Subhabrata Bhattacharya

1
Virtualization in Grid Computing

• Subhabrata Bhattacharya
• Grid Computing, SETLABS
• Infosys Technologies Ltd.

2
Agenda

• Grid computing - An Overview
• Problems and Challenges in Grids
• Need of execution environments
• Virtualization The context
• Techniques
• Benefits
• Issues
• Work _at_ Infosys
• Concluding notes

3
Grid Computing An Overview

• A software and hardware infrastructure, that
  allows flexible and seamless sharing of
  heterogeneous network of resources for compute
  and data intensive tasks and provides faster
  throughput at lower costs

4
What drives Grid ?

• High-throughput
• Scalability
• Globally distributed resources
• Infrastructure maintenance
• Disparate systems
• Efficient resource utilization
• Complete resource virtualization

5
Grid - Open Challenges

• Collective Administration
• On Demand Application Integration
• Resource monitoring and provisioning
• Flexibility and Security
• Guaranteeing execution
• Fault determination,tolerance, resolution
• Policy based management

6
Execution environments -Requirements

• Isolation
• Host from Jobs
• Jobs from Jobs
• Jobs from Host
• Resource provisioning
• Memory/CPU ticks/Storage/Network Bandwidth
• Security
• Fine-grained policy management
• Dynamic configuration

7
Execution Environments - Issues

• Latency in start-up
• Guaranteed provisioning
• Dynamic configuration management
• Ensuring safe execution
• Fault tolerance
• High availability
• Efficient utilization

8
Virtualization - A solution
A framework or methodology of dividing the
resources of a computer into multiple execution
environments, by applying one or more concepts or
technologies such as hardware and software
partitioning, time-sharing, partial or complete
machine simulation, emulation, quality of
service, and many others Source -
http//www.kernelthread.com
9
Virtual Machines
A system VM provides a complete, persistent
system environment that supports an operating
system along with its many user processes. It
provides the guest operating system with access
to virtual hardware resources, including
networking, I/O, and perhaps a graphical user
interface along with a processor and memory.

Source Architecture of Virtual Machines, Smith
Nair, Computer, May 2005, pp 32-38
10
Techniques

• Full System Virtualization (None)
• Para Virtualization (Xen)
• Hosted Virtual Machine (VM ware, UML)
• Shared Kernel virtualization (LinuxVserver,
  FreeVPS, BSD Jails)

11
Xen
Source http//www.cl.cam.ac.uk/netos/papers/2003
-xensosp.pdf, p-5

• Hypervisor (VMM) sits on top of H/W
• Ported to Linux/FreeBSD/NetBSD
• Hosted OS kernel modification required
• Near-native performance
• Highly scalable

12
User Mode Linux
Source http//user-mode-linux.sourceforge.net/sl
ides/als2000/img2.html

• Guest Kernel runs as user process on of host
  kernel
• Host kernel intercepts calls from guests,
  translates
• System calls implement Device drivers in guest
  kernel
• Highly secure and fault tolerant
• Significant performance overhead
• Used for Kernel development and debugging

13
Linux VServer

• Supports Linux kernel 2.4.x and 2.6.x
• Safe execution contexts for processes
• Based on Linux capability model
• Low performance overhead
• Highly scalable

14
Advantages

• Environment set-up for legacy systems
• Facilitating wide range of testing environments
• Reducing cost by eliminating high-end hardware
• Preventing resource under-utilization

15
Virtualization on Grid

• Grid integrates fragmented resources
• Provision through Virtualization
• SLA guarantees ensured
• Differentiated services
• Advance reservations
• Metering and Accounting (Based on VM uptime)
• Check-pointing
• Virtualization allows VM state save
• Jobs migration to standby VM
• VM migration to suitable grid node
• Security in Grid
• Virtualization provides isolation across VMs

16
Work _at_ Infosys

• Performance issues in virtualization
• Reduce start-up and tear down times of VMs
• Provide Metering and Accounting information to
  aid Grid Management
• Provide support for Fault tolerance, check
  pointing and recovery
• Seamlessly operate across virtual machines from
  different vendors
• Facilitate logical server consolidation through
  a policy level management tool

17
Conclusion

• Revised processor arch enables full system
  virtualization
• Reliability, high availability
• Scalable and secure infrastructure
• Low cost, on demand computing
• Facilitating cost based computing

18
References

• http//www-1.ibm.com/grid/
• http//www.gridforum.org
• http//www.freevps.com/docs/vps_concept.html
• http//linux-vserver.org/Linux-VServer-Paper
• http//www.vmware.com
• http//www.cl.cam.ac.uk/Research/SRG/netos/xen
• http//docs.freebsd.org/44doc/papers/jail/jail.ht
  ml
• http//www.freevps.com
• http//www.openvps.com
• http//www.sun.com/bigadmin/content/zones/

19
Thanks!
Questions?