Serverless network file systems

Abstract

In this paper, we propose a new paradigm for network file system design, serverless network file systems. While traditional network file systems rely on a central server machine, a serverless system utilizes workstations cooperating as peers to provide all file system services. Any machine in the system can store, cache, or control any block of data. Our approach uses this location independence, in combination with fast local area networks, to provide better performance and scalability than traditional file systems. Further, because any machine in the system can assume the responsibilities of a failed component, our serverless design also provides high availability via redundant data storage. To demonstrate our approach, we have implemented a prototype serverless network file system called xFS. Preliminary performance measurements suggest that our architecture achieves its goal of scalability. For instance, in a 32-node xFS system with 32 active clients, each client receives nearly as much read or write throughput as it would see if it were the only active client.

Hint-based cooperative caching

“…The simulator of the hint-based and ideal algorithms we wrote ourselves. We obtained the N-Chance simulator from the developers of that system, and modified it to incorporate additional functionality employed in the xfs file system [Anderson et al 1995]. In the modified system, a manager preferentially forwards a request to the client caches instead of a server, improving the cooperative caching hit rate and reducing the server load.…”

“…The xfs file system uses a refined version of N-Chance [Anderson et al 1995]. A subsequent paper by Feeley et al [1995] described the Global Memory Service (GMS) that provides better performance than N-Chance as well as reduced overhead.…”

Design and evaluation of remote memory disk cache

“…The researchers presented only simulation result, and their simulation result showed that cooperative cache reduces disk access by 50%, and read latency by at most 73%. Later, the cooperative cache was implemented in a Serverless Network File System [5]. 2) SSD as a Cache: A hybrid drive [6] is a NAND flash memory attached disk.…”

Notice of Violation of IEEE Publication Principles - Exploring Fault-tolerant Distributed Storage System using GE code

“…We use the centre server to organize all the storage nodes, but the centre server don't must be highly reliable. Many papers [12] [13] have discussed how to construct a serverless network file system by distribute the system service to all the storage nodes. Those methods are suitable for our system too.…”