What exactly is a “zero client”?

There are basically two different types of approaches to zero clients: zero clients for OS streaming and dedicated hardware zero clients for accessing certain specific virtual desktops and physical workstations.

1. Zero clients for OS streaming

“Zero OS” devices use a PXE (Preboot eXecution Environment) to download their operating systems from the network. This can either be a lean, minimal PC operating system or a lean, minimal thin client OS for establishing exclusive access to a given solution environment, such as Citrix XenApp / XenDesktop.

1.1.    Zero client for OS streaming

This type of zero client works like a “fat client” with local OS loaded onto its RAM, such as Windows XP Pro.

Problematic issue(s):

• The repeated boot-up of the extensive operating system requires additional server capacity and strains the network, especially at peak times (such as in the morning, at the start of the workday).

• Local operation of Windows requires “fat” hardware (a lot of RAM; a powerful, fast processor), making this approach no longer viable given today’s virtual desktop infrastructures (VDI).

1.2.    Zero client for thin client OS streaming (PC system)

These zero clients function like a thin client with a local OS loaded onto the RAM, such as for Citrix ICA or Microsoft Remote FX.

Problematic issue(s):

A lean, minimal thin client OS limits future viability of the investment: There is only proprietary (exclusive) access available to certain server environments (only ICA, PCoIP or RemoteFX).

2. Hardware zero clients

Hardware zero clients possess proprietary hardware designed to access a specific server solution or a corresponding, dedicated workstation.

2.1. Hardware zero clients with “zero CPU” and “zero RAM”

These zero clients possess minimal hardware without their own CPU (main processor) and RAM (working memory) sections. Their communications logic is permanently installed on the device hardware circuitry and as a result can only be updated with great difficulty, if at all (proprietary protocol).

Problematic issue(s):

• An exclusive, proprietary design: Access restricted to a certain server or VDI solution limits the future readiness and flexibility to just a few years.

• A lack of technological openness/flexibility: Such devices cannot make use of new protocols or protocol extensions (such as Microsoft RemoteFX or SPICE). The devices become obsolete faster.

• Limited potential for savings: Zero clients with their own proprietary methods of VDI access cannot access highly efficient, prevailing standard solutions such as Microsoft Windows Server 2008 R2 or Citrix XenApp.

• Greater power and resource requirements in the backend: These zero clients do not have any local computing and graphics capabilities or local software tools to decrease the workload on the central server. The result: What may look good when testing an individual unit does not work nearly as well once you reach just ten users (degraded performance).

2.2. Hardware zero clients for PCoIP: Is this really zero management?

These devices are optimized to function with the PCoverIP (PCoIP) transmission protocol developed by Teradici. They have a proprietary OS, a proprietary CPU as well as a block RAM chipset. They decode the image information from a VMware View desktop (as of Version 4 with PCoIP support) or a dedicated physical workstation by means of a Teradici® host chipset (1:1 connection).

Problematic issue(s):

• Proprietary hardware and software means a lack of secure future readiness.

• There are high resource demands on the server side.

• The 1:1 workstation connection solution requires an additional host PCoIP card. This relatively expensive solution is therefore only suitable for power users with particularly high system demands, such as for CAD applications.

Zero client concepts

Fig. 1: OS streaming; Fig. 2: Hardware zero clients with Pano Logic® or Teradici®/PCoIP® technology

Next Chapter: Functionalities: What Exactly is a Zero Client?