The pressing demand to deploy software updates without stopping running programs has fostered much research on live update systems in the past decades. Prior solutions, however, either make strong assumptions on the nature of the update or require extensive and error-prone manual effort, factors which discourage the adoption of live update. This paper presents Mutable Checkpoint-Restart (MCR), a new live update solution for generic (multiprocess and multithreaded) server programs written in C. Compared to prior solutions, MCR can support arbitrary software updates and automate most of the common live update operations. The key idea is to allow the running version to safely reach a quiescent state and then allow the new version to restart as similarly to a fresh program initialization as possible, relying on existing code paths to automatically restore the old program threads and reinitialize a relevant portion of the program data structures. To transfer the remaining data structures, MCR relies on a combination of precise and conservative garbage collection techniques to trace all the global pointers and apply the required state transformations on the fly. Experimental results on popular server programs (Apache httpd, nginx, OpenSSH and vsftpd) confirm that our techniques can effectively automate problems previously deemed difficult at the cost of negligible performance overhead (2% on average) and moderate memory overhead (3.9x on average, without optimizations).

GentleRain is a new causally consistent geo-replicated data store that provides throughput comparable to eventual consistency and superior to current implementations of causal consistency.GentleRain uses a periodic aggregation protocol to determine whether updates can be made visible in accordance with causal consistency. Unlike current implementations, it does not use explicit dependency check messages, resulting in a major throughput improvement at the expense of a modest increase in update visibility. Furthermore, GentleRain tracks causal consistency by attaching to updates scalar timestamps derived from loosely synchronized physical clocks. Clock skew does not cause violations of causal consistency, but may delay the visibility of updates. By encoding causality in a single scalar timestamp, GentleRain reduces storage and communication overhead for tracking causality.We evaluate GentleRain using Amazon EC2, and demonstrate that it achieves throughput equal to about 99% of eventual consistency, and 120% better than previous implementations of causal consistency.

Understanding the performance of data-parallel workloads when resource-constrained has significant practical importance but unfortunately has received only limited attention. This paper identifies, quantifies and demonstrates memory elasticity, an intrinsic property of dataparallel tasks. Memory elasticity allows tasks to run with significantly less memory that they would ideally want while only paying a moderate performance penalty. For example, we find that given as little as 10% of ideal memory, PageRank and NutchIndexing Hadoop reducers become only 1.2x/1.75x and 1.08x slower. We show that memory elasticity is prevalent in the Hadoop, Spark, Tez and Flink frameworks. We also show that memory elasticity is predictable in nature by building simple models for Hadoop and extending them to Tez and Spark.To demonstrate the potential benefits of leveraging memory elasticity, this paper further explores its application to cluster scheduling. In this setting, we observe that the resource vs. time trade-off enabled by memory elasticity becomes a task queuing time vs task runtime trade-off. Tasks may complete faster when scheduled with less memory because their waiting time is reduced. We show that a scheduler can turn this task-level tradeoff into improved job completion time and cluster-wide memory utilization. We have integrated memory elasticity into Apache YARN. We show gains of up to 60% in average job completion time on a 50-node Hadoop cluster. Extensive simulations show similar improvements over a large number of scenarios.

The pressing demand to deploy software updates without stopping running programs has fostered much research on live update systems in the past decades. Prior solutions, however, either make strong assumptions on the nature of the update or require extensive and error-prone manual effort, factors which discourage live update adoption.This paper presents Mutable Checkpoint-Restart (MCR), a new live update solution for generic (multiprocess and multithreaded) server programs written in C. Compared to prior solutions, MCR can support arbitrary software updates and automate most of the common live update operations. The key idea is to allow the new version to restart as similarly to a fresh program initialization as possible, relying on existing code paths to automatically restore the old program threads and reinitialize a relevant portion of the program data structures. To transfer the remaining data structures, MCR relies on a combination of precise and conservative garbage collection techniques to trace all the global pointers and apply the required state transformations on the fly. Experimental results on popular server programs (Apache httpd, nginx, OpenSSH and vsftpd ) confirm that our techniques can effectively automate problems previously deemed difficult at the cost of negligible run-time performance overhead (2% on average) and moderate memory overhead (3.9x on average).

No abstract