Various embodiments of the present disclosure are generally directed to an apparatus for retracting and extending sets of operational processing devices in a multi-device enclosure.
In accordance with some embodiments, an enclosed housing is provided with opposing first and second ends. Sleds are individually movable between a retracted position within the enclosed housing and an extended position in which the sled projects from the first end. Each sled supports a group of processing devices. A control board is disposed within the enclosed housing adjacent the second end. A plurality of flex circuits contactingly engages the processing devices to provide communication paths between the processing devices and the control board in both the retracted and extended positions of the sleds.
The present disclosure generally relates to an enclosure configuration that facilitates hot swapping of processing devices within the enclosure. The various embodiments can be implemented in a networked mass storage enclosure with a plurality of data storage devices, or other processing environments. In some cases, the processing environment may be a distributed object storage system for a cloud computing environment.
Mass storage systems often employ multiple data storage devices which are operationally arranged to provide a relatively large data capacity memory storage space. The devices may be grouped together into a storage enclosure or other module that can be removably installed into a rack system (e.g., a server cabinet, etc.).
Mass storage systems can take a variety of forms including servers, cloud storage modules, RAID (redundant array of independent discs) systems, object storage, extended memory systems (JBODs, or “just a box of drives”), point-to-point link systems, etc. The storage systems can be accessed locally or over a network including a local area network (LAN), a wide area network (WAN), the Internet, etc.
While operable to provide highly efficient computer storage, existing storage systems can be subject to a variety of limitations including the inability to remove and replace individual storage devices while maintaining the storage enclosure in a powered, operational condition (“hot swapping”).
Accordingly, various embodiments of the present disclosure are generally are directed to an apparatus and method for hot swapping individual processing devices within a multi-device storage enclosure. As explained below, in some embodiments a storage enclosure has a housing adapted to be mounted within a rack system between a cold aisle (front) and a warm aisle (back). The housing supports a number of components including multiple storage devices, fans, power supplies, control boards, etc.
Sets of the storage devices are mounted within retractable and extendable sleds, or trays, that are accessible from a single side of the storage enclosure such as the front (cold aisle) of the storage enclosure. Electrical interconnections with the storage devices are carried out via flexible conductive paths (referred to herein as “flex circuits”). This allows the individual sleds to be extended from the storage enclosure housing while the storage devices remain powered and in an operational state. One or more individual storage devices can be removed and replaced from an extended sled while the remaining storage device or devices remain powered and operational.
In some cases, a sled having multiple operational storage devices can be extended from a storage enclosure. At least one of the storage devices can be deactivated, removed and replaced with a new, replacement storage device while one or more of the remaining multiple operational storage devices in the sled remain in an operational state to service data access (e.g., read and write) operations with a host. The sled can then be retracted back into the storage enclosure.
The flex circuits can take a variety of forms, such as planar flex circuits (conductive traces on a flexible insulative backing layer), flexible wires, cables, jacketed cables, ribbon cables, etc. The flex circuits are arranged to maintain continuous electrical and mechanical interconnection with each of the multiple storage devices in each sled so that the non-affected storage devices continue to operate irrespective of whether the associated sled is extended or retracted.
In this way, service operations such as replacements and upgrades can be carried out from a selected side of the storage enclosures, such as from the cold aisle side of a storage facility, without the need to take the storage devices to an off-line and/or powered down condition. It is also generally unnecessary to remove the storage enclosure from the server cabinet or other installation location. As desired, individual device electronics such as indicator lights (LEDs, etc.), power switches, etc. can be provided for each device and incorporated into the sleds and/or the housing to allow user selection and control of individual storage devices.
While data storage devices are a particularly useful type of processing device to which the various techniques discussed herein can be applied, it will be appreciated that other processing devices, such as electronic systems, controllers, boards, operational modules, etc. can also benefit from the various applied techniques and are included within the present disclosure.
These and other features can be understood beginning with a review of
In some embodiments, the storage rack 108 is a 42U server cabinet with 42 units (U) of storage, with each unit comprising about 1.75 inches (in) of height. The width and length dimensions of the cabinet can vary but common values may be on the order of about 24 in.×36 in. Each storage enclosure can be a multiple of the storage units, such as 2U. As explained below, fully populating the rack 108 with storage enclosures 110 can provide well over 1 Petabyte (1015 bytes) of storage for the computer 104 and/or network applications.
A number of example configurations for the storage enclosures 110 are set forth in
The storage enclosure 110A includes dual power supplies 116, multiple airflow fans 118 and at least one controller board 120. The power supplies 116 provide electrical power for the storage enclosures 110. The fans 118 draw airflow from openings (not separately shown) in a front facing side 122 of the housing 114 and pull the airflow through the housing and out openings (not shown) in a rear facing side 124 of the housing.
The controller 120 may have one or more intelligent processors 126 and can take a variety of configurations including but not limited to a server, a controller (including dual redundant RAID controllers), a cloud controller, dual port controllers, an Ethernet drive controller, etc. Other support electronics and components can be provided in the storage enclosure 110A as well, including a boot storage memory device, storage control boards, device interface (I/F) boards, wiring cables, switches, brackets, LED indicators, etc. A redundant design may be provided so that substantially any failed component can be replaced while maintaining the storage enclosure 110A in an operational condition.
An alternative configuration for the storage enclosures is shown at 110B in
The storage devices 112 can take a variety of processing device configurations, such as but not limited to a hard disc drive (HDD), a solid state drive (SSD), a hybrid drive (solid state hybrid drive, SDHD), etc.
In the context of an HDD, the storage media 145 may take the form of one or more axially aligned magnetic recording discs which are rotated at high speed by a spindle motor. Data transducers can be arranged to be controllably moved and hydrodynamically supported adjacent recording surfaces of the storage disc(s). While not limiting, in some embodiments the storage devices 112 are 3½ inch form factor HDDs with nominal dimensions of 5.75 in×4 in×1 in.
In the context of an SSD, the storage media 145 may take the form of one or more flash memory arrays made up of non-volatile flash memory cells. Read/write/erase circuitry can be incorporated into the storage media module to effect data recording, read back and erasure operations. Other forms of solid state memory can be used in the storage media including magnetic random access memory (MRAM), resistive random access memory (RRAM), spin torque transfer random access memory (STRAM), phase change memory (PCM), in-place field programmable gate arrays (FPGAs), electrically erasable electrically programmable read only memories (EEPROMs), etc.
In the context of a hybrid (SDHD) device, the storage media 145 may take multiple forms such as one or more rotatable recording discs and one or more modules of solid state non-volatile memory (e.g., flash memory, etc.). Other configurations for the storage devices 112 are readily contemplated, including other forms of processing devices besides devices primarily characterized as data storage devices, such as computational devices, circuit cards, etc. that at least include computer memory to which secure erasure processing is applied.
The sled 115 in
The sleds can be retracted automatically or manually. For example, the sleds can have a rail and roller mechanism similar to a “drawer” support used in a dresser or desk to allow the sleds to extend from the enclosure, and a spring-biased locking/unlocking release mechanism so that a user can press a facing surface 172 of the sled in, allowing the sled to pop out a short distance and enabling the user to grasp the front of the sled and retract it from the housing 114. In other embodiments, a motor driven carriage (not separately depicted) can be used to drive the sled to an open position, in a manner similar to an optical disc tray in a personal computer. Extension and retraction of the sleds can additionally/alternatively be carried out via commands entered by the user in the local computer 104 (
It will be noted that rail kits are known in the art to enable entire storage enclosures to be extended and retracted relative to a server cabinet; such kits can be used in accordance with the storage enclosures 110 of the present disclosure to allow extension thereof from the server cabinet 108 (see
A flex circuit arrangement can be used to enable the sleds to be extended and retracted while maintaining the devices 112 in a powered up, connected condition.
The flex circuits 180, 182 in
The flex circuits 180, 182 operate in a manner similar to an “extension cord” in that the devices 112 can be moved while still maintaining power and communication connections with the control board 120 (
Another feature of the use of the flex circuits 180, 182 is the elimination of a main backplane or other rigid power board to interconnect the various devices 112. This reduces weight, improves interior airflow and cooling pathways, and simplifies the design including allowing for the installation of upgraded devices into the enclosure (including new devices with a different interface). Connectors 184, 186 mate with the devices and connectors 188, 190 interconnect with the control board 120.
The number and placement of the respective device connectors 188, 190 along each flex circuit 180, 182 will depend on the number, position and orientation of the devices 112 in each sled 115. It will be appreciated that the flex circuits may be rotated, folded or otherwise routed for clearance purposes so as to fit within the associated sleds and facilitate retraction and extension thereof while maintaining connection with the devices.
A general interconnection between the motherboard 210 and the storage control board 212 is represented by path 216, and the interconnection between the storage control board 212 and the drive I/F board is represented by path 218. In some cases, the three boards 210, 212 and 214 are placed in the rear of the storage enclosure 110 and a flex circuit 220 extends from an associated drive I/F board 214 to each storage device 112. In other cases, the drive I/F boards 214 may be located within the individual sleds 115, in which case the interconnection path 218 may alternatively or additionally be realized as a flex circuit as discussed above.
The hinge arm carrier assembly 230 includes at least two rigid hinge arms 232, 234 adapted to secure one or more flex cables 235 (see
The hinge arms 232, 234 can be articulated via joint members 236, 238 and 240. This allows extension of the distal joint member 240 as depicted in
In this way, the hinge arm carrier assembly 230 folds in a substantially flat condition when the sled 115 is retracted (
The flex circuit 260 comprises four (4) flex circuit segments 262A, 262B, 262C and 262D each terminating in a connector 264A, 264B, 264C and 264D adapted to physically interconnect with a corresponding connector on a respective data storage device 112. The flex circuit segments are contemplated as comprising flexible insulative substrates on which a series of spaced apart conductive traces are formed. The flex circuit segments may be single sided, double sided, laminated (multi-layer), etc. and may support control circuitry (not separately shown).
The flex circuit segments 262A, 262B, 262C and 262D are mechanically and electrically connected to a flex circuit extension 266 by way of an intervening interconnection member 268. The flex circuit extension 266 also comprises a flexible insulative substrate on which a series of spaced apart conductive traces are formed. The flex circuit extension 266 terminates at a connector 270 which physically interconnects with a corresponding connector on an I/F board 272 or other control signal mechanism.
The flex circuit segment 262D from
The hinge arm carrier 280 includes rigid arms 284, 286 which are connected by an intermediate joint member (elbow joint) 288 so that the arms may rotate relative to one another. A base joint member 290 is secured to the enclosure housing 114 (see e.g.,
The arms 284, 286 are planar members and include a plurality of securement rings 294 which are spaced along the respective arms 284, 286. As depicted in
As will be appreciated by those skilled in the art, storage systems such as 300 can be arranged to store data in the form of accounts, containers and objects. Accounts refer to individual users/entities that store data within the system (“account holders”). Each container is a set of data from a particular account holder arranged in a useful grouping. The individual data sets are referred to as objects. The objects can take a variety of forms such as files or other data sets.
A proxy server 302 communicates with computer based systems of the account holders (account holder devices 304) via a network 306 to store object data in the storage system 300. Data structures are organized as an account ring 308, container ring 310 and object ring 312. Each of these structures have associated physical components including respective account, container and object servers 318, 320 and 322, and respective account, container and object storage 328, 330 and 332. Without limitation, the various server and storage functions can be provided by the storage enclosures 110 discussed herein.
Generally, the account ring 308 provides control information regarding containers, the container ring 310 provides control information regarding objects, and the object ring 312 points to the storage locations of individual objects.
All data sets of whatever type are replicated by the system so as to be physically located in different locations (zones). This is exemplified by
Because of the relatively large amount of overhead processing involved in replication, updating and accessing the individual objects, it can be seen that hot swapping of individual data storage devices 112 while enabling remaining devices 112 within a given storage enclosure 110 to remain active can enhance the overall efficiency and availability of the storage system 300. Thus, the cold-aisle accessibility of the individual storage devices 112 for replacement and rebuild operations as discussed herein can have particularly valuable use in distributed object storage systems such as 300.
Initially at step 402, a set of storage devices 112 are electrically and mechanically coupled to sleds 115 within the storage enclosure 110A. The HDDs may be installed onto supports 160 and lowered into associated tray housings 162, and electrical interconnections are made using flex circuits such as in
The storage enclosure 110A thereafter enters a period of operational service whereby data are stored to and retrieved from the storage devices 112 as required by the local computer 104 and/or remote host/user devices (not shown) which access the memory space defined by the devices 112 via the network 108.
At some point during the operation of the storage enclosure 110A, one or more of the storage devices will be powered down, as indicated by step 406. This can take place due to a variety of reasons, such as a failure condition being experienced by the device or a commanded power down by a user for replacement/upgrade purposes. A detected imminent failure or reduction in reliability based on monitored parameters may result in the decision to power down and replace the device. It is contemplated albeit not necessarily required that, for each powered down device, one or more of the storage devices in the same sled 115 will remain powered up and remain in an operational mode to continue servicing access operations.
The sled(s) which house the affected device(s) is/are extended at step 408 such as depicted in
A replacement device is installed at step 412 by inserting the device into the vacant support 160, attaching the power/interface connections, and lowering the support. The sled is thereafter retracted at step 414 to return the storage enclosure to the condition previously set forth at step 404. As desired, data may be copied over to the new replacement device by the system, step 416, and the routine ends at step 418.
It is contemplated that only one sled will be extended at a time, but the system can accommodate the concurrent extension of multiple adjacent sleds as desired. It will be appreciated that regardless of the flex circuit configuration, each flex circuit will be supplied with sufficient slack length to enable transition of the associated sleds between the extended position and the retracted position while maintaining electrical and mechanical interconnection to each of the storage devices in the associated sled.
The storage enclosures as embodied herein are suitable for use in cloud computing environments as well as a variety of other environments. The use of flex circuits and sleds enhance interior airflow and uniformity of temperature gradients from the cold aisle to the warm aisle side, and facilitate ease of replacement of the individual devices without the need to power down and/or take offline the rest of the storage enclosure.
While various embodiments have been presented in the context of mass storage, such is merely exemplary and is not limiting, as the various enclosures can find applicability in a variety of processing environments.
It is to be understood that even though numerous characteristics and advantages of various embodiments of the present disclosure have been set forth in the foregoing description, together with details of the structure and function of various embodiments thereof, this detailed description is illustrative only, and changes may be made in detail, especially in matters of structure and arrangements of parts within the principles of the present disclosure to the full extent indicated by the broad general meaning of the terms in which the appended claims are expressed.
The present application makes a claim of domestic priority to U.S. Provisional Patent Application No. 61/833,635 filed Jun. 11, 2013, the contents of which are hereby incorporated by reference.
Number | Name | Date | Kind |
---|---|---|---|
3977756 | Rodondi | Aug 1976 | A |
5443312 | Schluter | Aug 1995 | A |
6137759 | Ogiro | Oct 2000 | A |
6459571 | Carteau | Oct 2002 | B1 |
6496366 | Coglitore et al. | Dec 2002 | B1 |
6551113 | Nishiyama | Apr 2003 | B1 |
6772887 | Audibert | Aug 2004 | B2 |
6854605 | Wrycraft | Feb 2005 | B2 |
7305458 | Hsue et al. | Dec 2007 | B2 |
7418623 | Elliott et al. | Aug 2008 | B2 |
7536586 | Ahmadian et al. | May 2009 | B2 |
7568122 | Mechalke et al. | Jul 2009 | B2 |
7877626 | Piszczek et al. | Jan 2011 | B2 |
8010713 | Matumura et al. | Aug 2011 | B2 |
8014144 | Cheng | Sep 2011 | B2 |
8120922 | Randall | Feb 2012 | B2 |
8191841 | Jeffery | Jun 2012 | B2 |
8296406 | Kasperson et al. | Oct 2012 | B2 |
8787012 | Michael | Jul 2014 | B2 |
9095070 | Ross | Jul 2015 | B2 |
9148980 | Moss | Sep 2015 | B2 |
20030082941 | Kuller | May 2003 | A1 |
20030227753 | Wrycraft | Dec 2003 | A1 |
20060171110 | Li | Aug 2006 | A1 |
20070045479 | Nguyen | Mar 2007 | A1 |
20070070595 | Chen | Mar 2007 | A1 |
20070227756 | Doerr | Oct 2007 | A1 |
20070230110 | Starr | Oct 2007 | A1 |
20090231800 | Franz et al. | Sep 2009 | A1 |
20100110628 | Barrenechea | May 2010 | A1 |
20110100933 | Kitten | May 2011 | A1 |
20120012714 | Rubenstein | Jan 2012 | A1 |
20120325127 | Adrain | Dec 2012 | A1 |
20130050963 | Zhou | Feb 2013 | A1 |
20130077219 | Fu | Mar 2013 | A1 |
20130151769 | Childs et al. | Jun 2013 | A1 |
20130161277 | Augsburger | Jun 2013 | A1 |
20130219101 | Hansen et al. | Aug 2013 | A1 |
20130286626 | Chen | Oct 2013 | A1 |
20130341471 | Yang | Dec 2013 | A1 |
20140153186 | Liu | Jun 2014 | A1 |
20140158834 | Chen | Jun 2014 | A1 |
20140183306 | Chen | Jul 2014 | A1 |
20160095246 | Noland | Mar 2016 | A1 |
20160150667 | Xu | May 2016 | A1 |
Number | Date | Country | |
---|---|---|---|
20140362515 A1 | Dec 2014 | US |
Number | Date | Country | |
---|---|---|---|
61833635 | Jun 2013 | US |