Abstract
We consider the problem of heterogeneous checkpointing in distributed systems. We propose a new solution to the problem that is truly heterogeneous in that it can support new architectures without any information about the architecture. The ability to support new architectures without additional knowledge or custom configuration is an important contribution of this work. This ability is particularly useful in mobile settings in which there is no a priori knowledge of the potential machines on which the application might execute. Our solution supports execution in unknown settings as long as there is compiler support for the high-level language in which the application is written. We precisely define what it means for a particular solution to be heterogeneous and discuss the heterogeneity of our solution and other solutions. We use code instrumentations at the source code level to provide heterogeneous checkpointing and recovery.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings of the IEEE Symposium on Reliable Distributed Systems |
| Pages | 56-65 |
| Number of pages | 10 |
| State | Published - 2001 |
| Event | 20th IEEE Symposium on Reliable Distributed Systems (SRDS 2001) - New Orleans, LA, United States Duration: Oct 28 2001 → Oct 31 2001 |
Other
| Other | 20th IEEE Symposium on Reliable Distributed Systems (SRDS 2001) |
|---|---|
| Country/Territory | United States |
| City | New Orleans, LA |
| Period | 10/28/01 → 10/31/01 |
ASJC Scopus subject areas
- Hardware and Architecture