This article was accepted into the corpus but its outbound wikilinks were never NER-processed — typical at the deepest BFS hop or when the run's entity cap was reached. No expansion funnel to show.
| Serial Attached SCSI | |
|---|---|
| Name | Serial Attached SCSI |
| Abbreviation | SAS |
| Developer | T10 Committee |
| Initial release | 2004 |
| Type | Storage interface |
Serial Attached SCSI is a point-to-point serial protocol for connecting and transferring data between storage devices and controllers, designed as the successor to parallel SCSI. It is used in enterprise EMC, Dell, HP and IBM storage arrays, supported by vendors such as Intel and Broadcom, and standardized by the INCITS T10 committee under the auspices of ANSI.
Serial Attached SCSI provides a high-performance, full-duplex, serial transport for block I/O between initiators and targets, intended for use in server, SAN, and data center environments alongside technologies from Cisco, NetApp, Hitachi, and Fujitsu. The technology addresses limitations of parallel interfaces seen in products from Seagate, Western Digital, and legacy Adaptec controllers by offering point-to-point links, improved signal integrity, and simplified cabling favored in deployments by Rackspace and Equinix. Major industry adopters include Apple (in select servers historically), hyperscalers like Google and Amazon Web Services, and research centers such as CERN.
The architecture defines endpoints such as host bus adapters (HBAs), expanders, edge devices, enclosures, and end devices produced by vendors like LSI (now part of Broadcom), Microchip, and Marvell. Key components include the host initiator, SAS expander fabric used in arrays from NetApp and Dell EMC, and end devices including disk drives by Seagate and Western Digital or solid-state devices by Samsung and SK Hynix. Enclosure management often uses standards linked to organizations such as SNIA and interfaces implemented by firms like Supermicro and Tyan for server integration.
SAS uses layers including PHY, link, transport, and application modeled and specified by the T10 committee and interoperates with command sets derived from SCSI standards referenced by the IETF for related networking protocols. Physical signaling employs differential serial lanes comparable in concept to technologies from PCI-SIG and leverages electrical characteristics validated by test suites from Tektronix and Keysight Technologies. The protocol supports frame-based SAS primitives, link negotiation, and discovery procedures similar in management scope to systems from Red Hat and Microsoft in server stacks.
SAS generations (e.g., 3.0, 4.0) offer successive link rates promoted by vendors Intel, Broadcom, and Marvell and deployed in environments run by Facebook and LinkedIn; these generations scale throughput, IOPS, and link aggregation techniques used in storage arrays from Hitachi and NetApp. Topologies using expanders allow large device counts suitable for hyperscale deployments at Microsoft Azure and enterprise SANs at Oracle, while multi-lane and wide-port configurations parallel bandwidth scaling strategies seen in NVIDIA interconnect roadmaps. Benchmarking by labs such as SPEC and vendors like Dell EMC demonstrate latency, queue depth, and throughput tradeoffs versus alternatives like NVMe.
Common implementations include SAS HBAs in servers from HPE and Lenovo, storage enclosures by Promise Technology and Supermicro, and backup targets in solutions by Commvault and Veeam. Use cases span transactional databases in deployments by SAP and Oracle, virtualization platforms from VMware and Microsoft Hyper-V, and large-scale archiving in institutions such as NIH and media companies like Disney and Netflix. Hybrid arrays combining SAS HDDs and SSDs are offered by Pure Storage partners and integrated into cloud-provider cold storage tiers at Backblaze.
SAS preserves command compatibility with parallel SCSI command sets standardized by T10 and coordinates interoperability with SATA devices through integrated compatibility layers adopted by Samsung, Western Digital, and Seagate drives. Industry interoperability testing involves consortia including SNIA and manufacturer programs run by Broadcom, Intel, and Microchip to ensure cross-vendor compatibility in mixed-drive enclosures from Supermicro and backplane vendors like Amphenol. Firmware and driver ecosystems include offerings by Red Hat, Microsoft, and vendor-specific stacks for enterprise OSes used by Oracle Linux and SUSE.
Development was led by the T10 committee under INCITS and ratified through ANSI processes, with successive revisions aligning with work from vendors such as LSI, Adaptec, Intel, and Seagate. Key milestones include the transition from parallel SCSI used by manufacturers like Sun Microsystems to serialized links adopted by enterprise arrays from EMC and standard updates published alongside related standards from IEEE and coordination with storage bodies such as SNIA. Profiles and conformance suites have been published and iterated in collaboration with corporate stakeholders including IBM and Hitachi.
Reliability features incorporate link-level error detection, power-loss protection strategies used by Seagate and Western Digital enterprise drives, and RAID integrations provided by vendors like DELL EMC and NetApp to ensure data integrity. Security practices in SAS deployments intersect with firmware signing and device authentication efforts advanced by Intel and platform security initiatives from Microsoft and Red Hat; operational reliability is supported by monitoring tools from SolarWinds and diagnostic suites from HPE and Dell Technologies.
Category:Computer storage interfaces