Real time distributed control system for information security in instrumentation networks. A distributed system can be defined as a set of homogeneous or heterogeneous nodes communicating through a LAN or WAN network. We need distributed systems because it can deliver High performance due to load balancing. Due to its unique architecture, system offer High availability since there is no single point of failure. Due to it's distributed architecture, managing resources at multiple node is bit tedious. Clocks of all the nodes should be synchronized. System should handle communication related errors arising due to out of order delivery of packets and packet loss. Tasks allotted to each node is independent of other nodes. To manage workload locally, each node have a resource manager. It also helps to migrate workload to remote nodes. Cooperation between the nodes, helps the system to meet the deadlines for a task.
Scheduling and Resource management in Distributed RT systems
- Local scheduling
- Resource management within a node
- Task scheduling
- Global scheduling
- Balancing load across nodes by transferring tasks from a heavily loaded local node to a lightly loaded node
- Transfer policy - When Tasks are migrated from a node to another
- Selection policy - Identifying the task to be migrated
- Information policy - Details of task that are going to be migrated
- Location policy - Destination of the migrated task
- Communication resource management
- QoS routing and Resource reservation for channel setup time
- Packet scheduling (run-time)
A node's load is measured by queue length and processor utilization. Before doing a Transfer, Transfer policy determines whether the current node is suitable to participate in a task migration either as a sender or as a receiver. It is the duty of selection policy decide tasks that should be migrated, once transfer policy is completed. While tasks are picked, make sure that total time taken for local decision, migration, remote decision and task’s execution time is less than task’s deadline. The concepts of reliability, and security add to trustworthiness of a control system. To achieve reliability, each components of the control system must exhibit a low failure rate. The resilient aspect helps continued performance of critical functions in case of failure of non critical sub tasks. Control systems should meet stated objectives of a system irrespective of external and internal disturbances such as process variation, component failure and communication degradation. To avoid unauthorized intrusion, system should block the paths that are vulnerable.