96MB Low End VPS Review Part XV – Secure Dragon
For those low end VPS fans, 15 seems to be a magic price point for many of the ultra-low-end VPs providers. Many of the companies have tried to shoot at this price point and almost every one of them failed. Most of them either oversold the resources so much that makes the box completely unusable, or the price points becomes so unsustainable that eventually the provider just collect cash and run.
Therefore, when Secure Dragon (which has nothing to do with Chinese dragon, I believe) offered a 96MB guaranteed memory low end VPS on LowEndBox.com, and disclaimed “No Overselling!” in bold on their home page, I was instantly curious to see how the performance of such a ultra-low-end VPS would be like.
General Information and Sign Up
As per the post on LowEndBox, here is the specifications of this low end VPS:
On their product homepage, there are actually more information about this package:
As you can see, the VPS provision was advertised as instant activation. However, what is more notable about them is the fact that they offer Money Back Guarantee for as long as one month!
Compare to most of the VPS providers, even the big names like BuyVM, which has clearly stated that they do not offer refunds, it is indeed a pretty bold move that Secure Dragon is taking.
It is also interesting to note that they have a live counter on the estimated time of ticket response on their website, which is something I have never seen before:
Although they do not guarantee response times, I have been checking this page at different times during the day and I have yet to see any response time for above an hour. Furthermore, unlike most of the ultra-low-end VPS providers, who only offer unmanaged services, Secure Dragon actually offer semi-managed services, which means they perhaps have more work to do than the unmanaged providers.
It is also interesting to note that their entire website is on HTTPs connection, which is really “secure” indeed.
Set up on the server is indeed instant, my Paypal confirmation comes in at 9:43pm and the new VPS information arrived within the same minute. It is good to note that all the password in the email is replaced by *OMITTED*.
The SolusVM address in the email is not on the standard HTTP/HTTPS port, however this announcement has indicated that they now have both secured and unsecured SolusVM access on standard port, which is something great.
After signing into the WHMCS control panel, the VPS control is, as with many other providers, under the services page:
As usual, the basic functionalities to control the VPS are available, and there are also bar graph showing the percentage of resources usage, however the actual resources time series graph are disabled to save on memory usage.
SolusVM shows a pretty standard screen after login using the user name that was assigned to you, which, on a side note, is a common practice by the VPS providers which I dislike. Basically, BuyVM uses your email address as the user names for SolusVM and Quickweb allow you to use anything after submitting a ticket to them, which are both great, I mean, at the end of the day, who is going to remember XYZ648 is my VPS A SolusVM login and ABC359 is my VPS B SolusVM login?
As with many other providers, rDNS is not instant and both built-in backup features are not working.
They have, however, a decent collections of OS templates, and instead of focusing on providing many distributions, they actually chose to focus on providing some ready-to-go packages:
The most interesting one in the list is the Debian minimal template, which according to the description, takes only 4MB of RAM and that is what I used for testing in this particular VPS.
Test on the Box
The VPS I have received, as indicated above, has 96MB of dedicated RAM (96MB burstable), 1 CPU core and 5GB disk drive space and according to this post, they own their hardware in GoRack data centre in Jacksonville, Florida.
Without installing any software after reloading the Debian minimal template, here is the RAM usage:
free -m
total used free shared buffers cached
Mem: 192 38 153 0 0 0
-/+ buffers/cache: 38 153
Swap: 0 0 0
As you can see, it is definitely way more than the 4MB RAM that was advertised. In fact, as shown in the top output below, SSH daemon + bash shell alone used more than 4MB of memory:
top - 21:48:04 up 4 min, 1 user, load average: 0.00, 0.02, 0.00
Tasks: 7 total, 2 running, 5 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 196608k total, 39912k used, 156696k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1387 root 16 0 8020 2708 2224 R 0.0 1.4 0:00.01 sshd
1390 root 15 0 2764 1484 1180 S 0.0 0.8 0:00.00 bash
1322 root 18 0 33136 1224 936 S 0.0 0.6 0:00.00 rsyslogd
1394 root 15 0 2260 1068 880 R 0.0 0.5 0:00.00 top
1331 root 18 0 5276 1032 680 S 0.0 0.5 0:00.00 sshd
1349 root 18 0 2040 788 636 S 0.0 0.4 0:00.00 cron
1 root 18 0 1984 688 588 S 0.0 0.3 0:02.12 init
However, I have to admit that they have definitely put in considerable effort to bring down the memory usage and this is definitely one of the shortest top output that I have ever observed.
The magic actually happened after I cleaned up the system a little bit using the famous Low End Script on LEB:
free -m
total used free shared buffers cached
Mem: 192 3 188 0 0 0
-/+ buffers/cache: 3 188
Swap: 0 0 0
Now we are seriously talking about 4MB of RAM usage, which is confirmed again by the top output:
top - 21:54:09 up 0 min, 1 user, load average: 0.16, 0.06, 0.02
Tasks: 7 total, 2 running, 5 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 196608k total, 4188k used, 192420k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1369 root 15 0 2764 1484 1180 S 0.0 0.8 0:00.00 bash
1368 root 15 0 2356 1240 828 R 0.0 0.6 0:00.08 dropbear
1373 root 15 0 2260 1068 880 R 0.0 0.5 0:00.00 top
1356 root 18 0 2040 788 636 S 0.0 0.4 0:00.00 cron
1 root 15 0 1984 688 588 S 0.0 0.3 0:02.12 init
1330 root 18 0 1884 680 564 S 0.0 0.3 0:00.00 syslogd
1338 root 18 0 2044 464 364 S 0.0 0.2 0:00.00 dropbear
Even with the full LNMP stack installed, the total memory usage is still about 60MB, which is really good:
free -m
total used free shared buffers cached
Mem: 192 59 132 0 0 0
-/+ buffers/cache: 59 132
Swap: 0 0 0
And the corresponding top output:
top - 23:19:45 up 13:24, 1 user, load average: 0.00, 0.00, 0.00
Tasks: 14 total, 2 running, 12 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.5%us, 0.5%sy, 0.0%ni, 98.9%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 196608k total, 61404k used, 135204k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
16169 www 15 0 15380 11m 852 S 0.0 6.0 0:19.84 nginx
11966 www 16 0 23968 6860 3408 S 0.6 3.5 1:13.15 php-fpm
11967 www 16 0 23968 6648 3408 S 0.0 3.4 1:17.99 php-fpm
11356 mysql 18 0 37048 5924 3220 S 0.0 3.0 0:00.00 mysqld
11965 root 15 0 23680 4080 1180 S 0.0 2.1 0:00.07 php-fpm
16167 root 19 0 4792 2196 1744 S 0.0 1.1 0:00.00 nginx
1368 root 15 0 2768 1492 1180 S 0.0 0.8 0:00.00 bash
1367 root 16 0 2464 1304 832 R 0.0 0.7 0:03.69 dropbear
17744 root 15 0 2328 1096 900 R 0.0 0.6 0:00.00 top
17554 root 15 0 2288 764 600 S 0.0 0.4 0:00.00 cron
1329 root 15 0 1884 708 588 S 0.0 0.4 0:00.01 syslogd
1 root 15 0 2028 688 600 S 0.0 0.3 0:03.07 init
10230 root 25 0 1744 532 448 S 0.0 0.3 0:00.00 mysqld_safe
1337 root 18 0 2044 468 372 S 0.0 0.2 0:00.00 dropbear
Vzfree shows no signs of overselling:
vzfree
Total Used Free
Kernel: 2048.00M 2.53M 2045.47M
Allocate: 192.00M 106.16M 85.84M (96M Guaranteed)
Commit: 96.00M 85.19M 10.81M (77.9% of Allocated)
Swap: 0.00M (0.0% of Committed)
The Inodes is set to pretty high limits, which is something great:
df -i Filesystem Inodes IUsed IFree IUse% Mounted on tmpfs 24576 3 24573 1% /lib/init/rw tmpfs 24576 1 24575 1% /dev/shm /dev/simfs 2621440 79931 2541509 4% /
There is not trace of disk space overselling as well:
df -h Filesystem Size Used Avail Use% Mounted on tmpfs 96M 0 96M 0% /lib/init/rw tmpfs 96M 0 96M 0% /dev/shm /dev/simfs 5.0G 1.9G 3.2G 38% /
vmstat shows the node is pretty lightly used:
vmstat procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu---- r b swpd free buff cache si so bi bo in cs us sy id wa 0 0 0 88224 0 0 0 0 0 29 0 181 2 0 98 0
which is again, proved by the output of uptime:
uptime 21:21:25 up 1 day, 11:26, 1 user, load average: 0.01, 0.02, 0.00
Although some of the beancounters values are not that high, we have to keep in mind that this box offers only 96MB of guaranteed memory:
cat /proc/user_beancounters
Version: 2.5
uid resource held maxheld barrier limit failcnt
226: kmemsize 2696742 5057456 2147483646 2147483646 0
lockedpages 0 411 999999 999999 0
privvmpages 27177 46705 49152 49152 0
shmpages 130 656 24576 24576 0
dummy 0 0 0 0 0
numproc 16 28 999999 999999 0
physpages 21159 32101 0 2147483647 0
vmguarpages 0 0 24576 2147483647 0
oomguarpages 21159 32101 24576 2147483647 0
numtcpsock 4 12 7999992 7999992 0
numflock 1 7 999999 999999 0
numpty 1 2 500000 500000 0
numsiginfo 0 6 999999 999999 0
tcpsndbuf 72344 398496 214748160 396774400 0
tcprcvbuf 65536 3502832 214748160 396774400 0
othersockbuf 9312 4275728 214748160 396774400 0
dgramrcvbuf 0 8752 214748160 396774400 0
numothersock 8 34 7999992 7999992 0
dcachesize 212214 334284 2147483646 2147483646 0
numfile 1043 1539 23999976 23999976 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
dummy 0 0 0 0 0
numiptent 14 14 999999 999999 0
The disk I/O is actually pretty good for this small box, and it is consistently above 50MB/s:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 18.9565 s, 56.6 MB/s
The results are pretty consistent:
dd if=/dev/zero of=test bs=64k count=16k conv=fdatasync 16384+0 records in 16384+0 records out 1073741824 bytes (1.1 GB) copied, 18.5315 s, 57.9 MB/s
It is interesting to note that in the description of the product, each user get half of a CPU core, I was wondering how it was done initially, but when I checked out the cpuinfo, it all make sense:
cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 15 model name : Intel(R) Xeon(R) CPU 5150 @ 2.66GHz stepping : 6 cpu MHz : 1329.997 cache size : 4096 KB physical id : 0 siblings : 2 core id : 0 cpu cores : 2 apicid : 0 fpu : yes fpu_exception : yes cpuid level : 10 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm syscall nx lm constant_tsc pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr lahf_lm bogomips : 5319.99 clflush size : 64 cache_alignment : 64 address sizes : 36 bits physical, 48 bits virtual power management:
Basically, the CPU is a Xeon 2.66GHz CPU that is capped at 1.33GHz, which is half of the clock speed.
Meminfo again does not show too many interesting bits:
cat /proc/meminfo MemTotal: 196608 kB MemFree: 88296 kB Buffers: 0 kB Cached: 0 kB SwapCached: 0 kB Active: 0 kB Inactive: 0 kB HighTotal: 0 kB HighFree: 0 kB LowTotal: 196608 kB LowFree: 88296 kB SwapTotal: 0 kB SwapFree: 0 kB Dirty: 98020 kB Writeback: 0 kB AnonPages: 0 kB Mapped: 0 kB Slab: 0 kB PageTables: 0 kB NFS_Unstable: 0 kB Bounce: 0 kB CommitLimit: 0 kB Committed_AS: 0 kB VmallocTotal: 0 kB VmallocUsed: 0 kB VmallocChunk: 0 kB HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 Hugepagesize: 2048 kB
One interesting thing I have noticed is their time sync output, which seems to be a lot higher than the time sync output I have seen, which I guess (and please correct me if I am wrong) primarily due to the lack of physical power of the box:
time sync real 0m0.323s user 0m0.000s sys 0m0.037s
The download speed is pretty good especially considering it is a 100Mbps port:
wget cachefly.cachefly.net/100mb.test --2011-07-24 21:23:10-- http://cachefly.cachefly.net/100mb.test Resolving cachefly.cachefly.net... 205.234.175.175 Connecting to cachefly.cachefly.net|205.234.175.175|:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 13.6M/s in 14s 2011-07-24 21:23:24 (7.11 MB/s) - `100mb.test' saved [104857600/104857600]
As you can see, I am getting pretty much close to the line speed.
In terms of the upload speed, BuyVM is the first VPS that I tested the upload, and the speed is not all that great:
wget 8.22.203.164/100mb.test --2011-07-29 09:15:46-- http://8.22.203.164/100mb.test Connecting to 8.22.203.164:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 4.45M/s in 29s 2011-07-29 09:16:15 (3.42 MB/s) - `100mb.test' saved [104857600/104857600]
Testing with a Nix Communications VPS from Montreal, Canada, however, the upload speed from the Secure Dragon VPS is a lot better:
wget http://8.22.203.164/100mb.test --2011-07-29 05:31:05-- http://8.22.203.164/100mb.test Connecting to 8.22.203.164:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 10.5M/s in 9.8s 2011-07-29 05:31:15 (10.2 MB/s) - `100mb.test' saved [104857600/104857600]
And the QuickWeb VPS in London, UK, gives a pretty slow speed too:
wget http://8.22.203.164/100mb.test --2011-07-28 22:13:46-- http://8.22.203.164/100mb.test Connecting to 8.22.203.164:80... connected. HTTP request sent, awaiting response... 200 OK Length: 104857600 (100M) [application/octet-stream] Saving to: `100mb.test' 100%[======================================>] 104,857,600 2.34M/s in 39s 2011-07-28 22:14:25 (2.58 MB/s) - `100mb.test' saved [104857600/104857600]
Therefore, it seems that the download speed is only good for east coast locations and is slow for the west coast and European locations.
Finally, given the fact that the VPS has only 96MB of dedicated RAM, I was pretty surprised that it actually managed to completed the UnixBench tests, although obviously the score is pretty low:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: sd: GNU/Linux
OS: GNU/Linux -- 2.6.18-238.12.1.el5.028stab091.1 -- #1 SMP Wed Jun 1 13:20:25 MSD 2011
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: Intel(R) Xeon(R) CPU 5150 @ 2.66GHz (5320.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
21:29:37 up 1 day, 11:34, 1 user, load average: 0.15, 0.08, 0.02; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Sun Jul 24 2011 21:29:37 - 22:00:38
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 7026307.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2258.9 MWIPS (10.1 s, 7 samples)
Execl Throughput 1846.0 lps (29.3 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 177335.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 55181.5 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 343895.9 KBps (30.0 s, 2 samples)
Pipe Throughput 396087.6 lps (10.0 s, 7 samples)
Pipe-based Context Switching 119821.6 lps (10.0 s, 7 samples)
Process Creation 5466.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3760.9 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 505.0 lpm (60.1 s, 2 samples)
System Call Overhead 317154.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 7026307.1 602.1
Double-Precision Whetstone 55.0 2258.9 410.7
Execl Throughput 43.0 1846.0 429.3
File Copy 1024 bufsize 2000 maxblocks 3960.0 177335.9 447.8
File Copy 256 bufsize 500 maxblocks 1655.0 55181.5 333.4
File Copy 4096 bufsize 8000 maxblocks 5800.0 343895.9 592.9
Pipe Throughput 12440.0 396087.6 318.4
Pipe-based Context Switching 4000.0 119821.6 299.6
Process Creation 126.0 5466.6 433.9
Shell Scripts (1 concurrent) 42.4 3760.9 887.0
Shell Scripts (8 concurrent) 6.0 505.0 841.6
System Call Overhead 15000.0 317154.5 211.4
========
System Benchmarks Index Score 445.7
Second test produced similar results:
# # # # # # # ##### ###### # # #### # #
# # ## # # # # # # # ## # # # # #
# # # # # # ## ##### ##### # # # # ######
# # # # # # ## # # # # # # # # #
# # # ## # # # # # # # ## # # # #
#### # # # # # ##### ###### # # #### # #
Version 5.1.3 Based on the Byte Magazine Unix Benchmark
Multi-CPU version Version 5 revisions by Ian Smith,
Sunnyvale, CA, USA
January 13, 2011 johantheghost at yahoo period com
1 x Dhrystone 2 using register variables 1 2 3 4 5 6 7 8 9 10
1 x Double-Precision Whetstone 1 2 3 4 5 6 7 8 9 10
1 x Execl Throughput 1 2 3
1 x File Copy 1024 bufsize 2000 maxblocks 1 2 3
1 x File Copy 256 bufsize 500 maxblocks 1 2 3
1 x File Copy 4096 bufsize 8000 maxblocks 1 2 3
1 x Pipe Throughput 1 2 3 4 5 6 7 8 9 10
1 x Pipe-based Context Switching 1 2 3 4 5 6 7 8 9 10
1 x Process Creation 1 2 3
1 x System Call Overhead 1 2 3 4 5 6 7 8 9 10
1 x Shell Scripts (1 concurrent) 1 2 3
1 x Shell Scripts (8 concurrent) 1 2 3
========================================================================
BYTE UNIX Benchmarks (Version 5.1.3)
System: sd: GNU/Linux
OS: GNU/Linux -- 2.6.18-238.12.1.el5.028stab091.1 -- #1 SMP Wed Jun 1 13:20:25 MSD 2011
Machine: i686 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: Intel(R) Xeon(R) CPU 5150 @ 2.66GHz (5320.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
22:11:22 up 1 day, 12:15, 1 user, load average: 0.04, 0.36, 0.84; runlevel 2
------------------------------------------------------------------------
Benchmark Run: Sun Jul 24 2011 22:11:22 - 22:42:15
1 CPU in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 6808840.7 lps (10.0 s, 7 samples)
Double-Precision Whetstone 2212.9 MWIPS (10.1 s, 7 samples)
Execl Throughput 1829.2 lps (29.6 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 171195.7 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 57220.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 330493.8 KBps (30.0 s, 2 samples)
Pipe Throughput 395309.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 120682.6 lps (10.0 s, 7 samples)
Process Creation 5305.6 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 3710.3 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 485.2 lpm (60.1 s, 2 samples)
System Call Overhead 315396.4 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 6808840.7 583.4
Double-Precision Whetstone 55.0 2212.9 402.3
Execl Throughput 43.0 1829.2 425.4
File Copy 1024 bufsize 2000 maxblocks 3960.0 171195.7 432.3
File Copy 256 bufsize 500 maxblocks 1655.0 57220.3 345.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 330493.8 569.8
Pipe Throughput 12440.0 395309.8 317.8
Pipe-based Context Switching 4000.0 120682.6 301.7
Process Creation 126.0 5305.6 421.1
Shell Scripts (1 concurrent) 42.4 3710.3 875.1
Shell Scripts (8 concurrent) 6.0 485.2 808.6
System Call Overhead 15000.0 315396.4 210.3
========
System Benchmarks Index Score 438.9
And finally, the GeekBench tests produced a resonable score considering the limited amount of resources available:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 2.6.18-238.12.1.el5.028stab091.1 i686
Model: Linux PC (Intel Xeon 5150)
Motherboard: Unknown Motherboard
Processor: Intel Xeon 5150
Processor ID: GenuineIntel Family 6 Model 15 Stepping 6
Logical Processors: 1
Physical Processors: 1
Processor Frequency: 1.33 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 4.00 MB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 15.6 GB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel Xeon 5150
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 1894 |||||||
multi-threaded scalar 1021 ||||
Text Compress
single-threaded scalar 1079 ||||
multi-threaded scalar 1023 ||||
Text Decompress
single-threaded scalar 949 |||
multi-threaded scalar 981 |||
Image Compress
single-threaded scalar 916 |||
multi-threaded scalar 898 |||
Image Decompress
single-threaded scalar 844 |||
multi-threaded scalar 862 |||
Lua
single-threaded scalar 1743 ||||||
multi-threaded scalar 1744 ||||||
Floating Point
Mandelbrot
single-threaded scalar 1008 ||||
multi-threaded scalar 1020 ||||
Dot Product
single-threaded scalar 2003 ||||||||
multi-threaded scalar 1971 |||||||
single-threaded vector 1345 |||||
multi-threaded vector 1546 ||||||
LU Decomposition
single-threaded scalar 2207 ||||||||
multi-threaded scalar 2254 |||||||||
Primality Test
single-threaded scalar 1559 ||||||
multi-threaded scalar 1253 |||||
Sharpen Image
single-threaded scalar 3193 ||||||||||||
multi-threaded scalar 3217 ||||||||||||
Blur Image
single-threaded scalar 2423 |||||||||
multi-threaded scalar 2375 |||||||||
Memory
Read Sequential
single-threaded scalar 1402 |||||
Write Sequential
single-threaded scalar 1944 |||||||
Stdlib Allocate
single-threaded scalar 1075 ||||
Stdlib Write
single-threaded scalar 666 ||
Stdlib Copy
single-threaded scalar 990 |||
Stream
Stream Copy
single-threaded scalar 1851 |||||||
single-threaded vector 2038 ||||||||
Stream Scale
single-threaded scalar 1912 |||||||
single-threaded vector 1961 |||||||
Stream Add
single-threaded scalar 1800 |||||||
single-threaded vector 2046 ||||||||
Stream Triad
single-threaded scalar 1951 |||||||
single-threaded vector 1520 ||||||
Integer Score: 1162 ||||
Floating Point Score: 1955 |||||||
Memory Score: 1215 ||||
Stream Score: 1884 |||||||
Overall Geekbench Score: 1522 ||||||
And again, the second time it showed a similar GeekBench score:
System Information
Platform: Linux x86 (32-bit)
Compiler: GCC 4.1.2 20070925 (Red Hat 4.1.2-33)
Operating System: Linux 2.6.18-238.12.1.el5.028stab091.1 i686
Model: Linux PC (Intel Xeon 5150)
Motherboard: Unknown Motherboard
Processor: Intel Xeon 5150
Processor ID: GenuineIntel Family 6 Model 15 Stepping 6
Logical Processors: 1
Physical Processors: 1
Processor Frequency: 1.33 GHz
L1 Instruction Cache: 0.00 B
L1 Data Cache: 0.00 B
L2 Cache: 4.00 MB
L3 Cache: 0.00 B
Bus Frequency: 0.00 Hz
Memory: 15.6 GB
Memory Type: N/A
SIMD: 1
BIOS: N/A
Processor Model: Intel Xeon 5150
Processor Cores: 1
Integer
Blowfish
single-threaded scalar 1907 |||||||
multi-threaded scalar 1022 ||||
Text Compress
single-threaded scalar 1088 ||||
multi-threaded scalar 1027 ||||
Text Decompress
single-threaded scalar 951 |||
multi-threaded scalar 975 |||
Image Compress
single-threaded scalar 917 |||
multi-threaded scalar 898 |||
Image Decompress
single-threaded scalar 859 |||
multi-threaded scalar 862 |||
Lua
single-threaded scalar 1756 |||||||
multi-threaded scalar 1742 ||||||
Floating Point
Mandelbrot
single-threaded scalar 1108 ||||
multi-threaded scalar 1041 ||||
Dot Product
single-threaded scalar 2014 ||||||||
multi-threaded scalar 1977 |||||||
single-threaded vector 1363 |||||
multi-threaded vector 1547 ||||||
LU Decomposition
single-threaded scalar 2211 ||||||||
multi-threaded scalar 2249 ||||||||
Primality Test
single-threaded scalar 1559 ||||||
multi-threaded scalar 1240 ||||
Sharpen Image
single-threaded scalar 3400 |||||||||||||
multi-threaded scalar 3207 ||||||||||||
Blur Image
single-threaded scalar 2444 |||||||||
multi-threaded scalar 2406 |||||||||
Memory
Read Sequential
single-threaded scalar 1394 |||||
Write Sequential
single-threaded scalar 1536 ||||||
Stdlib Allocate
single-threaded scalar 1094 ||||
Stdlib Write
single-threaded scalar 531 ||
Stdlib Copy
single-threaded scalar 1312 |||||
Stream
Stream Copy
single-threaded scalar 1857 |||||||
single-threaded vector 2008 ||||||||
Stream Scale
single-threaded scalar 1983 |||||||
single-threaded vector 2063 ||||||||
Stream Add
single-threaded scalar 1842 |||||||
single-threaded vector 2098 ||||||||
Stream Triad
single-threaded scalar 2055 ||||||||
single-threaded vector 1558 ||||||
Integer Score: 1167 ||||
Floating Point Score: 1983 |||||||
Memory Score: 1173 ||||
Stream Score: 1933 |||||||
Overall Geekbench Score: 1530 ||||||
Overall, the box has no sign of overselling, and besides download speed, it is generally pretty solid given the specifications.
Customer Service and Support
Although the VPS is advertised as semi-managed, what is included in semi-management is pretty vague and unlike Quality Servers, who has a pretty clear boundary on what is managed and what is not, Secure Dragon does not offer that kind of clarity on their website.
Having said that, however, their response time is pretty good, my first ticket with them is replied in 6 minutes (ticket submitted at 10:55AM, replied in 11:01AM) and the second follow-up ticket took 10 minutes (I replied in 11:24AM, ticket was responded at 11:34AM).
Furthermore, Secure Dragon actually put their estimated ticket response time on their website and as mentioned above, from what I have observed, the that time has rarely been more than an hour.
Conclusion
It is rare to see the ultra-low-end VPS providers who could provide a good VPS that is not heavily oversold and the performance are pretty solid with respect to the resources available, and I think Secure Dragon is definitely one of the success stories here. Fast response time, 30 days money back guarantee, and many other things make this deal a really sweet one. Besides the upload speed may not be that satisfactory, it is in general a pretty good ultra-low-end VPS, but once again, with respect to the resources available. For those who would like to sign up for one of these boxes, you have to realize that the resources you are assigned to is severely limited and hence you may not able to run everything you would like to run on it without very heavy optimization. Furthermore, although at the moment I do not think Secure Dragon is overselling any resources, posts like this one definitely make them to my watch list.
Very thorough review! I would like to point out a few things and hopefully clear up some confusion we might have caused you:
1) Our default Debian 5 install is now sitting at 4MB of RAM upon install. I don’t know why it jumped to 38MB but this was fixed when it was brought to our attention by one of our clients on the 21st.
# free -m
total used free shared buffers cached
Mem: 192 4 187 0 0 0
-/+ buffers/cache: 4 187
Swap: 0 0 0
2) About that overselling thread, that was created in response to numerous potential clients requesting that we offer them more resources than we had available and more resources than they are paying for “because it was OpenVZ”, i.e. overselling. After speaking with some of our clients on here, my partner, and the other users on WHT it was decided that we will stick to our “No Overselling EVAR!” policy. We decided that the type of users that are attracted to 5TB of bandwidth for a nickle isn’t the type of clients we want to attract because of the amount of time we spend in the help desk already.
3) Our semi-managed service is pretty extensive actually, it’s not supposed to be but we save the “we’re not fully managed” excuse as a last resort. So far we’ve handled a lot of tickets that some might consider out of the scope of semi-managed but the way we see it is if we know how to fix it and we have the ability to, then we’ll do it. We have the option for an additional $3/month you can have fully managed, but to be honest we try to treat all clients like that and short of making any changes to our nodes that might impact other users, we do our best. To give you an idea of some of the “semi-managed” support we’ve provided, here’s a list of tutorials we wrote specifically because these are some common tasks we were doing for clients: https://securedragon.net/knowledgebase.php?action=displaycat&catid=3
@Kujoe: Good to see the Debian template is fixed, I have just reloaded the OS template and I see here:
free -m
total used free shared buffers cached
Mem: 192 9 182 0 0 0
-/+ buffers/cache: 9 182
Swap: 0 0 0
which I definitely think it is very good, and here is the corresponding output from top:
top – 10:37:06 up 3 min, 1 user, load average: 0.00, 0.00, 0.00
Tasks: 7 total, 2 running, 5 sleeping, 0 stopped, 0 zombie
Cpu(s): 0.0%us, 0.0%sy, 0.0%ni,100.0%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 196608k total, 9692k used, 186916k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1354 root 16 0 8212 2808 2220 R 0.0 1.4 0:00.02 sshd
1356 root 15 0 2764 1488 1180 S 0.0 0.8 0:00.00 bash
1361 root 15 0 2260 1060 876 R 0.0 0.5 0:00.00 top
1322 root 18 0 5276 1032 672 S 0.0 0.5 0:00.00 sshd
1340 root 18 0 2040 788 636 S 0.0 0.4 0:00.00 cron
1316 root 15 0 2852 740 548 S 0.0 0.4 0:00.00 syslog-ng
1 root 15 0 1984 688 588 S 0.0 0.3 0:03.17 init
It is also assuring to know that you guys have decided not to oversell the resources, I think it is a very good move.
Thanks for the clarification!
@KuJoe That’s a very good decision (point 2), respect 🙂 I also don’t want to share my node with bandwidth rapist kind of users 😀
looking forward for your xen offer 😛
@circus: I think your wish has been granted! They have a Xen 96MB plan at the moment and if you are already with them on their OpenVZ plan, they even offer you a discount with the discount code: XEN4US, which makes you pay 16 bucks per year!
Impressed by your post..
Very Describe and even you post every test,etc 🙂
Awesomeeeeeeeeeeeee!!!!! work
Excellent Review.
I signed up with these guys and wow! Awesome service. The coupon did not work at first so I emailed them, they responded in no time flat and fixed it. I signed up and had access to the vps in a few minutes.
I then emailed them to say thank you and they replied in a few minutes.
@Luma: Did you sign up with their Xen or OVZ product line? I have just got another of their Xen box to play around, wondering how will that go, I’ll post a review once I get to it 🙂
I went with the OpenVZ because that is all I need and it runs great.
Good luck with your Xen VM 🙂
I am using both, XEN PV and OpenVZ and both rock! although I was expecting to see KVM from this provider too but I think their plans for KVM are a few months away or something like that.
Overall, very decent support, very nice VPS and you are paying dirt cheap for an excellent service 🙂
@Asim: I have just done some testing today on the box, I agree it seems pretty solid, my only complain is that it does not have a 32bit OS, for some reason people really hate putting up 32 bit OS on Xen PV 🙁 OpenITC does not do that, Secure Dragon does not do that as well 🙁
Last time when I asked that Joe explained that he purchased these templates from another company and they advised not to use 32bit on a 64bit system. I feel that 32bits IS the way to do it on small boxes.
I have seen this company make positive progress and I hope it will grow and we can proudly say one day that YES we are one of the first customers of this company 🙂
@Asim: Glad to know, I think 64 bit is just wasting memory out of nothing, it will do better for those monster systems (with >=3GB memory I think), but for the small low end VPS, it does more bad than good. Unfortunately I think the only one that I can find some 32 bit OS template for Xen is Quickweb, Up2VPS used to have some 32 bit templates at one point (although for some reason the kernel says 64 bit, but I tried to install 64 bit programs on it and it asked me to install the 32 bit versions instead).
Anyways, I do agree with you, so far, my experience with Secure Dragon is great, and I do hope they can keep things like this, making them a potential candidate to move 96MB to as well.
Just wanted to let you know that we do have 32bit OS templates for our Xen PV servers now. I have not tested them extensively (simple tests for network connectivity and such) but a lot of clients are running them and have reported no issues (as in they actually opened tickets to let us know they haven’t had any problems).
I’m glad to read good reviews like this and love the comments, it lets us know that we’re doing it right and will work hard to keep it up. 🙂
@KuJoe: Am I allowed to be more greedy and ask for a minimal template that is similar to the one you have in OVZ?
Never mind, just realised it is the famous minimal template, thanks KuJoe!
To be honest the templates on Xen are not customized at all and are directly from Stacklet. I’m not familiar with building Xen templates so I’m forced to purchase them for now. 🙁
That being said I’ll see if I can trim it down more and type up a tutorial or even write a script for our clients. 😉
Hi,
this becuse OS was installed with default kernel so it shows 64bit but we have new OS using PyGrub so when you are rebuilt to 32bit kernel will be shows 32bit not 64bit . please not . it’s not for all version or all OS only which use PyGrub
like centos 6 etc
Thanks You
UP2VPS
Pingback: 96MB Low End VPS Review Part XVII – Secure Dragon Xen VPS | | 96MB.com96MB.com
Pingback: SecureDragon – $18/Year 128MB OpenVZ VPS in Jacksonville – Low End Box