Thin Clients Booting over a Wireless Bridge
The boot process of a thin client is network-intensive, but once the operating system has been transferred, there is little traffic between the client and the server. As the time required to boot a thin client is a good indicator of the overall usability of the network, this is the metric we used in all our tests.
Our testbed consisted of a 3GHz Pentium 4 with 1GB of RAM as the PXE server. We chose Edubuntu 5.10 for our server, as this (and all newer versions of Edubuntu) come with LTSP included. We used six identical thin clients: 500MHz Pentium III machines with 512MB of RAM—plenty of processing power for our purposes.
When performing our tests, it was important that the results obtained were free from any external influence. A large part of this was making sure that the wireless bridge was not affected by any other wireless networks, cordless phones operating at 2.4GHz, microwaves or any other sources of Radio Frequency (RF) interference. To this end, we used the Azimuth 301w Test Chamber to house the wireless devices (see Resources). This ensures that any variations in boot times are caused by random variables within the system itself.
The Azimuth is a test platform for system-level testing of 802.11 wireless networks. It holds two wireless devices (in our case, the devices making up our bridge) in separate chambers and provides an artificial medium between them, creating complete isolation from the external RF environment. The Azimuth can attenuate the medium between the wireless devices and can convert the attenuation in decibels to an approximate distance between them. This gives us the repeatability, which is a rare thing in wireless LAN evaluation. A graphic representation of our testbed is shown in Figure 2.
We tested the thin-client network extensively in three different scenarios: first, when multiple clients are booting simultaneously over the network; second, booting a single thin client over the network at varying distances, which are simulated by altering the attenuation introduced by the chamber; and third, booting a single client when there is heavy background network traffic between the server and the other clients on the network.
As shown in Figure 3, a wired network is much more suitable for a thin-client network. The main limiting factor in using an 802.11g network is its lack of available bandwidth. Offering a maximum data rate of 54Mbps (and actual transfer speeds at less than half that), even an aging 100Mbps Ethernet easily outstrips 802.11g. When using an 802.11g bridge in a network such as this one, it is best to bear in mind its limitations. If your network contains multiple clients, try to stagger their boot procedures if possible.
Second, as shown in Figure 4, keep the bridge length to a minimum. With 802.11g technology, after a length of 25 meters, the boot time for a single client increases sharply, soon hitting the three-minute mark. Finally, our test shows, as illustrated in Figure 5, heavy background traffic (generated either by other clients booting or by external sources) also has a major influence on the clients' boot processes in a wireless environment. As the background traffic reaches 25% of our maximum throughput, the boot times begin to soar. Having pointed out the limitations with 802.11g, 802.11n is on the horizon, and it can offer data rates of 540Mbps, which means these limitations could soon cease to be an issue.
In the meantime, we can recommend a couple ways to speed up the boot process. First, strip out the unneeded services from the thin clients. Second, fix the delay of NFS mounting in klibc, and also try to start LDM as early as possible in the boot process, which means running it as the first service in rc2.d. If you do not need system logs, you can remove syslogd completely from the thin-client startup. Finally, it's worth remembering that after a client has fully booted, it requires very little bandwidth, and current wireless technology is more than capable of supporting a network of thin clients.
Fast/Flexible Linux OS Recovery
On Demand Now
In this live one-hour webinar, learn how to enhance your existing backup strategies for complete disaster recovery preparedness using Storix System Backup Administrator (SBAdmin), a highly flexible full-system recovery solution for UNIX and Linux systems.
Join Linux Journal's Shawn Powers and David Huffman, President/CEO, Storix, Inc.
Free to Linux Journal readers.Register Now!
- Tips for Optimizing Linux Memory Usage
- Secure Desktops with Qubes: Introduction
- Download "Linux Management with Red Hat Satellite: Measuring Business Impact and ROI"
- Working with Command Arguments
- Fancy Tricks for Changing Numeric Base
- Secure Desktops with Qubes: Installation
- CentOS 6.8 Released
- Linux Mint 18
- Seeing Red and Getting Sleep
- The Italian Army Switches to LibreOffice
Until recently, IBM’s Power Platform was looked upon as being the system that hosted IBM’s flavor of UNIX and proprietary operating system called IBM i. These servers often are found in medium-size businesses running ERP, CRM and financials for on-premise customers. By enabling the Power platform to run the Linux OS, IBM now has positioned Power to be the platform of choice for those already running Linux that are facing scalability issues, especially customers looking at analytics, big data or cloud computing.
￼Running Linux on IBM’s Power hardware offers some obvious benefits, including improved processing speed and memory bandwidth, inherent security, and simpler deployment and management. But if you look beyond the impressive architecture, you’ll also find an open ecosystem that has given rise to a strong, innovative community, as well as an inventory of system and network management applications that really help leverage the benefits offered by running Linux on Power.Get the Guide