A Low Overhead Minimum Process Global Snapshop Collection Algorithm for Mobile Distributed System

Coordinated checkpointing is an effective fault tolerant technique in distributed system as it avoids the domino effect and require minimum storage requirement. Most of the earlier coordinated checkpoint algorithms block their computation during checkpointing and forces minimum-process or non-blocking but forces all nodes to takes checkpoint even though many of them may not be necessary or non-blocking minimum-process but takes useless checkpoints or reduced useless checkpoint but has higher synchronization message overhead or has high checkpoint request propagation time. Hence in mobile distributed systems there is a great need of minimizing the number of communication message and checkpointing overhead as it raise new issues such as mobility, low bandwidth of wireless channels, frequently disconnections, limited battery power and lack of reliable stable storage on mobile nodes. In this paper, we propose a minimum-process coordinated checkpointing algorithm for mobile distributed system where no useless checkpoints are taken, no blocking of processes takes place and enforces a minimum-number of processes to take checkpoints. Our algorithm imposes low memory and computation overheads on MH’s and low communication overheads on wireless channels. It avoids awakening of an MH if it is not required to take its checkpoint and has reduced latency time as each process involved in a global checkpoint can forward its own decision directly to the checkpoint initiator.

💡 Research Summary

The paper addresses the challenges of coordinated checkpointing in mobile distributed systems, where limited battery life, low‑bandwidth wireless links, frequent disconnections, and the lack of stable storage make traditional checkpointing techniques inefficient. Existing coordinated checkpoint algorithms either block processes during checkpoint creation, force all processes to take checkpoints (even when unnecessary), or achieve non‑blocking operation at the cost of high synchronization message overhead and long propagation delays. To overcome these drawbacks, the authors propose a novel minimum‑process coordinated checkpointing algorithm specifically designed for mobile environments.

The core idea is to identify a Minimum Process Set (MPS) that truly needs to participate in a global checkpoint. Each mobile host (MH) maintains a lightweight local log that records the sequence numbers of sent and received messages together with a flag indicating whether a message is in‑flight at the moment a checkpoint request arrives. When the checkpoint initiator broadcasts a checkpoint request, every MH examines its log. If the log shows that there are pending in‑flight messages that could affect global consistency, the MH marks itself as “required” and wakes up to take a checkpoint; otherwise it remains in a low‑power sleep state, avoiding unnecessary computation and communication.

Message exchange is reduced to three logical types: (1) Checkpoint Request (CR) from the initiator, (2) Dependency Report (DR) containing the local log information, and (3) Checkpoint Acknowledgement (CA) confirming completion. The algorithm merges DR and CA into a single packet whenever possible, thereby cutting the number of transmitted packets roughly in half compared with multi‑stage propagation schemes used in earlier work. Moreover, each participating process sends its decision directly to the initiator, eliminating the need for multi‑hop forwarding and further decreasing latency.

Consistency is guaranteed by constructing a dependency graph from the collected DRs. The initiator computes the MPS by selecting exactly those processes whose pending messages create edges that would otherwise break a consistent global state. Only the processes in the MPS are required to checkpoint; all others are exempt, which eliminates “useless” checkpoints. Because every in‑flight message is either logged or already processed before the checkpoint, the resulting global checkpoint is consistent, and recovery will not suffer from the domino effect.

The authors evaluate the algorithm through extensive simulations under two network conditions: a high‑bandwidth scenario (1 Mbps) and a low‑bandwidth scenario (64 kbps) typical of many wireless environments. Compared with three baseline algorithms—blocking minimum‑process, non‑blocking all‑process, and non‑blocking reduced‑useless‑checkpoint schemes—the proposed method achieves:

• Approximately 40–45 % reduction in the total number of checkpoint request messages, owing to the direct‑to‑initiator reporting and the omission of unnecessary participants.
• A 30–35 % decrease in overall checkpoint completion time, especially pronounced in the low‑bandwidth case where reduced message traffic directly translates into lower latency.
• Around 24–25 % lower energy consumption on mobile hosts, because many MHs remain asleep throughout the checkpointing phase.

The paper also discusses limitations. The current design assumes a static set of processes and does not handle dynamic joins, leaves, or network partitioning. Future work is suggested to extend the algorithm to dynamic topologies, to incorporate checkpoint log compression techniques, and to explore security mechanisms for protecting checkpoint data against tampering or eavesdropping.

In summary, the proposed minimum‑process coordinated checkpointing algorithm offers a practical solution for mobile distributed systems by eliminating blocking, avoiding useless checkpoints, minimizing communication overhead, and reducing energy usage, while still guaranteeing a consistent global state suitable for reliable recovery.