ZERO CLIENT VS THIN CLIENT

WHILE THE TERM zero client is something of a marketing buzzword, it is a useful way of differentiating options for the devices that are used to access desktops. A zero client is similar to a thin client in its purpose—accessing a desktop in a data center—but requires a lot less configuration.

Zero clients tend to be small and simple devices with a standard set of features that support the majority of users. They also tend to be dedicated to one data center desktop product and remote display protocol. Typically, configuration is simple—a couple of dozen settings at the most, compared to the thousands of settings you see in a desktop operating system. Zero clients load their simple configuration from the network every time they are powered on; the zero clients at a site will all be the same. Zero clients support access to a variety of desktop types, terminal services, virtual desktop infrastructure (VDI) or dedicated rack mount or blade workstations.

The basic premise of zero clients is that the device on the user’s desk doesn’t have any persistent configuration. Instead, it learns how to provide access to the desktop from the network every time it starts up. This gives a lot of operational benefits, since the zero client devices are never unique. This contrasts with a thin client, which may have local applications installed and will hold its configuration on persistent storage in the device.

Thin clients became a mainstream product class shortly after Microsoft introduced Windows terminal Server and Citrix launched MetaFrame, both in 1998. To enter this market, PC manufacturers cut down their desktop hardware platforms. They repurposed their PC management tools, reusing as much technology as possible from existing PC business. This meant that a fairly customized Windows or Linux setup could be oriented toward being a thin client.

Over time optional features for USB redirection, a local Web browser, VOIP integration agents and multi-monitor display support were added. Each additional feature adds configuration and complexity to the thin client. After a few years, thin clients are really small PCs. Some even have PCI or PC Card slots added. These thicker thin clients get quite close to a full PC in terms of capabilities and complexity. Instead of simplifying management, IT administrators now needed to manage the device on the user’s desk as well as in the data center. Zero clients, then, are a return to the simpler devices on user’s desks—with simpler management.

Zero clients are much simpler to manage, configure and update. Zero client firmware images are a few megabytes, compared with the multiple gigabytes that thin client operating systems take up. The update process itself is much quicker and less intrusive on a zero client, possibly occurring every day when the client boots.

Thin clients need to be patched and updated as often as the desktop operating system they carry; since zero clients have no operating system, they need less frequent updates. Zero clients have few knobs and switches to turn—probably fewer than 100 configuration items in total—so they are simple to manage. Often, their bulk management is a couple of text files on a network share. Thin clients have a whole operating system to manage, with tens of thousands of settings necessitating complex management applications, usually on dedicated servers at multiple sites. A zero client is like a toaster. A consumer can take it out of its packaging and make it work. If the consumer is an employee at a remote branch, there are benefits to having that worker be able to deploy a new terminal. Sometimes, thin clients need special builds or customized settings applied to them before they are deployed. This obviously is not ideal for rapid deployment. The ability to rapidly scale can be important when it comes to something like opening a call center to accommodate an advertising campaign or a natural disaster response. Zero clients have lower power consumption. Thin clients have mainstream CPUs and often graphics processing units, but a zero client usually has a low-power CPU (or none at all), which cuts down on power consumption and heat generation.


The simplicity of zero clients also makes for a much smaller attack surface, so placing them in less trusted networks is not so worrying. Also, putting them in physically hostile locations is safe; lower power and usually passive cooling mean that heat, dust and vibration are less likely to cause maintenance problems. Zero clients are all the same. Models are released every couple of years rather than every few months, so your fleet will contain fewer models. That means there’s no need for help desk calls to move a device from one desk to another. Plus the user experience is consistent. Your supplier’s inventory of zero clients will also have fewer models, which should lead to better availability when you need new zero clients.

No comments:

Post a Comment