Date post: | 01-Nov-2014 |
Category: |
Technology |
Upload: | china-openstack-user-group |
View: | 867 times |
Download: | 0 times |
Introduction
XSmall SmallMedium
Large XLargeXXLarge
vCPU 1 2 2 4 4 8
MEM (GB) 1 2 4 8 16 32
DISK (GB) 30 60 120 240 480 960
Price ($/hr) 0.04 0.08 0.16 0.32 0.64 1.28
• Virtual Machines• az-1.region-a.geo-1• web-created• Ubuntu 11.04 64 bit• 3+ VM’s / model• total 20 VM’s
• Benchmark Suite• byte-unixbench• mbw• iozone• iperf• pgbench• Hadoop wordcount
• Data Filtering• best VM / model • average by 10
byte-unixbench
0
500
1000
1500
2000
2500
3000
3500
4000
4500
XSmal l Smal l Medi um Large XLarge XXLarge
Si ngl e ThreadMul t i Thread
• byte-unixbench index measures overall system performance• in multi-thread testing, n-Thread = n-vCPU• system with the same amount of vCPU exhibits similar performance• memory size does not have much impact on performance• 2 x vCPU => 1.5 x performance
mbw
• mbw 128 • MB/s• different systems exhibit similar memory performance
0
2000
4000
6000
8000
10000
12000
XSmal l Smal l Medi um Large XLarge XXLarge
MEMCPYDUMPMCBLOCK
iozone – os disk
• iozone -Mcew -i0 -i1 -i2 -s4g -r256k -f /io.tmp• KB/s• difference systems exhibit similar write performance • L / XL / XXL systems exhibit much better read performance• cgroup blkio throttling? QEMU blk throttle? Different disk types?
0
1000000
2000000
3000000
4000000
5000000
6000000
7000000
XSmal l Smal l Medi um Large XLarge XXLarge
wri terewri terandom wri tereadrereadrandom read
iozone – data disk
• iozone -Mcew -i0 -i1 -i2 -s4g -r256k -f /mnt/io.tmp• KB/s• difference systems exhibit similar write performance • XL / XXL systems exhibit much better read performance• cgroup blkio throttling? QEMU blk throttle? Different disk types?
0
1000000
2000000
3000000
4000000
5000000
6000000
XSmal l Smal l Medi um Large XLarge XXLarge
wri terewri terandom wri tereadrereadrandom read
iperf
• (x, y) represents the bandwidth between two systems• Mbps• bandwidth limited by the system with lower configuration • cisco quantum plugin?
XSmall Small Medium Large XLargeXXLarge
XSmall 25 25 25 25 25 25
Small 25 50 50 50 50 50
Medium 25 50 100 100 100 100
Large 25 50 100 200 200 200
XLarge 25 50 100 200 400 400
XXLarge 25 50 100 200 400 650
hadoop wordcount single node
• hadoop wordcount application provided in official distribution• test directory with 3 files, total file size is 2 GB.• test result shows the time needed to finish the calculation (s)
0
100
200
300
400
500
600
700
800
900
XSmal l Smal l Medi um Large XLarge XXLarge
2GB
hadoop wordcount multiple nodes
• dfs.replication = nNodes• test directory with 3 files, total file size is 2 GB.• test result shows the time needed to finish the calculation (s)
0100200300400500600700800900
1000
1 xXsmal l
2 xXSmal l
3 xXSmal l
4 xXSmal l
Smal l XXLarge
pgbench
• postgresql-9.1, postgresql-contrib-9.1• pgbench -i -s 16 pgbench• pgbench -t 2000 -c 16 –j n -U postgres pgbench • in multi-thread testing, n-Thread = n-vCPU
0
200
400
600
800
1000
1200
1400
1600
1800
XSmal l Smal l Medi um Large XLarge XXLarge
Si ngl e ThreadMul t i Thread
defects – pgbench single thread
0
200
400
600
800
1000
1200
XSmal l Smal l Medi um Large XLarge XXLarge
NormalDef ect
• defects were observed in all VM models • test results were smooth on the same VM instance• the following test results were not affected on defected instances
• mbw• iperf• byte-unixbench
defects – iozone write results
0
50000
100000
150000
200000
250000
300000
XSmal l Smal l Medi um Large XLarge XXLarge
NormalDef ect
• test performed on OS disks only• write performance seems to be the major problem
defects – iozone read results
• test performed on OS disks only• read performance is similar for all instances in both cases
0
1000000
2000000
3000000
4000000
5000000
6000000
7000000
XSmal l Smal l Medi um Large XLarge XXLarge
NormalDef ect
defect rate
• 7 defected instances were found out of 20 total instances • defect rate too high for deploying production systems • need extra caution when VM’s are auto-generated by API’s
%3520
7 =
conclusion
• HP defects were not directly related to OpenStack • OpenStack still lacks key functionalities for production deployment• building IaaS service is more complicated than installing OpenStack• open source IaaS software => IaaS support and service => $$$
先以欲勾牵,后令入佛智。鸠摩罗什大师译 《维摩诘所说经 .佛道品第八 》