Communications (rahul.net)
Navigation
Dedicated Servers | Rate Chart | FAQ | Demo 1 | Demo 2 | Getting Started | Activity Links | Site Design
On This Page
Top | Specifications | Additional Info | Notes | Performance | Misc Fees | Glossary | Sign up | Contact Us | Notice | Bottom
| Item Number | Monthly | Base Model | CPU | RAM | Disk |
|---|---|---|---|---|---|
| A5-SM | $179 | Supermicro 5013C-i (manuals) | 1 x 2.80 GHz Pentium 4, Hyperthreading, 1 M cache, 800 MHz bus | 1 G, ECC | 2 x 80 G 2M 7200 rpm, WDC WD800BB, IDE, same bus |
| A3-DP (out of stock) |
$163 | Dell PowerEdge 750 (manuals) | 1 x 2.80 GHz Pentium 4, Hyperthreading, 1 M cache, 800 MHz bus | 512 M , ECC | 2 x 80 G 8M 7200 rpm, Maxtor 6Y080M0, SATA, separate buses |
| C7-SM | $139 |
Supermicro 5012-BE (manuals) | 1 x 2.40 GHz Celeron, 128 k cache, 400 MHz bus | 896 M, ECC | 2 x 80 G 2M 7200 rpm, Samsung SP0802N, IDE, same bus |
| J3-GE (out of stock) |
$109 |
Generic with Supermicro P4SGE motherboard (manuals) | 1 x 2.00 GHz Celeron, 128 k cache, 400 MHz bus | 512 M, no ECC | 2 x 80 G 2M 7200 rpm, IDE, same bus |
TQMbox is a service mark of A2I COMMUNICATIONS.
CPU frequences are in decimal GHz as rated by the manufacturer. RAM and cache sizes are in binary kbytes or Mbytes. Disk capacities are in decimal gigabytes as rated by the manufacturer. See binary units and decimal units in the glossary. RAM refers to main memory.
All billing is synchronized to the 1st of each calendar month. Your first payment will be for the current partial month and, if your service begins on or after the 15th of a month, for the next full month.
There is a nonrefundable $35 setup fee. We reward you with a credit of double this setup fee after your server has remained in service, and you have remained a customer in good standing and made no late payments for any service from us, for 12 months.
Cancellation policy: During the first 30 days of your service, you may cancel at any time, and billing will be pro-rated for the number of days of service. Beyond the first 30 days, cancellation requires at least 5 days' notice and will be effective on the 1st of the following month.
Item |
Performance | Hardware |
OS | Availa- bility |
Monthly Volume |
|---|---|---|---|---|---|
| A5-SM | 1.06 (-j1) 1.17 (-j4) |
fan speed, voltages, temperatures | rh9, fc2, fc3, cen33, cen40, mdk10, mdk10.1, pldr | 5 days | 600 G |
| A3-DP | 1.09 (-j1) 1.27 (-j2) |
no | fc2, fc3, cen33, cen40, mdk10.1(*) | 15 days | 600 G |
| C7-SM | 0.47 (-j1) 0.47 (-j4) |
fan speed, voltages, temperatures | rh9, fc2, fc3, cen33, cen40, mdk10, mdk10.1, pldr | 5 days (limited quantity) | 600 G |
| J3-GE | to be announced | fan speed, voltages, temperatures | rh9, fc2, fc3, cen33, cen40, mdk10, mdk10.1, pldr | May 15, 2005 | 600 G |
Higher performance numbers corresponds to faster machines. Details of how we calculate performance numbers are provided below (see “Performance”).
The OS column lists operating systems that are believed to be compatible with each hardware configuration.
The Availability column gives the estimated number of business days for us to get your server online and ready for use after we accept your order.
Abbreviations: rh9 = RedHat Linux 9; fc2 = Fedora Core 2; fc3 = Fedora Core 3; cen33 = CentOS 3.3; cen40 = CentOS 4.0; mdk10 = Mandrake Linux 10; mdk10.1 = Mandrake Linux 10.1; pldr = PLD Rescue CD.
The notation “2 x 80 G 8M 7200 rpm” means: two disks of 80 decimal gigabytes each (total 160 decimal gigabytes); 8 M cache; and rotational speed 7200 rpm. "Separate buses" means that each of two disk drives is on a different IDE or SATA bus.
Hardware monitoring in the table above refers to the ability to get a reading of hardware data such as fan speed(s), various motherboard voltages, and CPU and motherboard temperatures. This is done using free software supplied by the server vendor. This software has been tested successfully with rh9, fc2, fc3, cen33, cen40, mdk10, and mdk10.1. Once you install the software, you will be able to get hardware status via the command line or via an X-Windows interface.
We also provide a free system health feature. See our example system health charts. At no charge to you, after you have installed the operating system and your server is about to enter production use, we will at your request install the above-mentioned monitoring software on your server, and will make ongoing system health graphs available to you. This is done at no charge the first time you make the request. (Subsequently, if you ask us to reinstall the monitoring software on your server or to re-initialize our charts, there is a $35 fee.)
(*): When installing mdk10.1 on the A3-DP server, disk partitioning must be done manually, and software RAID is not available. See “Why does it say “Not enough free space for auto-allocating?” in our FAQ document.
The following applies to all TQMbox servers listed above except where otherwise specified.
Performance numbers are provided for each server in the table above. These performance numbers are based on how long it takes to do a full compile of the Linux kernel and modules. This is a general-purpose benchmark that exercises the CPU, main memory disks, and associated caches. We run it as follows.
We do a fresh, default uninterruptible install of Fedora Core 2 from our menu. Without applying any further operating system updates, we run the following script:
#! /bin/sh
set -x
for n in 1 2 3 4; do
cd /usr/src/linux-2.6.8-1.521
make clean
make oldconfig
time make -j$n
echo :::: time above
done > log.bench 2>&1
This script rebuilds the kernel and modules with the commands “make -j1”, “make -j2”, “make -j3”, and “make -j3”. The -jN argument tells the make program to run N commands at a time. Thus “make -j4” will be done with 4 commands running at a time. Running more than one command at a time attempts to take advantage of parallelism between CPU processing and disk I/O (input/output), as well as any possible parallelism due to multiprocessing.
We then provide two performance numbers in our tables above. The first is always based on a build using the -j1 option. This represents the performance of the machine with no attempt to exploit any parallelism. The second is the best of the three numbers obtained using -j2, -j3, and -j4 and represents the performance gain, if any, observed due to parallelism.
We arbitrarily picked a total compile time of 39 minutes to correspond to a performance number of 1.00. All other performance numbers are relative to this.
Suppose the two performance numbers for a server are listed as “1.09(-j1)” and “1.27(-j2)”. This means that doing “make -j1” gave us the number 1.09; and doing “make -j2” gave us the performance number 1.27; and doing “make -j3” and “make -j4” did not give us a higher performance number than “make -j2” did.
Notice that on machines with a CPU that incorporates Intel’s hyperthreaded architecture, we do see a small but noticeable performance gain in our benchmark.
Please use our performance numbers only to compare our various servers among themselves. They are not meaningful if you attempt to apply them to any machine in any other environment.
Operating system installation. You may install/reinstall the operating system yourself from the boot-time menus, configuring the disk partioning yourself from our Server Configuration page. Only if you ask us to do this for you, our operating system install/reinstall fee is:
Note: The operating system install/reinstall listed above does not include the application of updates that were issued after the initial release of the operating system. To avoid security problems, you will need to promptly use yum, MandrakeUpdate, or other tool to bring the server up to date. We can do this for you—see below.
Update operating system. Bring operating system up to date to include all available officially-issued updates. You may do this yourself (see “How can I update the operating system?”). Only if you ask us to do this, our fee is $75.00 each time.
Lost root password recovery. You may reset a lost root password yourself, by using the boot-time menus to invoke the rescue mode for the operating system in use. If you ask us to do this for you, the fee to recover a lost root password depends on whether we can log into your server using our “a2iroot” facility (see “Why is there an ‘a2iroot’ account on my server?”).
Consulting. Consulting rate: $125 per hour. (For more information see FAQ.)
ECC. Error-Correcting Code. In its most common form, ECC memory is able to correct single-bit errors and detect double-bit errors occuring within a memory word. See also: “Why use ECC memory?.”
ATA, SATA. ATA (“AT attachment”) is an upward compatible disk interface standard that is a successor to IDE (“integrated drive electronics”). SATA (“serial ATA”) is a newer disk interface standard that is a successor to ATA.
RAM. Random access memory. As used in this document, this refers to main memory in a computer system.
SATA. See ATA, SATA.
Binary units. Memory sizes use binary prefixes, where binary k (also written kibi) means 1024, binary M or binary mega (also written mebi) means 1024*1024, and binary G (also written gibi) means 1024*1024*1024.
Decimal units. Disk capacities and CPU and bus speeds use decimal prefixes, where decimal M or decimal mega means 1000*1000, and decimal G or decimal giga means 1000*1000*1000.
IDE. See SATA.
Hyperthreading. A CPU design that allows a CPU to sometimes do two operations in parallel. See FAQ: “Why does my server have too many CPUs?”
To sign up for service, please see our Activity Links.
Contact us by telephone at +1 408-293-9706 or by email at support@rahul.net.
This document is part of a web site and should not be read in isolation. All information is subject to change without notice. Not everything described here applies equally to every possible combination of hardware and software. All products and services are provided on an “as is” basis; “with all faults” and “as available.” THERE ARE NO WARRANTIES, EXPRESSED OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THOSE OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. We are not responsible for the content of external web sites.