Home > Storage > PowerMax and VMAX > Storage Admin > Next-Generation PowerMax Family Overview > Key design features of second-generation PowerMax systems
At a high level, there are seven main components that comprise the bedrock upon which the new PowerMax systems are built. These are:
The PowerMax 2500 and PowerMax 8500 are built on modular building blocks called “nodes,” similar to the PowerBrick from the previous-generation PowerMax. Nodes contain the primary compute elements (CPU and memory) of the second-generation systems. Each second-generation PowerMax system has at least two nodes or a “node pair.” Each node has dual Intel Xeon Scalable processors, 24 DDR4 DIMM slots, and two 64 lane PCIe switches for front-end connections, among other advanced features.
The storage component for the PowerMax 2500 and 8500 is called the Dynamic Media Enclosure (DME). Each DME has 48 top loading slots for 2.5” U.2 based NVMe flash drives. Drives are side loaded into the DME slots. The side loading architecture provides better air flow for much improved cooling characteristics. This results in a considerably lower watts/GB of capacity ratio, greatly enhancing the overall efficiency of the system, especially at scale.
The second-generation PowerMax platform architecture has also been designed with sustainability and efficiency as a core tenent. As a result, both the PowerMax 2500 and 8500 are the most energy-efficient enterprise-level storage platforms Dell has ever produced. The second-generation platform delivers over 5x the effective capacity per watt consumed (PBe / watt) over the previous PowerMax generation.
The PowerMax 2500 can scale up to two node pairs and two DMEs, while the PowerMax 8500 can scale up to eight-node pairs and up to eight DMEs.
Both the PowerMax 2500 and 8500 are architected using new internal NVMe-oF topologies, featuring NVMe/RDMA using 100 GbE InfiniBand (NVMe/IB) and NVMe/PCIe based fabrics. These new NVMe-oF topologies are referred to as the PowerMax “Dynamic Fabric.” The dynamic fabric turns the compute and backend storage elements into individual independent endpoints on a large internal storage area network (SAN). These individual compute and storage endpoints can be placed into shared logical and physical resource pools, to disaggregate the storage and compute resources in the system. In this architecture, all node endpoints can access all storage endpoints in the DMEs using the system’s high speed NVMe-oF topologies, creating a true “active/active and share everything” system architecture. This system disaggregation decouples the compute and storage so that they be scaled and provisioned independently of each other to meet application requirements, rather than adhering to strict system architecture requirements. This disaggregated architecture makes the new PowerMax systems:
The second-generation PowerMax system is the first Dell Technologies storage platform to use Persistent Memory (PMEM) DIMMs (Dual In-Line Memory Module). The second-generation PowerMax uses PMEM for these primary purposes:
The primary value proposition of the use of PMEM in the second-generation PowerMax systems is that it enables an overall lower cost of system ownership because workload density and capacity can be increased using a smaller overall system footprint.
The disaggregation of storage and compute allows the PowerMax 2500 and 8500 to implement a new RAID distribution scheme called Flexible RAID Technology. At a high level, Flexible RAID provides customers with:
Flexible RAID provides all compute nodes in the system with active-active access to the storage resources that are distributed across all DMEs in the system. This model reduces the RAID overhead of the system, allowing for much higher system capacity while using fewer drives. This model also improves rebuild efficiency by over 2x when compared to the first-generation PowerMax, as 1 TB can be rebuilt in less than 10 minutes using the Flexible RAID model.
Flexible RAID offers flexible capacity expansion while meeting the availability and reliability numbers of traditional RAID.
Flexible RAID allows the new PowerMax systems to deliver the highest storage performance, resiliency, and efficiency. The technology provides more usable storage capacity by leveraging granular storage media, load balancing, and several RAID protection schemes built around RAID 1, 5, and 6.
The new PowerMax systems run a completely redesigned version of PowerMaxOS called PowerMaxOS 10. The latest PowerMaxOS 10 software builds on decades of software innovation to provide trusted, intelligent, secure storage for the most demanding mission-critical workloads while simplifying operations. The newly designed PowerMaxOS 10 brings forth new enhancements that make the PowerMax 2500 and PowerMax 8000 far more efficient at managing cache memory resources than the previous-generation PowerMax. This has allowed for a significant reduction in the amount of required DRAM cache for these systems, lowering overall supply chain dependencies, and reducing the customer’s total cost of ownership (TCO) long term.
Being able to intelligently automate storage administration tasks has become a paramount requirement for the modern data center, especially as storage infrastructure scales up and out. All modern storage platforms need to provide customers with the ability to use a REST API that provides access to the storage system so they can build automation scripts and playbooks using automation tools such as Ansible. The PowerMax platform is the first storage platform in the industry that can perform provisioning and other administrative tasks for open-system block, file, and mainframe workloads using a single unified comprehensive REST API toolkit.
Another core automation feature that applies to both the PowerMax 2500 and PowerMax 8500 is optimized workload placement. In a storage infrastructure with multiple PowerMax arrays, the systems will send information to Unisphere about storage usage and utilization. As a user provisions storage either manually through Unisphere or in an automated way using scripts, the user can allow the system to determine which PowerMax storage array is best suited to support the new workload.
The second-generation PowerMax also simplifies the storage provisioning for stateful applications on Kubernetes containers. Using an orchestration tool such as RedHat OpenShift, a user can select the PowerMax-SRDF storage class and embed it in a create stateful set code snippet. Upon running this code snippet, the user creates a complete storage environment on the PowerMax array, including remote replication for the stateful application.
Data is the life blood of the modern connected world. Cyberattacks and breeches in which data is stolen, altered, and rendered unusable can cost corporations many millions of dollars in lost revenue. The new PowerMax systems have been designed to provide our customers the highest levels of data protection and cybersecurity in the industry.
The Dell cybersecurity model aligns with the National Institute of Standards and Technologies (NIST) cybersecurity framework and is centered on the following principles:
The best cyber security strategies begin with prevention. The PowerMax is purpose-built to prevent unauthorized access to system resources. Each model incorporates intrinsic security features and comprehensive access controls to safeguard company data. These features include:
Dell CloudIQ is a powerful application used to track system health through pattern recognition and advanced analytics. Through CloudIQ cybersecurity, users can define legal configurations for PowerMax, monitor the system, and receive alerts if the array is out of compliance. CloudIQ can also track data patterns and detect anomalies, including changes to data reduction rates, to determine whether ransomware or malware may have infected the system. When suspicious anomalies are detected, CloudIQ alerts IT management to take corrective action.
Both first- and second-generation PowerMax systems use secure immutable snapshots to provide the industry’s most granular cyber recovery at scale, maximizing data recovery from a cyberattack. Administrators can set snapshot policies for up to 65 million secure snapshots to optimize recovery point objectives (RPO) and minimize data loss. Several options also exist for native cyber recovery from a secure vault for open systems and mainframe storage on PowerMax.
PowerMax is designed to consolidate demanding mixed workloads while delivering consistently high levels of performance. Its modern scale-up and scale-out architecture is ideal for relational databases, real-time analytics, demanding transaction processing workloads, and big data applications that require uncompromising uptime and extremely low latency.
PowerMax is the only platform in the industry in which a customer can run mainframe, open systems block, and file workloads natively on the same system without the use of gateways or third-party solutions. The ability to run mainframe workloads natively along with OS block and file is available on both the PowerMax 2500 and 8500, and on the PowerMax 8000. The PowerMax 2500 and 8500 are the only storage platforms in the industry which provide a guaranteed 3:1 data reduction ratio for mainframe workloads.
The PowerMax 2500 and 8500 also feature a completely redesigned 64-bit fully embedded NAS file platform for SMB and NFS workloads. This new file platform can provide four data movers (virtual machines acting as file servers) on the PowerMax 2500 and eight data movers on the PowerMax 8500. Each data mover can provide users with up to 512 TB of usable capacity which can be carved up into 64 TB file systems.
The following sections describe some specific PowerMax second-generation mainframe and file enhancements.