VAXcluster: a closely-coupled distributed system
ACM Transactions on Computer Systems (TOCS)
Fibre channel: gigabit communications and I/O for computer networks
Fibre channel: gigabit communications and I/O for computer networks
File server scaling with network-attached secure disks
SIGMETRICS '97 Proceedings of the 1997 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
VISA: Netstation's virtual Internet SCSI adapter
Proceedings of the eighth international conference on Architectural support for programming languages and operating systems
Infiniband
Obtaining High Performance for Storage Outsourcing
FAST '02 Proceedings of the Conference on File and Storage Technologies
Structure and Performance of the Direct Access File System
ATEC '02 Proceedings of the General Track of the annual conference on USENIX Annual Technical Conference
Supporting I/O-efficient scientific computation in TPIE
SPDP '95 Proceedings of the 7th IEEE Symposium on Parallel and Distributeed Processing
FAST '03 Proceedings of the 2nd USENIX Conference on File and Storage Technologies
Making the Most Out of Direct-Access Network Attached Storage
FAST '03 Proceedings of the 2nd USENIX Conference on File and Storage Technologies
ATEC '96 Proceedings of the 1996 annual conference on USENIX Annual Technical Conference
End system optimizations for high-speed TCP
IEEE Communications Magazine
On the elusive benefits of protocol offload
NICELI '03 Proceedings of the ACM SIGCOMM workshop on Network-I/O convergence: experience, lessons, implications
A case for virtual channel processors
NICELI '03 Proceedings of the ACM SIGCOMM workshop on Network-I/O convergence: experience, lessons, implications
Internet Protocol storage area networks
IBM Systems Journal
STICS: SCSI-to-IP cache for storage area networks
Journal of Parallel and Distributed Computing
Efficient Direct User Level Sockets for an Intel® Xeon" Processor Based TCP On-Load Engine
IPDPS '05 Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 9 - Volume 10
An Architecture for Software-Based iSCSI on Multiprocessor Servers
IPDPS '05 Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 9 - Volume 10
Performance of optimized software implementation of the iSCSI protocol
SNAPI '03 Proceedings of the international workshop on Storage network architecture and parallel I/Os
Performance evaluation of distributed iSCSI RAID
SNAPI '03 Proceedings of the international workshop on Storage network architecture and parallel I/Os
Analysis of iSCSI target software
SNAPI '04 Proceedings of the international workshop on Storage network architecture and parallel I/Os
Evaluating network processing efficiency with processor partitioning and asynchronous I/O
Proceedings of the 1st ACM SIGOPS/EuroSys European Conference on Computer Systems 2006
Server network scalability and TCP offload
ATEC '05 Proceedings of the annual conference on USENIX Annual Technical Conference
A scalable and high performance software iSCSI implementation
FAST'05 Proceedings of the 4th conference on USENIX Conference on File and Storage Technologies - Volume 4
TCP offload is a dumb idea whose time has come
HOTOS'03 Proceedings of the 9th conference on Hot Topics in Operating Systems - Volume 9
A nine year study of file system and storage benchmarking
ACM Transactions on Storage (TOS)
Deconstructing Network Attached Storage systems
Journal of Network and Computer Applications
SpringSim '09 Proceedings of the 2009 Spring Simulation Multiconference
Design and development of Ethernet-based storage area network protocol
Computer Communications
IsoStack: highly efficient network processing on dedicated cores
USENIXATC'10 Proceedings of the 2010 USENIX conference on USENIX annual technical conference
Hi-index | 0.00 |
This paper explores the effect of the current generation of hardware support for IP storage area networks on application performance. In this regard, this paper presents a comprehensive analysis of three competing approaches to build an IP storage area network that differ in their level of hardware support: software, TOE (TCP Offload Engine) and HBA (Host Bus Adapter). The software approach is based on the unmodified TCP/IP stacks that are part of a standard operating system distribution. For the two hardware-based approaches (TOE, HBA), we experimented with a range of adapters and chose a representative adapter for the current generation of each of the hardware approaches.The micro-benchmark analysis reveals that while hardware support does reduce the CPU utilization for large block sizes, the hardware support can itself be a performance bottleneck that hurts throughput and latency with small block sizes. Furthermore, the macro-benchmark analysis demonstrates that while the current generation of the hardware approaches may have the potential to provide performance improvements in CPU-intensive applications, overall the analysis does not demonstrate any performance benefits in database, scientific and email benchmarks. The analysis in this paper points out that a disparity in the processing power between the host and the adapter is the primary cause of the performance bottleneck in the current generation of the hardware approaches. The paper aims to guide the designers of the next generation of hardware-assisted adapters to better leverage the increasing processing power in the host. In particular, future adapters should be capable of handling small operations at wire speed.