Data reduction technologies play a critical role in environments to help reduce the amount of physical storage that is needed to save a dataset. Reducing the amount of data storage that is required not only reduces the number of drives that are required, but it also reduces the physical footprint of the solution. This ability helps reduce the total cost of ownership of the storage system. With Dell PowerStore, data efficiency techniques such as pattern matching, deduplication, and compression help to reduce the amount of data that is physically stored within the system. These data efficiency methods are always on, and their data reduction achievements can be viewed within various resources of the system.

This white paper discusses the data efficiency techniques and includes technical information about the data path through the system and the underlying efficiency technologies. Data efficiency savings and reporting is also discussed.

This white paper is intended for IT administrators, storage architects, partners, and Dell Technologies employees. It is also intended for any other individuals that are involved in the evaluation, acquisition, management, operation, or design of a Dell networked storage environment using PowerStore.

Dell Technologies and the authors of this document welcome your feedback on this document. Contact the Dell Technologies team by email.

Author

Ryan Poulin

Large amounts of data are created by applications daily and ensuring that data is saved in storage efficiently helps to reduce the overall solution cost. Data efficiency not only reduces the amount of data that is stored, but it also reduces the physical capacity that is required to store the data. Reducing the footprint of the system can also lead to power and cooling savings. PowerStore includes several data efficiency methods to help reduce the total space that is consumed by storage resources created within the system. Utilizing a combination of thin provisioning, deduplication, and compression techniques, PowerStore reduces the amount of data in the system and stores it as efficiently as possible. These features are controlled by the system and are always enabled.

In PowerStore, all incoming I/O follows the same data path. The PowerStore data path and hardware components involved varies by system model and includes deduplication and compression techniques to help reduce the amount of data that is stored within the system. This document describes the PowerStore data efficiency mechanisms in more detail and outlines how space saving is achieved.

PowerStore achieves new levels of operational simplicity and agility. It uses a container-based microservices architecture, advanced storage technologies, and integrated machine-learning to unlock the power of your data. PowerStore is a versatile platform with a performance-centric design that delivers multidimensional scale and always-on data reduction.

PowerStore brings the simplicity of public cloud to on-premises infrastructure, streamlining operations with an integrated machine-learning engine and seamless automation. It offers predictive analytics to easily monitor, analyze, and troubleshoot the environment. PowerStore is highly adaptable, providing the flexibility to host specialized workloads directly on the appliance and modernize infrastructure without disruption.

The following table provides definitions for some of the terms that are used in this document.

The following section describes PowerStore data efficiency techniques and how the data path handles read and write operations. The PowerStore 1000 – 9200 models, PowerStore 500, and PowerStore Gen 3 models have different I/O handling. The data path operations are described in detail in the sections below.

The following sections describe how PowerStore system models 1000 to 9200 provide inline data efficiencies. The data efficiency techniques used within PowerStore T and Q model appliances are the same unless otherwise noted.

Before the data efficiency techniques of PowerStore 1000, 1200, 3000, 3200, 5000, 5200, 7000, 9000, and 9200 are discussed, the system design and data path must be reviewed to explain how these PowerStore systems provide inline data efficiencies. Data path components in PowerStore 1000 to 9200 models shows a high-level diagram of the components within these systems that are part of the data path. In this figure, node A and node B are shown side-by-side for demonstration purposes. Within the system, DRAM memory is used as a caching layer as data enters and exits the system. All data passes through and interacts with DRAM memory. How the data interacts with DRAM memory depends on the I/O type, primarily whether it is a read or a write. How a read and write I/O passes through DRAM memory is explained later in this section.

Another major component of the data path are the data drives. The drives, which are outlined in orange in the figure below, provide the physical capacity to the system to store data. If additional enclosures are attached to the system, they similarly add to the usable capacity of the system. Within PowerStore, any data drives in a PowerStore 1000 to 9200 system contribute to a single, large, usable capacity. Drives used for the NVMe NVRAM cache, or SCM drives dedicated for metadata tiering, are excluded from the usable capacity of the system. This space is shared for all resources within the system.

PowerStore 1000 to 9200 model systems include NVRAM drives, which are part of the PowerStore system's data path. These drives are circled in green in the lower right of the figure above. The NVRAM drives are hard drives located in the far-right slots of the base enclosure of the appliance. Depending on the system model, these two or four NVRAM drives are used as an additional write I/O cache location for protection purposes within the system. As writes enter the system, a copy of each write is placed in DRAM memory and the NVRAM drives before the host is acknowledged that the write has been saved within the system. After the host is acknowledged, data is passed through the deduplication and compression logic before being saved to the data drives of the system.

NVRAM drives in PowerStore shows more information about the NVRAM drives. As mentioned previously, each appliance contains either two or four NVRAM drives. If the system only contains two NVRAM drives, they are populated in drive slots 23 and 24. If four NVRAM drives are present, slots 21 through 24 are occupied. NVRAM drives are deployed in mirrored pairs to guard against a single drive failure. Drives 21 and 22 are configured in a mirrored pair, as well as drives 23 and 24. During an NVRAM drive failure, write caching continues to be enabled, and the mirrored pair is reestablished when the faulted drive is replaced.

The appliance also provides battery backup to the NVRAM drives using internal batteries within each node. Slots 21 and 23 are connected to Node A's internal battery device, while Node B provides power to slots 22 and 24. Battery backup is required as NVRAM drives contain both volatile and nonvolatile media. The volatile media provides fast access speeds and is used as a backup location for write caching within the system while the appliance is under normal operation. If power to the appliance is interrupted or the system is powered off, the volatile write cache is destaged to the nonvolatile media within the NVRAM drive. When the write cache information is safely stored, power is removed from the drives and the system completes its shutdown operation. The NVRAM design and operations replace the need to protect the contents of DRAM write cache.

All writes entering a PowerStore 1000, 1200, 3000, 3200, 5000, 5200, 7000, 9000, and 9200 appliance follow the same data path, regardless of the resource type. A write I/O entering PowerStore 1000 to 9200 model system displays a high-level diagram of a write I/O entering the system. The following figure also shows the steps that are taken to store the information within the system before acknowledging the host.

Step 1 depicts an I/O entering Node B. The I/O is determined to be a write, and the information is stored within write cache in the node's DRAM memory (Step 2). A copy of the write is also written to the NVRAM drives, shown as Step 3 in the figure below. As mentioned previously, the appliance NVRAM drives are used as an additional write cache location within the system for redundancy purposes. Each NVRAM drive is also dual ported, meaning that each node has access to the drives through physical internal connections and the information that is contained in them. If required, the peer node can access the data as needed.

Step 4 occurs for each write I/O that enters the system. Information is passed between the nodes to update the peer that a new write has been received and that it has the newest copy of the data for the resource. This token includes information about the I/O such as what resource was updated, the address within the resource that was updated, and the location where the I/O was saved to within the write cache. If the information is requested through the peer node, the node can access the data within the NVRAM drives by its own internal channels.

The last step (Step 5) in completing the I/O is to acknowledge the host. After the information is safely stored within the appliance and all previous actions are complete, the host is acknowledged. The data is passed through the deduplication and compression algorithms at a later time.

Unlike the PowerStore 1000, 1200, 3000, 3200, 5000, 5200, 7000, 9000, 9200 appliances, the PowerStore 500 does not support NVMe NVRAM drives. Instead, the system uses node DRAM memory as a write caching layer to temporarily store data before it is passed through the deduplication and compression process. A write I/O entering PowerStore 500 model system is a high-level diagram of a write I/O entering a PowerStore 500 system and the steps that are taken before acknowledging the host.

Step 1 depicts a write request entering Node B. The information is stored within the node's DRAM memory, as shown in Step 2. The information is then mirrored to the peer node's DRAM memory, as depicted by Step 3. A full copy of the data and information about the request is passed through the internal connections between the nodes in the system. After the data is mirrored to the peer node, the host is acknowledged that the write has been persisted on the system, as shown in Step 4. The data is passed through the deduplication and compression algorithms at a later time.

Within a PowerStore 500 appliance, Battery Backup Units (BBUs) are used to protect the contents of the appliance's volatile DRAM memory. If power is interrupted to the system, the BBUs ensure the cached contents within DRAM memory are safely persisted to a nonvolatile M.2 SATA device contained within the node. When all data is protected, the system finishes its graceful power down procedure. When power is resumed, cached data is restored to DRAM memory during the boot process, thus preserving the previous cache contents. Normal operations continue and data is passed through deduplication and compression algorithms at a later time.

The following sections describe how PowerStore Gen 3 system models (1500, 5500, and 9500) provide inline data efficiencies. The data efficiency techniques used within PowerStore Gen 3 appliances are the same across all three models.

Before the data efficiency techniques of PowerStore Gen 3 are discussed, the system design and data path must be reviewed to explain how these PowerStore systems provide inline data efficiencies. The PowerStore Gen 3 data path includes DRAM memory and Remote Direct Memory Access (RDMA) cards as major components.

Within PowerStore Gen 3 systems, DRAM memory is used as a caching layer as data enters and exits the system. All data passes through and interacts with DRAM memory. How the data interacts with DRAM memory depends on the I/O type, primarily whether it is a read or a write. The DRAM memory is protected by one or more Battery Backup Units (BBUs) and Software Defined Persistent Memory (SDPM) technology. If power is interrupted to the system, the BBUs and SDPM ensure the cached contents within DRAM memory are safely persisted to the internal M.2 non-volatile drives found within each node. When power is restored, cached data is restored to DRAM memory during the boot process.

Another major component of the PowerStore Gen 3 data path are the Remote Direct Memory Access (RDMA) cards. RDMA is a networking technology that allows direct memory-to-memory data transfers between systems without involving the CPU or operating system. In PowerStore Gen 3, physical RDMA cards on each node are used to mirror data between memory locations across nodes. This enables high-speed, low-latency cache mirroring between nodes, ensuring data availability if one node fails.

All writes entering a PowerStore Gen 3 appliance follow the same data path, regardless of the resource type or which model the write is entering. The following paragraphs describe the steps that are taken to store the information within the system before acknowledging the host/client.

Step 1 in the following figure depicts a write request entering Node 2. The information is stored within write cache in the node's DRAM memory (Step 2). A full copy of the write is also mirrored to the peer node's DRAM memory (Step 3). This mirroring is accomplished using the RDMA cards, which provide high-speed, low-latency data transfer between nodes without CPU involvement. The peer node is also notified about the information, including what resource it is intended for. After the data is mirrored to the peer node, the peer node acknowledges the write (Step 4). The last step (Step 5) in completing the I/O is to acknowledge the host that the write has been persisted on the system. Once the data is stored in both nodes' DRAM memory, the host is acknowledged. The data is passed through the deduplication and compression algorithms at a later time.

Deduplication and compression within a PowerStore appliance occur as data is copied from the appliance’s DRAM memory-based write cache to the data drives within the system. This is true for PowerStore Gen 1, Gen 2, and Gen 3 models. During this process, data is stored within the back-end storage in full stripe writes, which are created with the data remaining after it is passed through the deduplication and compression process.

All writes to the system pass through deduplication and compression logic to ensure the data is stored as efficiently as possible. These software features are part of PowerStore data path operations and cannot be disabled by the user. Workloads are also dynamically prioritized. Deduplication operations may be deferred to a period with a less extreme I/O demand. If this occurs, compression remains enabled and continues to achieve space savings. Once deduplication is resumed, blocks that were skipped are passed through the deduplication algorithm to determine if deduplication can occur. Used space is recovered if deduplication can be achieved.

When data is passed through the deduplication and compression logic, it first passes through common pattern detection, reviewing the data for all zeros or ones. If one of these patterns is matched, deduplication occurs and the metadata within the resource is updated to state what data pattern was found in that location. The metadata for a resource tells the system how it can re-create or locate data for an address within a storage resource. For example, if a block of zeros were deduplicated starting at a particular address within the resource, the metadata would have information to provide zeros as a response for that address if it is requested.

If a pattern was not found, the data is passed through the deduplication algorithm. This algorithm first creates a fingerprint for each block of data within the resource using a hashing algorithm. PowerStore Gen 1 and Gen 2 models use a 4 KB block size while PowerStore Gen 3 models use an 8 KB block size. Once created, the fingerprint is compared to other fingerprints which represent data within the PowerStore single deduplication domain within the appliance. The deduplication domain allows data for any resource within the appliance to deduplicate. If a match is found, the physical storage is single-instanced. If a match is not found, the data is sent for compression.

PowerStore Gen 3 models include unaligned deduplication, which extends the deduplication capability beyond standard page boundaries. Unaligned deduplication scans sequential data blocks to find duplicate data that may span across two consecutive blocks of data. This approach helps improve data reduction savings for datasets with unaligned sequential data patterns that would not be detected by standard page-aligned deduplication.

The data reduction process in PowerStore Gen 3 models operates in a specific order. Page-aligned deduplication is always attempted first and is preferred over unaligned deduplication. If page-aligned deduplication is not achieved and multiple consecutive blocks have been written, unaligned deduplication is then run. If neither deduplication method is successful, compression is applied. Unaligned deduplication cannot be disabled by the user, and like all deduplication and compression algorithms in PowerStore, it is always enabled.

If deduplication is not possible, compression is attempted. PowerStore will only compress a block if space savings are achieved. The resulting block, either in compressed or full form, is placed into a full stripe write to be written to the drives. In PowerStoreOS 4.0 and later, the compression algorithm allows multiple blocks of data to be compressed together. This change potentially allows for higher savings in blocks that otherwise would not see it. This feature is available for all PowerStore models.

In PowerStore, the fingerprint cache consumes a fixed region of space within system memory but can expand into the data drives to store fingerprints as needed. This allows the system to retain all fingerprints created on the appliance. The entire fingerprint cache is also mirrored within the data drives, protecting it against unforeseen issues.

Deduplication example in PowerStore shows an example of the result of the deduplication and compression logic within the PowerStore system. In this example, Resource 1 and Resource 2 have written four blocks of data each, and they are stored within the system. Deduplication within the PowerStore system allows blocks from different resources to deduplicate, reducing the amount of space consumed by duplicate data within the system. In this example, Block A is unique to Resource 1, and it is only referenced by Resource 1. Block E is present in both resources, and has deduplicated within the system, thus saving space. Further savings can be achieved if the data is compressible, as the resources can both reference a compressed block.

PowerStore is included in the Dell Technologies Data Reduction Guarantee. More information about the program can be found in the Dell Technologies Data Reduction Guarantee Terms And Conditions. In PowerStoreOS 4.2 and later, a link to the Data Reduction Guarantee can be found in Settings > Support > General Support under Support Contracts.

When a read operation is sent to a resource within the PowerStore system, the system must first determine where the latest copy of the data being requested is located. If the latest copy is in the node DRAM memory, the information is returned to the host or application requesting the data, and the read operation completes. If not, the resource metadata is reviewed to determine where the current data is located. If the data was a simple pattern, such as all zeros or ones, the block is re-created in DRAM memory and sent to the host. If the data being requested resides in the data drives within the system, the data is read and re-created in DRAM memory and sent to the host. Read operations to data residing within the drives do not cause the data to be changed in any way, thus preserving any deduplication and compression savings.

When an update to a previously written location is received, the write follows the same data path as a new write into the system. The data is saved in write cache and the host is acknowledged to complete the I/O operation. The new data that is received eventually passes through the deduplication and compression algorithms. If needed, the data is written to the back-end drives as part of the normal cache-cleaning operation to a new location within the storage, and the location previously used is invalidated.

By using multiple space-savings techniques, PowerStore attempts to use the least amount of disk space to store user data. For reporting purposes, space-savings information is available at the cluster, appliance, and resource level, and includes multiple space-saving metrics. Depending on the object being viewed, these metrics may include an overall efficiency value, and thin, snapshot, and data reduction savings values. At the appliance level, savings information is aggregated for any resources on the appliance. Likewise, at the cluster level, space savings information is aggregated from each of the appliances within the cluster.

The overall efficiency value is a computed ratio of the total provisioned space to the physical used space. The total provisioned space not only includes the provisioned size for each of the resources within the system or cluster, such as the size of the volumes, file systems, and all thin clones, but also any snapshots. The physical used space is the total used space on the drives to store the data after thin, deduplication, and compression savings are achieved. The used space is not the total amount of space that is consumed from a host or client perspective before space savings is applied. The overall efficiency value is available for the entire cluster and each appliance within the cluster.

All resources within PowerStore are thinly provisioned, which means the system does not reserve the total provisioned size of a resource from the storage, but rather allocates space as needed. The thin savings achieved for a cluster, appliance, or resource is the ratio of the total provisioned size to the logical space used. The logical space that is used in this instance is the total space that is consumed from a host or client perspective without the effects of deduplication or compression. As shown in Thin savings example, if a 100 GB volume has 25 GB of written data, the savings due to the resource being thin is 4:1.

Snapshot savings, which are displayed as snap savings within PowerStore Manager, is the ratio of logical used space for the snapshots to data that is uniquely owned by the snapshots. The logical used space is the amount of space consumed by data which is in common between the snapshot and the parent resource when the snapshot was created. For example, a volume has 5 GB of data consumed and a snapshot is then taken. The logical used space by the snapshot in this instance is 5 GB, which is shared with the parent resource. If 1 GB of data is overwritten in the volume, the amount of data that is uniquely owned by the snapshot is now 1 GB. In this example, also shown in Snapshot savings example, the snap savings is 5 GB/1 GB or 5:1.

Another space efficiency metric is the Data Reduction Ratio (DRR). This ratio is a comparison between the amount of space a dataset would have consumed if no space savings were achieved (logical used), compared to the amount of physical space occupied after passing the data through deduplication and compression (physical used). If 125 GB of data only needs 25 GB of physical storage after deduplication and compression, then the data reduction savings is 5:1. This example is shown in Data reduction savings example. In PowerStoreOS releases prior to version 4.0, the data reduction ratio is only displayed for the cluster and appliance levels.

Additional space savings metrics are also available starting in the PowerStoreOS 4.0 release. Users can view the Overall Data Reduction Ratio (Overall DRR), Reducible Data Reduction Ratio (Reducible DRR), and the amount of Unreducible Data. These metrics are available for the cluster and each individual appliance within a cluster, along with volumes, volume groups, storage containers, virtual machines, file systems, and NAS servers. For volumes, volume groups, file systems, and NAS servers, these metrics are reported for the family, such as Family Overall DRR or Family Reducible DRR.

Following are the definitions for the metrics:

Overall DRR

Ratio of space which would have been taken if deduplication and compression are not applied to the physical space occupied after deduplication and compression. The Overall DRR value is the same value as the Data Reduction Ratio from releases prior to version 4.0.

Reducible DRR

The Data Reduction Ratio for the reducible data within the cluster, appliance, or resource. This metric ignores any Unreducible Data in the calculation.

Unreducible Data

The amount of data written within the cluster, appliance, or resource that could not be reduced by PowerStore’s deduplication and compression logic.

In PowerStore, the Physical Capacity, Historical Usage, and Data Savings metrics can be viewed for the cluster, and the cluster top resource consumers from the Dashboard > Capacity tab. The cluster metrics are aggregated from the metrics of its appliances and resources, such as volume groups, volumes, file systems, and virtual machines. An example of the Capacity tab for the cluster is shown in PowerStore Manager: Dashboard > Capacity tab. Under the Data Savings title, the Overall Efficiency, Snap Savings, and Thin Savings for the cluster can be viewed. A graphic exists showing the relationship between the Logical Used and Physical Used space, along with the Data Reduction savings that were achieved. In PowerStoreOS 4.0, hovering over the Data Reduction graphic allows the user to see the Overall DRR, Reducible DRR, and Unreducible Data values for the cluster.

In PowerStoreOS version 4.2 and later, the layout of the information provided within the Data Savings box on the Capacity tab for the cluster has been updated. All information previously found within the previous codes is available in a table format, as shown in Data Savings Chart (PowerStoreOS 4.2 and later) below. Users no longer need to hover over the chart to view savings information. Data Savings also includes a ? icon, which links to PowerStore Manager online help.

When hovering over the Physical Capacity chart, the user is provided with more information about the Used and Free capacity within the system. For each metric, User and System space is also reported. For Used, the User and System space metrics state how much capacity is currently in use for User data and by the System. The System value includes the amount of space used by system metadata. For Free, the User and System values report how much free space exists.

When a resource is deleted, capacity that was used by user data is freed back to user space and any metadata space is freed to system space. When additional capacity is required for new user data or metadata, the system uses the free capacity within the system. If there is no free system space to create additional metadata, user space will be reduced to satisfy the need. This information is available within the Physical Capacity chart for the cluster and each individual appliance. An example is shown in PowerStore Manager: Dashboard > Capacity tab > Physical Capacity.

The capacity metrics for an appliance can be viewed from the Hardware > Appliances > [appliance] > Capacity tab. On this screen, as shown in PowerStore Manager: Hardware > Appliances > [appliance] > Capacity tab, the Physical Capacity, Historical Usage, and Data Savings information for the appliance can be viewed. Within the Data Savings block, the Overall Efficiency, Snap Savings, and Thin Savings for the appliance can be viewed. As mentioned previously, this information is aggregated, along with information for other appliances within the same cluster, to create the cluster metrics. Lastly, the Data Reduction savings ratio is presented within the graphic displaying the comparison between Logical Used and Physical Used space for the appliance. In PowerStoreOS 4.0, hovering over the Data Reduction graphic allows the user to see the Overall DRR, Reducible DRR, and Unreducible Data values for the appliance. If the cluster contains a single appliance, the capacity information on this page matches what is displayed on the Dashboard > Capacity tab.

In PowerStoreOS version 4.2 and later, the layout of the information provided within the Data Savings box on the Capacity tab for the appliance has been updated. All information previously found within the previous codes is available in a table format, as shown in Data Savings Chart (PowerStoreOS 4.2 and later) below. Users no longer need to hover over the chart to view savings information. Data Savings also includes a ? icon, which links to PowerStore Manager online help.

Thin Savings and Snap Savings can be viewed for volumes and file systems within a PowerStore system. This information can be found on the Capacity tab within the properties of the resource. Within the Usage box, the Used and Free space are displayed, along with the Thin Savings and Snap Savings. The amount of Snapshot/Thin Clone Space being consumed, and the amount of Family Unique Data can be viewed on this page. The Historical Usage of the resource is also displayed on the Capacity tab, which shows the Provisioned size and Used space for the resource. An example of this page for a volume is found in PowerStore Manager: Volume Properties > Capacity tab. Usage information for file systems is also available on the Capacity tab. The information includes Thin Savings, Snap Savings, Snapshot/Thin Clone space, and File System Family Unique Data. The Historical Usage chart is also available for file systems.

The Overall DRR, Reducible DRR, and Unreducible Data metrics can be viewed for volumes, volume groups, storage containers, virtual machines, file systems, and NAS servers. PowerStoreOS version 4.0 and later also allows a user to add columns for these metrics to the table view. This allows a user to easily compare the data across the resources within the cluster. PowerStore Manager: Storage > Storage Containers page shows the Family Overall DRR, Family Reducible DRR, and Family Unreducible Data columns on the Volumes page.

Capacity and data efficiency information can also be viewed within the topology views for volumes, volume groups, and file systems. To view this page, select the checkbox for a resource then More Actions > View Topology. PowerStore Manager: Volume > More Actions > View Topology > Capacity tab is an example of this page for a volume with the Capacity tab selected.

Data efficiency features within the PowerStore system are directly part of the read and write code path. All blocks written to the drives within the system can be shared among all resources within the appliance. Therefore, all resources including their snapshots and thin clones support the thin, deduplication, and compression efficiency features of PowerStore. Other features that transfer data into the system, such as import, replication, and migration, support the data efficiency features of the system. These features are passed through the same code path as a host write into the system.

When data is transferred out of the PowerStore system, it must first be returned to its full size. For instance, when a host or client requests data, the data must first be re-created within DRAM memory before it can be passed to this host. This rule is also true for data replicating to a remote cluster, data moving from one appliance to another within the same cluster due to an internal migration, and for Storage Direct backups. For replication, migration, and Storage Direct backups, when the data reaches its destination, it is subject to the data efficiency algorithms of the destination appliance.

In PowerStoreOS 4.0 and later, synchronous replication can be configured on block and file resources. When configured, each write to the local resource must be written to the destination system before the host is acknowledged. Each write is transferred to the remote system in its original size, and deduplication and compression savings are handled by the remote system, independently of the source. For more information about synchronous replication, see the Dell PowerStore: Replication Technologies white paper.

PowerStore offers powerful data efficiency techniques which help reduce the amount of storage that is required to store datasets for the various resources within the system. Through the always-on features of thin resources, deduplication, compression, and the intuitive management of data that is written within the storage of the system, PowerStore uses its space as efficiently as possible. By reducing the amount of storage needed to store a dataset, data efficiency techniques help to further reduce the total cost of ownership of a solution.

The Dell Technologies Info Hub > Storage site provides expertise that helps to ensure customer success with Dell storage platforms.

Dell.com/powerstoredocs provides detailed documentation about how to install, configure, and manage Dell PowerStore systems.