IP-networked storage protocols such as NFS and iSCSI have become increasingly common in today 's LAN environments. In this paper, we experimentally compare NFS and iSCSI performance for environments with no data sharing across machines.

Data storage plays an essential role in today's fast-growing data-intensive network services. New standards and products emerge very rapidly for networked data storages. Given the mature Internet infrastructure, overwhelming preference among IT community recently is using IP for storage networking because of economy and convenience. iSCSI is one of the most recent standards that allows SCSI protocols to be carried out over IP networks. However, there are many disparities between SCSI and IP in terms of speeds, bandwidths, data unit size, and design considerations that prevent fast and efficient deployment of SAN (Storage Area Network) over IP. This paper introduces STICS (SCSI-To-IP Cache Storage), a novel storage architecture that couples reliable and highspeed data caching with low-overhead conversion between SCSI and IP protocols. A STICS block consists of one or several storage devices such as disks or RAID, and an intelligent processing unit with CPU and RAM. The storage devices are used to cache and store data while the intelligent processing unit carries out caching algorithm, protocol conversion, and self-management functions. Through efficient caching algorithm and localization of certain unnecessary protocol overheads, STICS can significantly improve performance, reliability, manageability, and scalability over current iSCSI systems. Furthermore, STICS can be used as a basic plug-and-play building block for data storage over IP. Analogous to "cache memory" invented several decades ago for bridging the speed gap between CPU and memory, STICS is the first-ever "cache storage" for bridging the gap between SCSI and IP making it possible to build efficient SAN over IP. We have carried out a partial implementation and simulation experiments to study the performance ...

This paper presents our thesis that the advantages of the internet protocol framework make it the best choice for communications protocols to, and between, networkattached peripherals (NAPs). Moreover, the IP suite is more appropriate than the specialized protocol stacks being developed for commercial NAPs. The benefits of using IP include support for heterogeneous network media, wide-area connectivity, and reduced research and development effort. We examine the issues for use of the internet protocols (TCP/UDP/IP) for NAPs, address commonly held concerns regarding its performance, and describe the Netstation project’s prototype implementations of IP peripherals. 1

In many clusters today, the local disks of a node are only used sporadically. This paper describes the software support for sharing of disks in clusters, where the disks are distributed across the nodes in the cluster, thereby allowing them to be combined into a high-performance storage system. Compared to centralized storage servers, such an architecture allows the total I/O capacity of the cluster to scale up with the number of nodes and disks. Additionally, our software allows customizing the functionality of the remote disk access using a library of code modules.

We propose to make nodes in a cluster double as compute nodes and network attached storage (NAS) nodes. This allows for a flexible and customizable storage system as the NAS control software is handled by regular workstations. The nodes can still be efficient compute nodes if networks with remote DMA capabilities are used, as such networks remove the processor from the data forwarding loop. We demonstrate this through measurements of a prototype implementation.

This paper concerns the sharing of storage devices in clusters, where the storage devices are distributed across the nodes in the cluster. Compared to centralized storage servers, such an architecture allows the total I/O capacity of the cluster to scale with the number of nodes and disks. The Scalable Coherent Interface (SCI) cluster interconnect allows such storage devices to be shared efficiently, as the processor of the node hosting the disk need not participate in the actual data transfer. Thus, SCI makes it feasible for the nodes in the cluster to act both as compute nodes and as flexible network-attached storage nodes.

file object NAS SAN NAS File system data structures (files and directories) Nonvolatile storage volume (blocks) "myfile" "bob" NASD esse molestaie cons ipscing elit, sed diam n eisand tempor inciden NASD ielt esse molestaie son cons ipscing diam my eisand tempor inciden NASD vel eum irure dolor prehenderit voluptate esse molestaie cons equat illium dolore t nulla pariatur.Lorem ipse amet, consecte pscing elit, sed diam nonu ielt esse molestaie son cons nulla pariatur.Lorem sit amet, consectetur ipscing elit, sed diam derit voluptate veielt quat illium dolore fugi nulla pariatur.Lore sem amet, consetetur adipscing elit, sed derit voluptate veielt quat vel illium dolore fugi nulla pariatur.Lore sem sit amet, consetetur adipscing elit, esse molestaie son cons t nulla pariatur.Lorem amet, consectetur ipscing elit, sed diam eum irure dolor prehenderit voluptate ve ielt esse molestaie son cons equat illium dolore nulla pariatur.Lorem ipse amet, consecte pscing elit, sed...

The iSCSI protocol provides access to SCSI devices over a TCP transport. Using the iSCSI protocol enables one to build a Storage Area Network using standard Ethernet infrastructure and standard networking management tools. This paper outlines how we implemented a family of iSCSI initiators on a common core. The initially supported initiators were on the Windows NT and the Linux Operating Systems. Code for a version of the Linux iSCSI initiator has been released as Open Source. Initial testing indicates that iSCSI can provide reasonable performance relative to traditional storage environments.

This paper presents our thesis that the advantages of the internet protocol framework make it the best choice for communications protocols to, and between, network-attached peripherals #NAPs#. Moreover, the IP suite is more appropriate than the specialized protocol stacks being developed for commercial NAPs. The bene#ts of using IP include support for heterogeneous network media, wide-area connectivity, and reduced research and development e#ort. We examine the issues for use of the internet protocols #TCP#UDP#IP# for NAPs, address commonly held concerns regarding its performance, and describe the Netstation project's prototype implementations of IP peripherals.

This paper revisits the historical concept of link-layer, block-oriented storage networking in light of commoditybased, high-bandwidth local area networks. It examines how advances in data communication technology such as virtual LANs and switched gigabit Ethernet make it relatively simple and inexpensive to re-centralize the storage resources used by clusters of workstations and servers. In particular, the work is motivated by an interest in how to support scalable and efficient access to read-only and private read-write data such as root file systems, swap partitions, log files and static web pages. These techniques complement, but do not replace, higher level distributed file systems whose primary goal is to provide coherent access to shared read/write data. This paper describes the design and implementation of a very simple, link-layer protocol, the Ethernet Block Device (EBD), for accessing remote block devices. It compares the EBD prototype to a locally attached disk and to similar, network-based techniques that use TCP/IP such as the Linux Network Block Device (NBD), as well as higher level distributed file systems such as NFS. Functionally, the implementation is compared with a local disk to determine what restrictions and visible differences exist. The performance evaluation is based on a series of standard disk and file access benchmarks run on commodity servers and standard networking equipment. The performance results show that for large sequential reads and random reads and writes EBD generally outperforms comparative network storage technologies by 15% to 30%, and performance is best when the data set fits into the server's memory. On a benchmark that is very metadata-intensive and does small sequential operations, EBD and NBD perform similarly. On certain...