Search Results (2)

NVMe-over-Fabrics (NVMeoF) is expected to have high-performance and be highly scalable for disaggregating NVMe SSDs to High-Speed Network (HSN)-attached storage servers, thus the aggregated NVMe SSDs in storage servers can be elastically allocated to remote host servers for better utilization. However, due to the well-known connection scalability issue of RDMA NICs (RNICs), RDMA-enabled HSN can only provide a limited scale of performant Queue Pairs (QPs) for NVMeoF I/O queues to transfer capsule and data between the storage server and remote host servers. However, in current NVMeoF implementations, multiplexing multiple NVMeoF I/O queues onto a single RNIC QP is not supported yet. In this paper, we investigate how NVMeoF capsule and data transfers are performed efficiently over HSN with a limited number of RNIC QPs, and propose SPANoF, a Scalable and Performant Architecture for NVMe-over-Fabrics. SPANoF dissolves the intrinsic one-to-one mapping relationship between NVMeoF I/O queues and RNIC QPs, allocates a dedicated send-list for each NVMeoF I/O queue rather than for each RNIC QP, transfers NVMeoF capsules and data in send-lists with a QP-centric manner to remove lock-contention overhead, and polls for transfer completion notifications to remove interrupt-caused context switch overhead. We implemented SPANoF in the Linux kernel and evaluated it by the FIO benchmarks. Our experimental results demonstrate that SPANoF can avoid the performance collapses for commercial RNICs with a limited number of performant QPs and avoid the system crash for domain-specific RNICs with only limited-scale available QPs. Compared with the native NVMeoF implementation in Linux kernel, SPANoF can saturate an RNIC of the storage server with only three RNIC QPs of the remote host server. Compared with lock-based QP-sharing mechanisms, SPANoF improves bandwidth by up to 1.55× under 64 KB sequential write requests, improves throughput by up to 4.18× and reduces the average latency by 28.31% under 4 KB random read requests. Full article

Distributing Big Data for science is pushing the capabilities of networks and computing systems. However, the fundamental concept of copying data from one machine to another has not been challenged in collaborative science. As recent storage system development uses modern fabrics to provide faster remote data access with lower overhead, traditional data movement using Data Transfer Nodes must cope with the paradigm shift from a store-and-forward model to streaming data with direct storage access over the networks. This study evaluates NVMe-over-TCP (NVMe-TCP) in a long-distance network using different file systems and configurations to characterize remote NVMe file system access performance in MAN and WAN data moving scenarios. We found that NVMe-TCP is more suitable for remote data read than remote data write over the networks, and using RAID0 can significantly improve performance in a long-distance network. Additionally, a fine-tuning file system can improve remote write performance in DTNs with a long-distance network. Full article