In the first post, we gave an high level overview of the product (https://volumes.blog/2020/05/05/what-is-powerstore-part-1-overview/) Now, lets dive a little bit deeper: There are ten different models PowerStore Models PowerStore is […]
In the first post, we gave an high level overview of the product (https://volumes.blog/2020/05/05/what-is-powerstore-part-1-overview/)
Now, lets dive a little bit deeper:
There are ten different models PowerStore Models
PowerStore is designed from the ground up to utilize the latest in storage and interface technologies in order to maximize application performance and eliminate bottlenecks. Each PowerStore appliance has two nodes and uses NVMe to take full advantage of the tremendous speed and low latency of solid-state devices, with greater device bandwidth and queue depth. PowerStore has been architected to maximize performance with NVMe flash storage and supports the even greater demands of Intel Optane Storage Class Memory (SCM) which provides performance approaching the speed of DRAM.
This performance-centric design enables PowerStore to deliver 6x more IOPs and 3x lower latency for real-world workloads compared to previous generations of Dell midrange storage.
PowerStore is a flexible design built to meet the requirements of different storage applications with support for high availability. The PowerStore platform design includes two major configurations: PowerStore T and PowerStore X. The table displays the available models and specifications for each platform.
There are ten different models within the PowerStore product line: Five PowerStore T models and five PowerStore X models. The higher the model number, the more CPU cores and memory per system. PowerStore systems consist of nodes, one or more base enclosures, one or more expansion enclosures, and appliances.
For high availability, PowerStore systems have:
- Two redundant power supplies
- Multiple redundant network ports with system bond
- Two redundant nodes
- RAID-protected disk drives
PowerStore T systems support clusters of up to four appliances for:
- Constant uptime with intracluster migrations
- Scale up
- Simplified management
- Automatic data placement
PowerStore Back Enclosure – Back View
The back view shows I/O modules and port that provide connectivity for system management, to front-end host, and back-end Expansion Enclosures (shelves)
Management port (in red) is only used only with PowerStore T appliances. Two ports on the mezz card are used for management traffic with PowerStore X appliances.
The Base Enclosure supports only NVMe devices with twenty-five (25) slots that are labeled Slots 0 to24. SAS SSDs can only be added to Expansion Enclosures. In the Base Enclosure, the first 21 slots, slots 0 through 20 can be populated with either NVMe SSD or NVMe SCM drives for data storage.
The same drive types must be populated in the 21 slots. You cannot mix NVMe SSD, and NVMe SCM drives in the same base enclosure. Minimum of 6 SSDs must be used. The last two or four slots (dependent on the model) must be populated with NVMe NVRAM devices and are used for write cache and vaulting. On the PowerStore 1000 and 3000 the last two slots (23 and 24) are reserved for two (2) NVMe NVRAM devices.
PowerStore supports four types of drives: NVMe SSD (Flash), NVMe Storage Class Memory (SCM) SSD, NVMe NVRAM, and SAS SSD (Flash). The drive types must be installed in specific locations and enclosures.
The NVMe flash and NVMe SCM drives on the left are supported in the Base Enclosure, slots 0 to 20. The NVMe NVRAM type drives are supported in the Base Enclosure, slots 21 to 24 for write cache. SAS flash drives shown on the right are only supported in the PowerStore Expansion Enclosures.
Connect PowerStore to a pair of Ethernet switches to ensure high availability, not single-switch configurations. This requirement applies to switches used for iSCSI, file, intercluster management, and intercluster data. Dell EMC does not process PowerStore orders that include only a single switch.
Each node must have at least one connection to each of the Ethernet switches. Multiple connections provide redundancy at the network adapter and switch levels.
It is recommended that you deploy the switches with Multi-Chassis Link Aggregation Group (MC-LAG). The Dell version of this is called Virtual Link Trunking interconnect (VLTi) topology. Alternative connectivity methods—including reliable L2 uplinks and dynamic LAG—should be used only a solution like VLTi is not a possibility.
The PowerStore supports Dell EMC Networking Top-of-Rack (ToR) switches running OS10 Enterprise Edition (OS10EE). Third-party switches with requisite features are supported. See the Support Matrix for a list of supported switches.
Dell EMC recommends the following supported Dell EMC PowerSwitch Ethernet switches:
(For information about OS10EE, go to Dell Support and search for OS10EE.)
PowerStore T and PowerStore X Switches
PowerStore T and PowerStore X have different switch configurations.
Considerations for OOB management configuration:
- At least one OOB management switch is recommended for PowerStore T configurations. PowerStore X does not support OOB management.
- Can be configured with or without a management VLAN.
- Switch ports must support untagged native VLAN traffic for system discovery.
you can take a virtual tour by clicking the screenshot below
you can also download the introduction to the platform white paper which contains much more in-depth information, by clicking the screenshot below
in the 3rd post, we are going to cover PowerStore with AppsON