The IDC team has done an updated lab validation of the PowerMax platform. You can find the full document here. The report highlights the simplicity, efficiency, and resilience of the PowerMax family of arrays. The slide layout makes it easy to rapidly digest.
The report highlights a number of PowerMax features and functions. Some of these are things we have done for a long time (like Remote RAID), while others are fairly recent (like SRDF/Metro Smart DR). The management of PowerMax has a great GUI (Unisphere for PowerMax), as well as REST API, cloud management (with CloudIQ), and of course a CLI. The report highlights data efficiency, local replication tools, and drills into some of the great remote replication features.
As part of the discussion on management and ease of use, there are details on the REST API. The REST functions have grown over several releases, helping customers to move from CLI scripts for repeated tasks to automated workflows. The API can serve up details on configuration, performance, and system health, as well as actively managing array resources. It is easy to consume in multiple ways, including a Python library and Ansible playbooks (all written in Python and available for customers to edit as desired). Since PowerMax arrays tend to be large, the code is fully re-entrant, allowing for many parallel non-overlapping operations to be in flight at the same time.
PowerMax snapshot functionality is very advanced. Snaps have an automatic time to live, after which they will be expired (no need to chase them down later). The snaps can be secure, protecting them from accidental or malicious deletion before their planned expiration period completes. For long-term retention, snaps can be shipped to the cloud. With snapshot policies, snap protection service levels are applied to groups of devices (storage groups). Policies automate the frequency of snaps, the number of copies to keep, the type of each snap, and the location. By applying policies to the various storage groups, customers can be assured that their desired protection policies are being met automatically in a way that is tailored to the business value of each group. And given the rise in malware and other threats, automated protection provides real peace of mind.
PowerMax Remote RAID is a core function of SRDF that is sometimes overlooked. SRDF provides a remote version of a local group of devices. The remote version is not a copy - the two instances are linked together at the track level, just like a locally mirrored pair of devices would be. Both arrays know which tracks are current on each system. And in the event of a local RAID failure, SRDF will automatically read any data it does not have locally from the remote instance. This is not something that needs to be invoked – it is always present for any mode of SRDF replication. PowerMax devices with local 7+1 RAID 5 are highly available (planned for over 7x9s, and with proactive replacement the practical availability is virtually 100%). With SRDF active, the added planned availability becomes over 1000x better than having local RAID 6. And since most other arrays use remote copy technology that does not provide this same sort of remote resilience benefit, the other arrays lead off with recommending RAID 6 (or other dual parity structures) to ensure data integrity. PowerMax with RAID 5 and SRDF delivers far better availability without all the RAID 6 overhead.
SRDF/Metro provides a true active/active copy of data – customers can write to any block of any device on either array and the other will be updated with only a single round trip delay. With SRDF/Metro, customers can extend cluster-aware applications such as ESX clusters or Oracle RAC and allow the applications to function as if they are in the same building sharing access to a single array. SRDF/Metro Smart DR adds a single, simple remote copy for both sides of a Metro device pair. By adding Smart DR, devices from both sides are linked up to a single device in the remote array. Updates that happen will be sent by one array or the other, using the same bandwidth as being linked to a single source but offering the protection of connectivity to both. Should anything happen to either production site, the surviving array will keep up the remote copy without the loss of a single transaction.
The IDC team did a great job of detailing many key differentiators for PowerMax in the market. The paper delivers some great detail on various implementation details around those features. Contact your local Dell team for more, or reach out to me at [email protected]. Thanks for your interest in PowerMax.
