OVH minicloud: Hello world bench!

INTRODUCTION

The french hosting provider OVH is about to add some cloud related offers:

- minicloud: 1 small virtual machine, but very cheap 1,99euros a month

- coreCloud: 1 to 10 virtual machines instances. 9,99e/month

- myCloud: the most promising offer: 1 to 48 instances for 49,99e/month. You have your own cloud in which you can dynamically create virtual machines, distribute charge, etc..

WARNING: The OVH framework is yet not released. We are still waiting for more details to be provided.

Nothing more except a fixed price is yet provided about billing details.

MINICLOUD PRESENTATION

I had to chance to beta-test the minicloud offer. Basically you have one virtual machine, which only these caracteristics are described:

- OS: Debian 5.0 Lenny 64 bits
- RAM: 512 Mo
- CPU: 1 x64
- HDD: 5 Go

GOING FURTHER

Since we only have one virtual machine available, let us go further by discovering some details about it, and benchmarking it. However, keep in mind that since resources are shared within the cloud virtual machines, the following benchmarks actually depends of the cloud load.

# UNIX BENCH 5.1

========================================================================

BYTE UNIX Benchmarks (Version 5.1.2)

System: v12347.ovh.net: GNU/Linux

OS: GNU/Linux -- 2.6.32.2-xxxx-grs-ipv4-64 -- #1 SMP Tue Dec 29 14:41:12 UTC 2009

Machine: x86_64 (unknown)

Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")

CPU 0: Intel(R) Xeon(R) CPU E5504 @ 2.00GHz (3990.0 bogomips)

x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET

20:06:37 up 1 day, 19:57, 1 user, load average: 0.10, 0.03, 0.50; runlevel 2

------------------------------------------------------------------------

Benchmark Run: Sat Feb 20 2010 20:06:37 - 20:37:55

1 CPU in system; running 1 parallel copy of tests

Dhrystone 2 using register variables 7092797.6 lps (10.5 s, 7 samples)

Double-Precision Whetstone 2344.2 MWIPS (10.0 s, 7 samples)

Execl Throughput 1253.3 lps (30.0 s, 2 samples)

File Copy 1024 bufsize 2000 maxblocks 267298.5 KBps (30.0 s, 2 samples)

File Copy 256 bufsize 500 maxblocks 90045.7 KBps (31.0 s, 2 samples)

File Copy 4096 bufsize 8000 maxblocks 567304.8 KBps (31.0 s, 2 samples)

Pipe Throughput 770806.2 lps (10.8 s, 7 samples)

Pipe-based Context Switching 142454.6 lps (11.1 s, 7 samples)

Process Creation 3915.6 lps (30.3 s, 2 samples)

Shell Scripts (1 concurrent) 1834.1 lpm (60.0 s, 2 samples)

Shell Scripts (8 concurrent) 240.8 lpm (60.2 s, 2 samples)

System Call Overhead 1342105.8 lps (10.9 s, 7 samples)

System Benchmarks Index Values BASELINE RESULT INDEX

Dhrystone 2 using register variables 116700.0 7092797.6 607.8

Double-Precision Whetstone 55.0 2344.2 426.2

Execl Throughput 43.0 1253.3 291.5

File Copy 1024 bufsize 2000 maxblocks 3960.0 267298.5 675.0

File Copy 256 bufsize 500 maxblocks 1655.0 90045.7 544.1

File Copy 4096 bufsize 8000 maxblocks 5800.0 567304.8 978.1

Pipe Throughput 12440.0 770806.2 619.6

Pipe-based Context Switching 4000.0 142454.6 356.1

Process Creation 126.0 3915.6 310.8

Shell Scripts (1 concurrent) 42.4 1834.1 432.6

Shell Scripts (8 concurrent) 6.0 240.8 401.4

System Call Overhead 15000.0 1342105.8 894.7

========

System Benchmarks Index Score 507.1

Surprisingly, this is a pretty good score compared to more expensive cloud offers:

- Amazon: 210

- Slicehost: 295

- Rackspace: 305

- Linode x86_64: 559

- Linode i686: 723

Ref: VPS comparison

Then let us take some time to discover some features of an OVH minicloud offer:

#OS

root@v12347:~# uname -a

Linux v12347.ovh.net 2.6.32.2-xxxx-grs-ipv4-64 #1 SMP Tue Dec 29 14:41:12 UTC 2009 x86_64 GNU/Linux

#HDD

root@v12347:~# df

Filesystem 1K-blocks Used Available Use% Mounted on

/dev/sda1 5201532 673584 4265808 14% /

tmpfs 1026496 0 1026496 0% /lib/init/rw

udev 10240 2652 7588 26% /dev

tmpfs 1026496 0 1026496 0% /dev/shm

root@v12347:~# hdparm -t /dev/sda1

/dev/sda1:

Timing buffered disk reads: 126 MB in 3.03 seconds = 41.62 MB/sec

root@v12347:~# hdparm -T /dev/sda1

/dev/sda1:

Timing cached reads: 7326 MB in 2.00 seconds = 3664.46 MB/sec

#CPU

root@v12347:~# cat /proc/cpuinfo

processor : 0

vendor_id : GenuineIntel

cpu family : 6

model : 26

model name : Intel(R) Xeon(R) CPU E5504 @ 2.00GHz

stepping : 5

cpu MHz : 1995.001

cache size : 4096 KB

fpu : yes

fpu_exception : yes

cpuid level : 11

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 syscall nx rdtscp lm constant_tsc up arch_perfmon pebs bts rep_good xtopology tsc_reliable nonstop_tsc aperfmperf pni ssse3 cx16 sse4_1 sse4_2 popcnt hypervisor lahf_lm

bogomips : 3990.00

clflush size : 64

cache_alignment : 64

address sizes : 40 bits physical, 48 bits virtual

power management:

root@v12347:~# tiobench --size 384

Run #1: /usr/bin/tiotest -t 8 -f 48 -r 500 -b 4096 -d . -TTT

Unit information

================

File size = megabytes

Blk Size = bytes

Rate = megabytes per second

CPU% = percentage of CPU used during the test

Latency = milliseconds

Lat% = percent of requests that took longer than X seconds

CPU Eff = Rate divided by CPU% - throughput per cpu load

Sequential Reads

File Blk Num Avg Maximum Lat% Lat% CPU

Identifier Size Size Thr Rate (CPU%) Latency Latency >2s >10s Eff

---------------------------- ------ ----- --- ------ ------ --------- ----------- -------- -------- -----

2.6.32.2-xxxx-grs-ipv4-64 384 4096 1 ###### 102.2% 0.001 0.14 0.00000 0.00000 2909

2.6.32.2-xxxx-grs-ipv4-64 384 4096 2 ###### 194.6% 0.002 8.38 0.00000 0.00000 1574

2.6.32.2-xxxx-grs-ipv4-64 384 4096 4 ###### 385.0% 0.005 20.04 0.00000 0.00000 727

2.6.32.2-xxxx-grs-ipv4-64 384 4096 8 ###### 413.6% 0.008 32.05 0.00000 0.00000 774

Random Reads

File Blk Num Avg Maximum Lat% Lat% CPU

Identifier Size Size Thr Rate (CPU%) Latency Latency >2s >10s Eff

---------------------------- ------ ----- --- ------ ------ --------- ----------- -------- -------- -----

2.6.32.2-xxxx-grs-ipv4-64 384 4096 1 ###### 129.6% 0.001 0.03 0.00000 0.00000 1953

2.6.32.2-xxxx-grs-ipv4-64 384 4096 2 ###### 67.56% 0.001 0.04 0.00000 0.00000 3906

2.6.32.2-xxxx-grs-ipv4-64 384 4096 4 ###### 123.8% 0.002 3.41 0.00000 0.00000 1953

2.6.32.2-xxxx-grs-ipv4-64 384 4096 8 ###### 56.57% 0.001 0.04 0.00000 0.00000 3906

Sequential Writes

File Blk Num Avg Maximum Lat% Lat% CPU

Identifier Size Size Thr Rate (CPU%) Latency Latency >2s >10s Eff

---------------------------- ------ ----- --- ------ ------ --------- ----------- -------- -------- -----

2.6.32.2-xxxx-grs-ipv4-64 384 4096 1 39.91 11.80% 0.021 421.55 0.00000 0.00000 338

2.6.32.2-xxxx-grs-ipv4-64 384 4096 2 30.80 29.35% 0.061 2149.02 0.00203 0.00000 105

2.6.32.2-xxxx-grs-ipv4-64 384 4096 4 29.44 52.49% 0.112 2117.31 0.00407 0.00000 56

2.6.32.2-xxxx-grs-ipv4-64 384 4096 8 34.61 50.22% 0.163 5115.43 0.00203 0.00000 69

Random Writes

File Blk Num Avg Maximum Lat% Lat% CPU

Identifier Size Size Thr Rate (CPU%) Latency Latency >2s >10s Eff

---------------------------- ------ ----- --- ------ ------ --------- ----------- -------- -------- -----

2.6.32.2-xxxx-grs-ipv4-64 384 4096 1 16.75 2.572% 0.005 1.64 0.00000 0.00000 651

2.6.32.2-xxxx-grs-ipv4-64 384 4096 2 44.54 9.122% 0.005 4.06 0.00000 0.00000 488

2.6.32.2-xxxx-grs-ipv4-64 384 4096 4 40.43 18.63% 0.007 10.73 0.00000 0.00000 217

2.6.32.2-xxxx-grs-ipv4-64 384 4096 8 18.86 -13.5% 0.003 0.63 0.00000 0.00000 -140

# RAM

Not tested.

# NETWORK (not really representative)

# using iperf from a RPS limited to 100Mbit/s

[ 3] 0.0-10.0 sec 113 MBytes 94.7 Mbits/sec

# ideally I would have to rent another minicloud to check the effective performance (because I guess it is limited by my RPS). I guess result would be closer to 10Gb/s or 1Gb/s