The present embodiments relate generally to auditing the power consumption in a data storage system as individual utilizations of data storage elements vary and more particularly, but without limitation, to reducing the power output of a power supply unit during intermittent delays in data transfer processing with the selected storage elements.
The combination of multiple storage devices into distributed data storage capacity has proliferated in response to market demands that enormous amounts of data be readily available in a fast, reliable, and efficient manner. Although types of individual storage elements vary, some things are common like electrical components that not only advantageously perform desired functionalities but also consume significant amounts of energy and generate heat. Typically, multiple storage element arrays require some type of cooling devices, from simple fans to complex refrigeration cooling systems, to remediate the thermal load.
With continued demands for ever increased levels of storage capacity and performance, there remains an ongoing need for improvements in the manner in which the storage elements in such data storage arrays are operationally powered to optimize energy usage and heat generation. It is to these and other improvements that preferred embodiments of the present invention are generally directed.
Some embodiments of the present invention contemplate a data storage system having an enclosure containing a plurality of drives that are individually selectable to transfer data corresponding to an execution of input/output (I/O) commands between the data storage system and another device. A memory in the enclosure temporarily stores unexecuted I/O commands, and a power supply device is capable of simultaneously operating all of the plurality of drives in support of multiple transfers of data. A power management device operably reduces a power output of the power supply device in response to a forecasted interruption in the transfer of data with one of the drives at a time when an unexecuted I/O command for the one of the drives resides in the memory.
Some embodiments of the present invention contemplate a method that includes: operating a data storage system having a plurality of drives that are individually selectable to transfer data corresponding to an execution of I/O commands between the data storage system and another device, a power supply device capable of simultaneously operating all of the plurality of drives in support of multiple data transfers of data, and a memory temporarily storing unexecuted I/O commands; forecasting an interruption in the transfer of data with one of the drives at a time when an unexecuted I/O command for the one of the drives resides in the memory; and in response to the forecasting, reducing a power output of the power supply device.
Some embodiments of the present invention contemplate a data storage library having a frame supporting a shelf system to queue a plurality of magazines and a plurality of tape cartridges each removably supported by one of the plurality of magazines. A plurality of drives is each adapted to engage one of the tape cartridges at a time in a data transfer relationship. A transport system selectively moves the tape cartridges between the queue in the shelf and the data transfer relationships in one of the plurality of drives. A memory within the frame temporarily stores I/O commands driving the data transfer relationships between each of the drives and the tape cartridges. A power supply device within the frame is capable of operating all of the plurality of drives simultaneously in support of the data transfer relationships. Computer code stored in memory is implemented to reduce a power output of the power supply device in response to a forecasted interruption in the data transfer relationship with one of the drives at a time when a pending I/O command for the one of the drives resides in the memory.
Initially, it is to be appreciated that this disclosure is by way of example only, not by limitation. The energy audit concepts herein are not limited to use or application with any specific system or method for using storage element devices. Thus, although the instrumentalities described herein are for the convenience of explanation, shown and described with respect to exemplary embodiments, it will be appreciated that the principles herein may be applied equally in other types of storage element systems and methods involving the storage and retrieval of data.
To illustrate an exemplary environment in which preferred embodiments of the present invention can be advantageously practiced,
Each array 104 preferably includes a pair of controllers 108 (denoted A1, A2 and B1, B2) for redundancy sake, and a set of data storage devices 110. It is further contemplated that in some embodiments the A client 102 and the A data storage array 104 can be physically located at a first site, the B client 102 and B storage array 104 can be physically located at a second site, and the C client 102 can be yet at a third site, although such is merely illustrative and not limiting.
A fabric interface (I/F) 120 communicates with the dual controllers 108 and the clients 102 via the fabric 106, and a drive I/F 122 communicates with the drives 112. The I/F circuits 120, 122 and a path controller 124 form a pass-through communication path for commands and data between the storage array 104 and the client(s) 102, such as by employing the cache memory 116. Again, although illustrated discretely, it will be understood that each path controller 124 and the corresponding I/F circuits 120, 122 can be unitarily constructed.
The power management device 128 advantageously reduces the power output of the PSU 126 when one or more of the drives 112 is predictably not being utilized or is being utilized in a way resulting in a reduced power consumption. For example, the power management device 128 executes computer instructions stored in memory to recognize a forecasted interruption in the respective transfer of data with a selected one or more of the drives 1121, 1122, . . . 112n at a time when it is also known from the cache manager 130 that ultimately, after the interruption, I/O command processing with the selected one or more drives 1121, 1122, . . . 112n will resume. An indication that I/O command processing with the selected one or more drives 1121, 1122, . . . 112n will ultimately resume can be that an unexecuted I/O command for the selected one or more drives 1121, 1122, . . . 112n resides in the cache 116 (
The total number of drives 112 that are powered up at any given time can be parametrically determined in a way that optimizes data throughput and power usage. For an extreme example, if there is only one pending I/O command in cache 116 it would be a waste of energy, and concomitant needless excess generated heat, to power more than one drive 112 to execute that one pending I/O command. In that event the PSU 126 can be derated to output only enough power to operate the one drive 112 in accordance with embodiments of the present invention.
In that same vein of reasoning, if some data transfers can be delayed to an off-peak time, such as during the middle of the night, then it can be acceptable to execute those I/O commands at a slower data throughput rate than what is required during the normal office hours when users otherwise expect data transfers to occur as instantly as possible. For example, where redundant storage methodologies are employed such as a redundant array of independent tapes (RAIT), redundant copies of primary data can at least to some extent be held in a nonvolatile memory in the array 104 and transferred via the drives 112 during off-peak operational time periods. Ultimate control of time shifting of the I/O command load can be governed by rules 132 that can be empirically derived by characterizing the I/O command load over time, and/or the rules 132 can be user-defined by a graphical user interface allowing the user to set and adjust the governing parameters.
Each of the tape cartridges 136 is selectively loadable into one of the drives 112 to cooperatively form an operable data transfer relationship to store data to and/or retrieve data from the tape cartridge 136. Each drive 112 can have a MAM device reader/writer 140 to store data to and/or retrieve data from the MAM device. In these illustrative embodiments the drive 112 establishes wireless communications 142 with the MAM device, such as by radio frequency communication, although neither the disclosed embodiments nor the claimed embodiments are so limited to those illustrative embodiments. The MAM device data can advantageously include access occurrence data, such as timestamp data indicating when the tape cartridge 136 is loaded in a drive 112, load count data indicating how long a tape cartridge 136 is loaded in the drive 112, validity data indicating any data and/or portions of the storage medium in a tape cartridge 136 of questionable integrity, and the like. Besides, or in addition to, storing data on the MAM devices, a larger system memory 144 can accommodate information, such as the access occurrence data, load data, validity data, and the like, from each of a plurality of MAM devices associated with respective tape cartridges 136. Computational routines on the data stored in the MAM devices and in the system memory 144 can be under the top-level control of a central processing unit (“CPU”) 146. A graphical user interface (“GUI”) 147 provides helpful tabular and graphical information to a user of the tape library for providing inputs thereto and receiving useful outputs therefrom.
The tape library can advantageously have a shelving system 148 capable of processor-based archiving the magazines 138 within the tape library. A transport unit 150 shuttles magazines 138 between the shelving system 148 and the drives 112, and picks and places a particular tape cartridge 136 from a shuttled magazine 138 to/from a desired drive 112. Again, although
The tape library is not necessarily limited to using a fixed number of tape cartridges 136. Rather, an access port 152 is configured to cooperate with an external transport system (not shown) to deliver or remove individual tape cartridges 136 or magazines 138.
Top-level control is provided by the CPU 146 in communication with all the various components via a computer area network (not shown). Data, virtual mappings, executable computer instructions, operating systems, applications, and the like are stored to the system memory 144 and accessed by one or more processors in and/or under the control of the CPU 146. The CPU 146 includes macroprocessors, microprocessors, memory, and the like to logically carry out software algorithms and instructions.
As one skilled in the art will recognize, the illustration of the tape library in
In the tape library of
In block 166 it is determined whether any forecast exists of an upcoming interruption of data transfer capability with a drive, such as due to the need to change out a tape cartridge in a selected drive with another tape cartridge as discussed. If the determination of block 166 is “no,” then control returns to block 162 for further execution of the pending I/O commands. If, however, the determination of block 166 is “yes,” then in block 168 the power management device converts the forecasted interruption to a power savings value calculated in terms of operating the PSU during at least a portion of the interruption at a reduced power output level. For example, without limitation, the power management device can calculate the power savings resulting from operating the PSU at a predetermined reduced power mode, producing less power than the maximum output mode but nonetheless making the drive fully capable of all operations necessary to change out the existing tape cartridge with another tape cartridge. In equivalent alternative embodiments the power management device can calculate the reduced power mode in terms of even lower predetermined power output modes such as a minimal output sleep mode or even a power off mode. Generally, the more power savings that are gained from the reduced power mode, the longer will be the recovery period for returning the drive to data transfer duty after the replacement tape is mounted.
In block 168 the power management device then compares the calculated power savings value ΔP1 to a predetermined first threshold value T1. If the power savings value is greater than the first threshold value then the power management device reduces the power output mode of the PSU to the predetermined reduced power mode in block 170, such as the power-off mode depicted, for an interval of time associated with the forecasted interruption. Preferably, the interval of time during which the power supply is operated at the reduced power mode is correlated to the forecasted interruption duration so that the PSU is returned to the appropriate operational capability before the replacement tape is fully made ready to transfer data again. For that purpose it can be advantageous to calculate the power saving mode in terms of a fractional portion of the interruption duration, either empirically or manually set and/or adjusted, such as calculating the power savings value over 90% of the entire interruption interval.
If, on the other hand, the comparison in block 168 concludes that the power savings value is not greater than the first threshold, then control passes to block 172 where the power management device converts the forecasted interruption to a second power savings value different than the first. The second power savings value is also a reduced power mode, but one that produces comparatively more power than the other reduced power mode for which the first power savings was calculated. For example, without limitation, in the illustrative embodiments of
In block 172 the power management device then compares the calculated power savings value ΔP2 to a predetermined second threshold value T2. If the power savings value is greater than the second threshold value then the power management device derates the PSU to the reduced power mode in block 174 for an interval of time associated with the interruption. As above, the interval of time during which the PSU is operated at the reduced power mode is preferably correlated to the forecasted interruption duration so that the PSU is returned to full operational capability before the replacement tape is made fully ready to transfer data again; such as calculating the power savings value over 90% of the entire interruption interval. In either event, when the replacement tape is ready in block 176 the power management device resumes power control post-interruption by control looping back to block 162 and the process continuing.
As discussed previously, the power management device 128 has visibility of the I/O load from the cache manager 130. The power management device 128 correlates one or more command queues of the I/O commands to the corresponding drives 112 to which they are related to produce a map 180 of an ongoing forecast of the utilization of each drive 112. For example,
The map 180 depicts by “Xs” those intervals of time during which an interruption is forecast to occur with one of the drives 112, such as during the time that a tape cartridge is switched out as discussed before. For example, the map 180 informs the power management device 128 that such an interruption will occur with drive D11 during the time interval t4-t6. The power management device 128 can calculate the power savings potential as discussed above from derating PSU1 from an output capacity of four drives 112 to a lower predetermined output capacity of only the three drives (D12, D13, D14) that will be required during t4, and the further power savings potential to an even lower predetermined output capacity of two drives (D13, D14) that will be utilized in the set of drives 1121 during t5-t6.
The power management device 128 performs the same power savings potential analysis for the other set of drives 1122 which likewise only needs to power two drives (D21, D24) during the interval t5-t6. In some embodiments, where the power savings threshold analysis is satisfied, the power management device 128 can derate both PSU1 and PSU2 from maximum power output to a predetermined reduced power mode for powering only two drives 112 instead of four drives 112. However, in the event that each PSU operates most efficiently at maximum output load then the power management device 128 powers off one of the two PSUs in this case and enables the other PSU, such as PSU1 depicted in
The example above that each PSU operates most efficiently at maximum loading is illustrative and not limiting of the contemplated embodiments of this invention. Generally, the loading at which any particular PSU operates most efficiently will either be known from the manufacturer's specifications or it can be empirically determined by measuring the ratio of input power (Pin) to output power (Pout) over the range of possible different loadings. If, for example in alternative equivalent embodiments, the PSU is known or observed to be most efficient when powering three of the four drives (¾ loading), then the power supplies would be managed as above but for aiming to keep all the powered PSUs operating at ¾ load for maximum operating efficiency.
Embodiments of the present invention can be commercially practiced in a Spectra Logic T-950 tape library manufactured by Spectra Logic of Boulder Colo.
It is to be understood that even though numerous characteristics and advantages of various embodiments of the present invention have been set forth in the foregoing description, together with the details of the structure and function of various embodiments of the invention, this disclosure is illustrative only, and changes may be made in detail, especially in matters of structure and arrangement of parts within the principles of the present invention to the full extent indicated by the broad general meaning of the terms in which the appended claims are expressed. For example, multiple or even predetermined pluralities of tape drives can be managed in the energy audit process for example, while still maintaining substantially the same functionality without departing from the scope and spirit of the claimed invention. Another example can include using these techniques across multiple library partitions, while still maintaining substantially the same functionality without departing from the scope and spirit of the claimed invention. Further, though communication is described herein as between a client and the library communication can be received directly by a tape drive, via the interface device 120, for example, without departing from the scope and spirit of the claimed invention. Further, for purposes of illustration, a first and second tape drive and tape cartridges are used herein to simplify the description for a plurality of drives and tape cartridges. Finally, although the preferred embodiments described herein are directed to tape drive systems, and related technology, it will be appreciated by those skilled in the art that the claimed invention can be applied to other systems, without departing from the spirit and scope of the present invention.
It will be clear that the claimed invention is well adapted to attain the ends and advantages mentioned as well as those inherent therein. While presently preferred embodiments have been described for purposes of this disclosure, numerous changes may be made which readily suggest themselves to those skilled in the art and which are encompassed in the spirit of the claimed invention disclosed and as defined in the appended claims.
It is to be understood that even though numerous characteristics and advantages of various aspects have been set forth in the foregoing description, together with details of the structure and function, this disclosure is illustrative only, and changes may be made in detail, especially in matters of structure and arrangement to the full extent indicated by the broad general meaning of the terms in which the appended claims are expressed.