Home > Storage > Data Storage Essentials > Storage Admin > The Next Generation PowerMax - Balancing Performance with Efficiency and Sustainability > Industry Leading Storage Density
Storage density plays a crucial role in power-efficient storage systems due to its direct impact on energy consumption, physical space utilization, and overall system efficiency. Storage density refers to the amount of data that can be stored within a given physical space or storage medium. By leveraging higher storage densities, organizations can optimize energy usage, reduce operational costs, and ensure sustainable growth for their storage infrastructure. The following section discusses how the next generation PowerMax delivers the remarkable improvements in storage density and storage power efficiency over the previous generation PowerMax systems.
The next generation PowerMax uses a newly designed disk storage enclosure called the Dynamic Media Enclosure (DME). The next generation PowerMax DME is designed to maximize storage density, supporting up to 48, side loading, high capacity 2.5” NVMe SSDs allowing for up to over 4 PBe in a 2U form factor. This new side loading design provides customers with dramatically improved storage density, far more efficient cooling, temperature control, as well as improved flexibility and scalability. This is particularly beneficial for data centers and enterprise storage environments where maximizing storage capacity is crucial.
Side loading drive carriers, also known as horizontal drive carriers, offer advantages in terms of storage capacity per rack unit compared to other drive carrier orientations. Side loading drive carriers allow for a higher density of drives within a given rack unit. By positioning the drives horizontally, side by side, the rack unit is fully utilized in terms of width, height, and depth. Drives can be layered across the entire depth and width of the enclosure rather than being only limited to its width. By layering drives in this manner, more capacity per rack unit can be achieved as more drives can be added without having to increase the height of the enclosure.
The next generation PowerMax supports a variety of NVMe SSD capacity sizes (3.84 TB, 7.68 TB, 15.36 TB, and 30.72 TB). Drive sizes can be intermixed within a DME with the only requirement that the drives be one size apart (i.e., 3.84 TB drives can be intermixed with 7.68 TB drives).
PowerMax DME Enhanced Adaptive Cooling
The side loading nature of the DME allows for an innovative thermal design which effectively manages the increased heat generated for its high-density NVMe SSD configuration. The new DME uses an adaptive cooling design featuring six variable speed fan packs, each with dual 60 mm fans, which push a high velocity airflow from the front of the unit, around the sides of the NVMe SSD drives, and out the back of the enclosure. The fan packs variable speeds are controlled by an onboard BMC and sensors which are constantly measuring air pressure, air speed, and temperatures within the unit and of each individual NVMe SSD in the enclosure.
The next generation PowerMax DME has its fan packs mounted in the front of the enclosure creating a high velocity airflow which traverses the enclosure from the front to the back. With high density storage systems such as the next generation PowerMax, airflow is of paramount importance in the drive enclosures as proper airflow plays a crucial role in maintaining optimal operating temperatures, ensuring the reliability and longevity of the drives, and maximizing overall system performance. This airflow along with the side loading alignment of the drives greatly improves the heat dissipation of the unit as the drive carriers are aligned with the airflow and not perpendicular to it. This provides several additional cooling efficiency advantages over traditional enclosures with the drives in the front and fans in the rear. Some of these advantages are:
Air Intake: Front-facing fans pull in cool air from the front of the enclosure, where ambient air temperatures are typically lower. This allows the fans to draw in fresh, cooler air directly onto the drives and other components, facilitating more efficient cooling. Rear fans, on the other hand, often face warmer air due to the heat generated by the components within the enclosure.
Airflow Direction: Front-facing fans create a direct path of airflow that passes over the drives and other heat-generating components before exiting through the rear of the enclosure. This airflow pattern helps to remove heat from the drives more effectively, as the air is forced to flow across the entire surface area of the drives. Rear fans, by comparison, often have a less direct airflow path. This creates a less efficient and inconsistent cooling coverage for all drives.
Heat Dissipation: Drives in an enclosure tend to generate heat predominantly on their front-facing side where the circuitry and components are located. Front-facing fans are positioned to target this area directly, allowing for efficient heat dissipation. Rear fans, while still contributing to overall airflow, may not provide as focused cooling to the front side of the drives, potentially resulting in less effective heat dissipation.
Hot Spot Prevention: Front-facing fans help prevent the formation of hot spots within the enclosure. By drawing in cooler air from the front, they can evenly distribute airflow across all drives, minimizing temperature variations and preventing localized heat concentrations. This is particularly important in high-density storage environments such as the next generation PowerMax where the drives are closely packed together.
Maintenance Accessibility: Front-facing fans are more accessible for maintenance and cleaning. Since they are at the front of the enclosure, they can be easily accessed and cleaned without removing drives or other components. This accessibility ensures that the fans can be kept free from dust and debris, optimizing their cooling performance.
In summary, airflow management is critical for high-capacity drive enclosures such as the PowerMax DME to maintain optimal operating temperatures, extend drive lifespan, and ensure reliable performance. The DME’s advanced airflow and innovative active cooling design contributes to the overall system stability, reduces the risk of failures, and helps customers maximize the value and longevity of their storage infrastructure.
The next generation PowerMax introduced a new distributed active-active RAID (redundant array of independent disks) protection architecture called Flexible RAID. Flexible RAID is one of the key components that drive the next generation PowerMax efficiency improvements regarding storage capacity per watt and storage density.
The foundational element of Flexible RAID is shared physical resource pools called disk groups. Disk groups are comprised of physical drives of the same technology, size, and performance characteristics. In Flexible RAID, RAID protection is assigned at the disk group level. Flexible RAID supports RAID1 1+1, RAID5 3+1, RAID5 8+1, RAID 5 12+1, and RAID6 12+2 protection schemes. A disk group can have only one protection scheme assigned to it. Disk groups are broken down into smaller groupings of drives called RAID clusters. A disk group can contain multiple RAID clusters which inherit the parent disk group's attributes, such as its RAID protection scheme. The number of drives in each RAID cluster (RAID width) is determined by its RAID protection scheme. The host presentable namespaces are carved from the capacity contained within a RAID cluster.
The physical location within the DMEs of the RAID cluster drives is determined algorithmically by PowerMaxOS to ensure the highest fault tolerance, accessibility to the nodes, and available capacity utilization. This automatic configuration by PowerMaxOS means that a customer does not need to predetermine the physical layout of the drives in the system to achieve the highest levels of fault tolerance and performance. As drives are added into the system, PowerMaxOS automatically selects which disk group and RAID cluster the additional capacity will be allocated to.
In the next generation PowerMax, disk groups can span across multiple DMEs in the system. A key enabler for Flexible RAID disk groups to span multiple DMEs is the Dynamic Fabric (discussed in detail in the next section). The dynamic fabric allows for any physical drive in the disk group to be accessed by any node in the system in a distributed, balanced manner regardless of which DME the drive is located in. The combination of Flexible RAID with the dynamic fabric allows for the following benefits over previous generation PowerMax systems:
More capacity using fewer drives: Flexible RAID uses a universal sparing algorithm in which spare space is distributed across all the drives in a RAID Cluster, eliminating the need for a dedicated spare drive
Flexible upgrades: Flexible RAID allows customers to add capacity to the system in single drive increments as the individual drives are treated as unique endpoints on the dynamic fabric.
Higher resilience with faster rebuilds: With Flexible RAID, drive rebuilds occur 2x faster than the previous PowerMax generation. This is because of the distributed nature of the dynamic fabric allows for all drives to be accessed by all the nodes in the system. This allows for the rebuild process to be spread across all the compute resources in the system, increasing the computing power, which allows for increased parallel processing and concurrent execution of the rebuild tasks.
Data reduction plays a crucial role in achieving energy efficiency in storage arrays. It involves techniques and algorithms that aim to reduce the amount of data stored, transmitted, or processed. Storage arrays with data reduction provide the following efficiency benefits to customers:
Reduced Storage Requirements: Data reduction techniques such as compression, deduplication, and thin provisioning help minimize the amount of physical storage space required. This allows organizations to reduce their storage infrastructure leading to lower power consumption for cooling, reduced hardware costs, and a smaller physical data center footprint.
Lower Data Transfer Overhead: Data reduction techniques can significantly reduce the amount of data that needs to be transmitted over networks or across storage systems. This reduction in data transfer results in lower network traffic and decreased energy consumption in data transmission.
Extended Hardware Lifespan: By reducing the amount of data stored and processed, data reduction techniques can help extend the lifespan of storage hardware. This translates to energy savings by minimizing the manufacturing, deployment, and disposal of storage devices.
Environmental Impact: Energy-efficient storage arrays with data reduction capabilities contribute to overall sustainability efforts. By reducing energy consumption, organizations can lower their carbon footprint and minimize their environmental impact, aligning with green IT initiatives and promoting a more sustainable approach to data management.
Data reduction plays a significant role on how the next generation PowerMax delivers its efficiency achievements. Both the PowerMax 2500 and 8500 systems delivers a powerful 5:1 data reduction guarantee for Open Systems block storage and are the only storage platforms in the industry that deliver a 3:1 data reduction guarantee for Mainframe block storage.
Both the PowerMax 2500 and 8500 use inline hardware compression using the Adaptive Compression Engine (ACE). The ACE engine was originally developed for first generation PowerMax and VMAX all flash models but has been significantly enhanced for the next generation PowerMax systems. ACE provides PowerMax with a powerful data reduction method that provides negligible performance impact while delivering the highest space saving capability. The following design factors make the Adaptive Compression Engine unique:
Intelligent compression algorithms – Intelligent compression algorithms determine the best compression ratios to be used and provide the ability to dynamically modify storage backend layout for the highest data compression efficiencies.
Inline hardware data compression – Inline hardware data compression greatly inhibits the compression function from consuming critical PowerMax system core resources.
Activity Based Compression – Activity Based Compression (ABC) focuses the compression function on the least busy data in the system, while allowing the most active data in the system to bypass the compression workflow. This ensures that all data in the system receives the appropriate compression focus while maintaining optimal response time.
Fine Grain Data Packing – Fine Grain Data Packing which includes a zero reclaim function that prevents the allocation of buffers with all zeros or no actual data.
Enhanced Compression – There is an additional compression algorithm found in PowerMaxOS called Enhanced Compression (EC). The EC algorithm scans already compressed data which the system determines to have been not accessed for a long period of time. The EC algorithm then tries to further reduce this data, to a larger compression ratio, to realize additional capacity savings.
As said earlier, the Adaptive Compression Engine has been available for open systems block environments since original PowerMax and VMAX All Flash, however beginning with the next generation PowerMax systems ACE also became available for mainframe environments. The next generation PowerMax is the only storage platform in the industry which provides data compression for both mainframe and open systems environments.
The next generation PowerMax employs inline hardware deduplication to identify repeated data patterns on the array and store those repeated patterns using a single instance in the array’s usable capacity. Depending on customer workloads, inline deduplication along with inline compression give PowerMax the ability to achieve an industry leading data reduction ratio with negligible performance impact. The following are the important design factors for deduplication on PowerMax:
Inline hardware data deduplication – Inline hardware-based data deduplication prevents the consumption of critical PowerMax system core resources, limiting performance impact. The deduplication and compression functions are performed on the same hardware module in the PowerMax system.
Deduplication Algorithm – PowerMax uses an advanced algorithm to perform the deduplication function. This algorithm produces a unique data identifier for each item of data which is processed through the compression engine. These unique identifiers are stored in a Hash ID table on the PowerMax system.
Hash ID Table – The Hash ID table stores all the Hash IDs for the data processed through the compression engine. When a new write enters the compression engine, the Hash ID created for the write is compared to the Hash IDs already in the table. If the Hash ID is found to already exist in the Hash ID table, the write is not written to the storage.
Dedupe Management Object (DMO) – The DMO is a 64-byte object which serves as the connection (pointer) between the devices and the single instance of actual data. DMOs are stored in the PowerMax global cache.
Inline deduplication is available to all open systems next generation PowerMax customers, but it is not currently supported for mainframe environments.
In summary, storage density is a critical factor in power-efficient storage systems as it directly influences power consumption, cooling efficiency, space utilization, and scalability. By leveraging the next generation PowerMax higher storage densities, organizations can optimize energy usage, reduce operational costs, and ensure sustainable growth for their storage infrastructure.