Transcript OS Project A database and retrieval structure for

OS Project Review-2
A database and retrieval
structure for check pointing and
restarting cloud applications
Shivapriya B.J
MT2009011
Shifali Choubey
MT2009028
Koteswar Rao M
MT2009081
Divya Muthavarapu
MT2009091
Partha Sarkar
MT2009134
Proposed Architecture
Architecture(cont..)
1. In order to checkpoint the state of the
Virtual Machine, Node controller will
periodically issue a command to all the
Virtual Machines.
2. Node Controller will then take the snapshot
of the Virtual Machine and store it in the
database associated with it.
3. When any Virtual Machine goes down, the
Node Controller will issue a command to
install a new Virtual Machine.
Implementation

The process is divided into two parts
Database manager
Application manager
Database Manager

The major function of Database manager
is to
-Check the virtual machines which
are running inside a particular node
controller
- Checkpointing these virtual
machines periodically.
Application Manager
•
Its major function is to:
- Check the status of the Virtual
machine as to whether it is crashed or
not.
- To restore the Virtual Machine
when crashed.
Implementation(Cont..)
Implementation(Cont..)
1.
2.
3.
4.
Database Manager will issue a request to Virtual
Machine for checkpointing its state.
Database Manager will fetch the checkpointed
data(RAM Image,Disk Image) from Virtual
Machine and stores it into the database.
Application Manager get the checkpointed data
of a virtual machine which was crashed.
Application Manager restores the virtual
machine back to its running state.
Approach Followed
•
•
•
Initially, we check pointed the entire Virtual
Machine.
DrawBack:
It is a very difficult task to checkpoint
and restore the entire VM each time because
of its huge size.
Solution to the above DrawBack:
Pull Mechanism and
Incremental/Differential BackUp mechanisms
are used.
Incremental backup
•
•
•
In incremental backup, each original piece
of backed up information is stored only
once, and then the successive calls backup
only the information that changed since a
previous backup.
This makes incremental back-up highly
efficient and fast.
Using “rdiff-backup”, incremental backup
is implemented.
Sharable Storage System
•
•
•
These distributed databases are
implemented using NFS(Network File
System).
NFS allows hosts to mount partitions on a
remote system and use them as though they
are local file systems.
This allows the system administrator to
store resources in a central location on the
network, providing authorized users
continuous access to them.
MOUNTING PROCESS
Work Done
Successfully Checkpointed the Virtual
Machine
 Successfully restored the checkpointed
VM on different systems.
 Successfully Implemented Incremental
Backup.
 Successfully Implemented the Sharable
Storage System.

Future Targets
To make the Database Manager and
Application Manager run periodically.
 To automate the Database Manager and
Application Manager.

Project Link

http://code.google.com/p/cloud-6e
THANK YOU