fault tolerance in corba and wireless corba chen xinyu 18/9/2002
TRANSCRIPT
Fault Tolerance in CORBA and
Wireless CORBA
Chen Xinyu
18/9/2002
Outline
Introduction to CORBA and Wireless CORBA
What is Fault Tolerance
Fault Tolerant CORBA
Fault Tolerance in Wireless CORBA
Conclusion
Future Work
What is CORBA
Common Object Request Broker Architecture• A Distributed Object Computing (DOC) open standard
– Compare to platform/language specific alternatives
– e.g., Java RMI, Microsoft’s DCOM
• A language-neutral environment
• A middleware infrastructure specification
Administered by the Object Management Group • a.k.a., the OMG
Wireless CORBA Architecture
Encapsulates, forwards or ignores incoming GIOP messages Decapsulates and forwards messages from the GIOP tunnel Generates mobility events Lists available services
Similar to the Access Bridge Does not provide forwarding Generates mobility events Does not list services
Abstract transport-independent tunnel for GIOP messages Concrete tunnels for TCP/IP, UDP/IP and WAP. Only one GIOP tunnel
Keeps track of the associated access bridges Redirects requests for services on the terminal
Source: Telecom Wireless CORBA, OMG Doucment dtc/01-06-02
Wireless CORBA
CORBA
Key:TCP/IP Network CORBA objects may be
invoked anywhere along the “end to end” path
IIOP
GTP Tunnel
GIOP
GIOPGTP
Access Point
Fault, Error and Failure
Fault Error Failure
Fault tolerant mechanisms
an anomalous condition occurring in the system hardware or software
the part of the system state that is liable to lead to a failure
occurs when the delivered service of a system or a component deviates from its specification
Fault tolerance is the ability of a system to continue providing its specified service despite
component failure
Fault tolerance is the ability of a system to continue providing its specified service despite
component failure
Fault Tolerant CORBA Architecture
Source: Bell Labs Research
Object Replication Styles
Passive Replication• Only one replica processes each request,
other replicas are available as backups
• Lower memory and processing costs
• Slower recovery from faults
• Duplicate message detection during recovery from faults
Active Replication• Several replicas process each request
• Faster recovery from faults
• State transfer to initialize new replicas
ORB ORB ORB ORB ORB
ORB ORB ORB
Passive Replication
Clientinvokes a method of
Server AServer A
Server B
Reliable totally ordered multicast
Primaryreplica
Primaryreplica
Only primary replica of Server A executes the method
Reply returnedfrom primary replica of Server Bto primary replica of Server A
Only primary replicaof Server Bexecutes the method
Reliabletotally orderedmulticastfor state transfer
ObjectObject Object Object Object
Object Object Object
Source: Eternal Systems, Inc
Active Replication
Object
ORB ORB ORB ORB ORB
ORB ORB ORB
Clientinvokes a method of
Server AServer A
Server B
Reliable totally ordered multicast
STOP STOP
Duplicate invocationssuppressed
Reliabletotally orderedmulticasts forrequests and replies
Object Object Object Object
Object Object Object
Duplicate repliessuppressed
STOPSTOP
Source: Eternal Systems, Inc
Device, Wireless & Mobile Issues
Device Issues• Slow processor
• Small memory
• Small disk space
• Low power supply
• Physical damage
Applying Mobile Host as Stable Storage
a large number of system messages or a large size of information carried in a message
Applying Access Bridge as Stable Storage
Uncoordinated checkpointing Pessimistic message logging
Checkpoints and Logs collection
Mobile Issue• Handoff
Wireless Issues
• High bit error rate
• Little bandwidth
• Long transfer delay
Recovery Scheme
Uncoordinated checkpointing• time
• predefined number of messages
Pessimistic message logging• no extra communication overhead
Independent rollback recovery• only failed objects rollback
Fault Tolerance Architecture
Client Object
Terminal Bridge
Recovery Mechanism
ORB
Platform
Mobile Host
Recovery Mechanism
Logging Mechanism
Platform
Access Bridge
Mobile Side
Fixed Side
Access Bridge
ORB
Recovery Mechanism
Logging Mechanism
ORB
Platform
Remote Server
GIOP Tunnel
Multicast Messages
Server Replica
Checkpoint and Logs Collection Strategies
Pessimistic• checkpoint and logs are transferred during handoff
• generates heavy volume of data transfer
Lazy• creates a linked list of Access Bridges
• complicated recovery
Frequency-based• the number of handoffs
Distance-based • the distance between mobile host and the Access
Bridge carrying its latest checkpoint
Mobile Host Crash
Access Bridge 1
Access Bridge 2
Access Bridge 3
Home Location
Agent
HandoffLocation Update
Mobile Host Crash
Access Bridge 1
Access Bridge 2
Access Bridge 3
Home Location
Agent
HandoffLocation Update
Mobile Host Crash
Access Bridge 1
Access Bridge 2
Access Bridge 3
Home Location
Agent
Mobile Host Crash
Access Bridge 1
Access Bridge 2
Access Bridge 3
Home Location
Agent
Collect last checkpoint and succeeded message logs
Sorted by Ack. SN
Reconnect
Messages Replay
Conclusion
Fault Tolerant CORBA is based on Object Replication
Fault tolerance in Wireless CORBA is based on Rollback-Recovery Protocol
Checkpoint and message logs collection is important in Wireless CORBA
Future Work
Low-cost Checkpointing Algorithm• forces a minimum number of objects to take
checkpoints
• minimizes the number of synchronization messages
• makes checkpointing nonblocking
Failure Detection in Wireless Environment
Question and Answer
Thank You