LCCFS: a lightweight distributed file system for cloud computing without journaling and metadata services

Abstract

The major usage of a file system integrated with a cloud computing platform is to provide the storage for VM (virtual machine) instances. Distributed file systems, especially those implemented on top of object storage have many potential advantages over traditional local file systems for VM instance storage. In this paper, we make an investigation in the requirements imposed on a file system in cloud computing scenario, and claim that the implementation of a file system for VM instance storage could be reasonably simplified. We demonstrate that on top of an object storage with simple object-granularity transaction support, a lightweight distributed file system, which requires neither journaling nor dedicated metadata services, can be developed for cloud computing. We have implemented such a distributed file system, called LCCFS (lightweight cloud computing file system), based on the RADOS (reliable autonomic distributed object storage) object storage. Our experimental results show that for the main workloads in cloud computing, LCCFS achieves almost the same or slightly higher performance than CephFS (ceph filesystem), another published distributed file system based on RADOS. Compared to CephFS, LCCFS has only one tenth of its LOCs (lines of code). This theoretical simplicity makes it easy to implement LCCFS correctly and stably by avoiding the sheer design and implementation complexity behind CephFS, thereby making LCCFS a promising candidate in the cloud computing production environment.