The Panasas ActiveScale Storage Cluster - Delivering Scalable High Bandwidth Storage

By Hong Tang, Aziz Gulbeden, Jingyu Zhou, William Strathearn, Tao Yang, Lingkun Chu. Proceedings of the ACM/IEEE SC2004 Conference, 2004

399 citations recorded by the source corpus.

Subfields: Computer Networks and Communications

View Paper: https://www.semanticscholar.org/paper/9cdda3f5457fd8637e06f7ed9fee40060ddd15f4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

Citations in the corpus

ICTCP: Incast Congestion Control for TCP in data center networks (2010, methodology): “TCP incast has been identified and described by Nagle et al. [ 12 ] in distributed storage clusters.”
Fastpass (2014, background): “With small router queues, response packets are lost, triggering delay-inducing retrans-missions [26], whereas large queues cause delays to other bulk trafﬁc.”
Measurement and Analysis of TCP Throughput Collapse in Cluster-based Storage Systems (2008, methodology): “Cluster-based storage systems are becoming an increasingly important target for both research and industry [3, 30, 15, 24 , 14, 10].”
Remote sensing big data computing: Challenges and opportunities (2015, background): “Recently, great efforts have been towards parallel file systems, such as OrangeFS [59], PVFS [29], Lustre [60], PanFS [61] and GPFS [62].”
Safe and effective fine-grained TCP retransmissions for datacenter communication (2009, methodology): “These timeouts and the resulting delay can reduce application throughput by 90% (Section 3) or more [ 25 , 28].”
CRUSH: Controlled, Scalable, Decentralized Placement of Replicated Data (2006, background): “The Sor-rento [Tang et al. 2004] storage system’s use of consistent hashing [Karger et al. 1997] most closely resembles CRUSH, but lacks support for controlled weighting of devices, a well-balanced distribution of data, and failure domains for improving data safety.”
DRILL: Micro Load Balancing for Low-latency Data Center Networks (2017, background): “…Saturn does not appear to be broadly representative of most datacenter fabrics. study the problem within a cluster (hosts connected via one switch or a tree topology), and naturally focus exclusively on overrun of the last hop buffer (connected to the receiver) [29, 30, 33, 43, 55, 60, 69–72].”