Hokkaido University Academic Cloud:
   Largest Academic Cloud System in Japan
Masaharu Munetomo
Information Initiative Cen...
Information Initiative Center, Hokkaido University
• A key institute of “high performance computing infrastructure” (HPCI)...
Hokkaido University Academic Cloud
• Largest Academic Cloud System in Japan: 43TFlops
• More than 2,000 VMs can be deploye...
An overview of software architecture
HW Boot/VM Image: SAN Data: NAS
RedHat/CentOS Hypervisor ( XenServer / VMWare)
Cloud&...
Why we employ CloudStack?
Why we employ CloudStack?
Why we employ CloudStack?
• High quality and usability in user portal and management
console
Why we employ CloudStack?
• High quality and usability in user portal and management
console
• Easy to control resources v...
Why we employ CloudStack?
• High quality and usability in user portal and management
console
• Easy to control resources v...
VM Hosting Services: Hosting & Project servers
• Hosting servers: for web-hosting, etc. including CMS
middleware packages ...
In-house developed cloud portal
Selection of service-level
Selection of packages
# of VMs for cluster
Authentication Config...
Network configuration
• Assign a VLAN to each user (laboratory)
Physical Node #1 Physical Node#2
Virtual

Router

for user ...
Automated Deployment of VM clusters
• Customizing scheduling policies in CloudStack to balance I/O overheads for
cluster p...
Shared Storage #1
Resource Pool #1
HyperVisor #2
HyperVisor #1
Virtual(
DiskVM(
Shared Storage #2
Resource Pool #2
HyperVi...
Deployment time for clusters
• Fast deployment: 5 min. for 1 VM, around 10 min. for 20 VMs
• A medium size cluster consist...
Heterogeneous hybrid cloud deployment test
• CloudStack@Hokkaido University - OpenNebula@TokyoTech
We developed a VPN clie...
SHINCLOM (Simple Heterogeneous INter-CLOud
Manager): Inter-cloud manager for academic cloud systems
SHINCLOM (Simple Heterogeneous INter-CLOud
Manager): Inter-cloud manager for academic cloud systems
SHINCLOM (Simple Heterogeneous INter-CLOud
Manager): Inter-cloud manager for academic cloud systems
Upcoming SlideShare
Loading in …5
×

Hokkaido University Academic Cloud: Largest Academic Cloud System in Japan

1,210 views

Published on

Hokkaido university academic cloud which started services in 2011, is the largest academic cloud system in Japan. Its peak performance is 43TFlops and HPC/Hadoop cluster instances can be deployed automatically.

Published in: Technology, Business
0 Comments
3 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,210
On SlideShare
0
From Embeds
0
Number of Embeds
66
Actions
Shares
0
Downloads
27
Comments
0
Likes
3
Embeds 0
No embeds

No notes for slide

Hokkaido University Academic Cloud: Largest Academic Cloud System in Japan

  1. 1. Hokkaido University Academic Cloud:    Largest Academic Cloud System in Japan Masaharu Munetomo Information Initiative Center, Hokkaido University, Sapporo, JAPAN. munetomo@iic.hokudai.ac.jp Cloud Technical Leadership Forum @San Francisco Intercontinental Hotel Apr. 18th, 2012
  2. 2. Information Initiative Center, Hokkaido University • A key institute of “high performance computing infrastructure” (HPCI) in Japan. • Founded in 1962 as a national supercomputing center. • University R&D center for Supercomputing, Cloud computing, Networking, IT systems for education • Supercomputer (172TFlops, #95 in TOP500) & Cloud System (43TFlops)
  3. 3. Hokkaido University Academic Cloud • Largest Academic Cloud System in Japan: 43TFlops • More than 2,000 VMs can be deployed • High-performance cloud system: each physical node has 40-cores, 128GB memory. Network: 10GbE x 2, Shared Storage: 760TB Hitach BladeSymphony BS2000 Xeon E7 8870 2.4GHz (10-core) x 4 128GB memory / 10GbE x 2 Hitach NAS Storage AMS2300: 260TB AMS2500: 500TB
  4. 4. An overview of software architecture HW Boot/VM Image: SAN Data: NAS RedHat/CentOS Hypervisor ( XenServer / VMWare) Cloud&Middleware&(CloudStack) Portal&system& (In9house&developed Univ.&SSO& CloudStack& API access
  5. 5. Why we employ CloudStack?
  6. 6. Why we employ CloudStack?
  7. 7. Why we employ CloudStack? • High quality and usability in user portal and management console
  8. 8. Why we employ CloudStack? • High quality and usability in user portal and management console • Easy to control resources via API
  9. 9. Why we employ CloudStack? • High quality and usability in user portal and management console • Easy to control resources via API • Open source
  10. 10. VM Hosting Services: Hosting & Project servers • Hosting servers: for web-hosting, etc. including CMS middleware packages (WordPress, MediaWiki, etc.) • Project servers: for research projects • Cluster packages are also available: Hadoop/MPI Class Cores Memory Storage JPY/month S 1 3GB 100GB 525 M 4 12GB 100GB 2,100 L 10 30GB 100GB 5,250 XL 40 128GB 2TB 21,000
  11. 11. In-house developed cloud portal Selection of service-level Selection of packages # of VMs for cluster Authentication Config. FW initial Config.
  12. 12. Network configuration • Assign a VLAN to each user (laboratory) Physical Node #1 Physical Node#2 Virtual
 Router
 for user A eth0% eth0% eth0.200 eth0.201 eth0% eth0.800 eth0% eth0.200eth0.201 Virtual
 Router
 for user B eth1% eth0% eth0%eth0%eth1% eth0% VLAN+200!VLAN+800! VLAN+201! VLAN+200!VLAN+201! eth0.800 VLAN+800! Virtual%Interface Bridge Send%with% VLAN%tag %Private%Network%for%User%A %Private%Network%for%User%B %Global%Network Virtual%NIC
  13. 13. Automated Deployment of VM clusters • Customizing scheduling policies in CloudStack to balance I/O overheads for cluster packages (Hadoop / MPI / Torque). Storage #3 Virtual( Disk Storage #4 Virtual( Disk Storage #2 Virtual( Disk Zone! POD! Shared Storage #1 Resource Pool #1 HyperVisor #2 HyperVisor #1 Virtual( DiskVM( Balancing!overheads!of!disk!I/O!with! round8robin!assignment!of!Virtual!disks.! Storage #1 VM( VM( VM( VM( Virtual( DiskHadoop Cluster Shared Storage #2 Resource Pool #2 HyperVisor #4 HyperVisor #3 Virtual( Disk VM( Shared Storage #3 Resouce Pool #3 HyperVisor #6 HyperVisor #5 Virtual(( Disk VM( Shared Storage #4 Resouce Pool #4 HyperVisor #8 HyperVisor #7 Virtual( Disk VM(
  14. 14. Shared Storage #1 Resource Pool #1 HyperVisor #2 HyperVisor #1 Virtual( DiskVM( Shared Storage #2 Resource Pool #2 HyperVisor #4 HyperVisor #3 Virtual( Disk VM( Shared Storage #3 Resouce Pool #3 HyperVisor #6 HyperVisor #5 Virtual(( Disk VM( Shared Storage #4 Resouce Pool #4 HyperVisor #8 HyperVisor #7 Virtual( Disk VM( CloudStack+ (ManagementServer)+ CloudStack+ DB+ Planner+ Storage+load+balancing VM+Alloca<on Planner+ Request VM+Deployment+ Job Deployment( Informa7on Host Storage Deployment( request Deploy(VMs CloudStack+ Portal
  15. 15. Deployment time for clusters • Fast deployment: 5 min. for 1 VM, around 10 min. for 20 VMs • A medium size cluster consisting of 257 VMs can be deployed in a little more than 1 hour ! • It greatly speedup starting research projects since conventional physical cluster needs more than a week or even a month to finish deployment including purchase process. y"="0.2267x"+"4.1748" R²"="0.9989 0" 10" 20" 30" 40" 50" 60" 70" 0" 50" 100" 150" 200" 250" 300" Deployment*+me*(min) #*of*VMs
  16. 16. Heterogeneous hybrid cloud deployment test • CloudStack@Hokkaido University - OpenNebula@TokyoTech We developed a VPN client to establish connection via API (Python + OpenSwan + l2tpd) Gateway Management Server Computing Node Gateway Virtual Router VPN Client VPN Server CloudStack Virtual Route VPN
  17. 17. SHINCLOM (Simple Heterogeneous INter-CLOud Manager): Inter-cloud manager for academic cloud systems
  18. 18. SHINCLOM (Simple Heterogeneous INter-CLOud Manager): Inter-cloud manager for academic cloud systems
  19. 19. SHINCLOM (Simple Heterogeneous INter-CLOud Manager): Inter-cloud manager for academic cloud systems

×