Dell EMC Announces VMAX Enhancements

Disclaimer: I recently attended Dell EMC World 2017.  My flights, accommodation and conference pass were paid for by Dell EMC via the Dell EMC Elect program. There is no requirement for me to blog about any of the content presented and I am not compensated in any way for my time at the event.  Some materials presented were discussed under NDA and don’t form part of my blog posts, but could influence future discussions.

Dell EMC today announced some enhancmenets to its VMAX platform, along with a new all-flash model. I thought I’d run through the highlights.



Dell EMC are saying that the 950F offers a “68% increase over the VMAX 850F”.  They’re also saying that they’re seeing “sustained low latency across most workloads”.



The primary reason behind this improvement is the V-Brick engines being refreshed with the latest Intel Broadwell 18-core CPUs.  Each V-Brick has 4-18 core CPUs, bringing the total scale of a 950F up to 576 CPU cores. This is a whole lot of cores to work with and provides some tangible benefits, including reduced response time (compared to the VMAX 850F). A side benefit of this is that you can now get the same performance and capacity as the 850F in a 25% smaller footprint.



The cache on the 950F is based on newer high-speed DDR4 memory, the same as the 250F.  As with the 250F, different V-bricks can have difference cache sizes (1TB or 2TB). You can also jam up to 4PB of effective capacity in these things by leveraging the latest high-capacity 15.36TB  SSDs.  The 950F uses 16 links of 6Gb SAS for connectivity to the drives which provides the same bandwidth per engine as all other VMAX All Flash systems.

V-Brick scaling for open systems is consistent with the 450F and 950F with V-Bricks starting with 53TB of flash capacity and scaling out storage via 13TB Flash Capacity packs.  Host connectivity ports on a fully loaded system is 192 for open and 256 for FICON if running all mainframe.



D@RE External Key Manager

If you’re into D@RE, you might just be excited to hear this news.

  • Allows customers to use a standardized key management infrastructure for D@RE across their entire environment.
  • Support for external key managers using KMIP (Key Management Interoperability Protocol).  KMIP defines message formats for the manipulation of cryptographic keys on a key management server.
  • Allows customers to integrate VMAX with existing key management infrastructure.
  • Leverage external key manager platforms such as Gemalto (SafeNet) KeySecure, IBM Secure Key Lifecycle Manager and other KMIP compatible servers (in the future).   The client runs securely on the VMAX Management Module Control Station (MMCS) inside the VMAX engine.

This simplifies key management (i.e. key generation, escrow, recovery) for VMAX and other KMIP compatible encryption solution like Unity. It also:

  • Simplifies security compliance across the DC;
  • Provides a centralised audit log and FIPS 140-2 Level 3 compliance with Hardware Security Module (HSM) integration; and
  • Allows multiple key server appliances to be clustered (even in geographically dispersed DCs) providing even higher levels of availability.

This gives the security keys to the array back to the teams who need them.

Customers who’ve been using the embedded key manager on VMAX 3 or VMAX All Flash can non-disruptively migrate those keys to the external key manager. Dell EMC tell me this is both “painless and secure”.


Secure Snaps with TimeFinder SnapVX

Secure Snaps are a cool feature that provide a secure option to protect against accidental deletion or malicious attacks on data.


ProtectPoint for VMAX

ProtectPoint support now extends to Microsoft SQL Server and Microsoft Exchange. You can now backup Microsoft SQL and Exchange directly to Data Domain.

Also new is the Instant Recovery feature for File Systems. Once a file system restore is initiated the data is immediately accessible to the host. Any application recovery workflow that uses file systems can benefit from this.



In what I think is great news, RecoverPoint is now available for VMAX3 and VMAX AF platforms. With this you get:

  • Heterogeneous replication, with support for XtremIO, UNITY, VNX, VMAX, VMAX2, and third-party arrays via VPLEX;
  • Granular recovery from particular points in time (meaning RPOs as low as a minute).

RecoverPoint uses VMAX TimeFinder SnapVX local snapshot technology. The SnapVX-based replication relies on the delta between consecutive snapshots. Note that the data differential is further compressed and deduped to minimize the WAN bandwidth consumption.


  • Is supported on both VMAX3 (100K/200K/400K) and VMAX All Flash platforms: VMAX 250F, 450F, 850F and 950F;
  • Supports flexible fan-in and fan-out configurations; and
  • Requires Gen 5 or Gen 6 hardware and RecoverPoint 5.1.



VMAX isn’t the sexiest array on the block any more. There’s nothing wrong with that though. It does what it does really well. And the people buying these arrays are interested in performance and reliability. So why would you buy VMAX now when there’s a slew of alternative solutions available? Some of it is sweat equity. Shops running VMAX invariably have a big investment in processes and tools to support the VMAX environment. They work in places that don’t necessarily move at a rapid pace, or they do things that require a reliable, repeatable result. That doesn’t mean they don’t need useful things like external key management systems though. Financial services places are rife with security folk ready to put the boot into hapless storage operators in the name of compliance. But I digress.

There aren’t a million new features in this release, but it’s nice to see some of the features finally make it to VMAX. I’m hopeful that Dell EMC continue to invest in the platform, and continue to integrate useful protection tools such as ProtectPoint and Secure Snaps across the portfolio. The VMAX 950F is orderable this month and will be generally available in June. You can read more about it in a blog post from Dell EMC here.

EMC announces new VMAX range




Powerful, trusted, agile. That’s how EMC is positioning the refreshed range of VMAX arrays. Note that they used to be powerful, trusted and smart. Agile is the new smart. Or maybe agile isn’t smart? In any case, I’m thinking of it more as bigger, better, more. But I guess we’re getting to the same point. I sat in on a pre-announcement briefing recently and, while opinionalysis isn’t my strong point, I thought I’d cover off on some speeds and feeds and general highlights, and leave the rest to those who are good at that kind of thing. As always, if you want to know further about these announcements, the best place to start would be your local EMC account team.

There are three models: the 100K, 200K and 400K. The 100K supports

  • 1 – 2 engines;
  • 1440 2.5″ drives;
  • 2.4PB of storage; and
  • 64 ports.

The 200K supports

  • 1 – 4 engines;
  • 2880 2.5″ drives;
  • 4.8PB of storage; and
  • 128 ports.

Finally, the 400K supports

  • 1 – 8 engines;
  • 5760 2.5″ drives;
  • 9.6PB of storage; and
  • 256 ports.

*Note that the capacity figures and drive counts are based on code updates that are scheduled for release in 2015.

Hypermax Operating System is a significant enhancement to Enginuity, and is built to run not just data services inside the box, but services coming in from outside the box as well. This includes an embedded data storage hypervisor allowing you to run services that were traditionally run outside the frame, such as management consoles, file gateways, cloud gateways and data mobility services.

Dynamic Virtual Matrix is being introduced to leverage the higher number of cores in the new hardware models. In the largest 400K, there’ll be 384 CPU cores available to use. These can be dynamically allocated to front-end, back-end or data services. Core / CPU isolation is also an available capability.

While they look like an ultra-dense 10K, they’re not. You can have two engines and drives in a single cabinet. All models support all-flash configurations. If money’s no object, you could scale to 4PB of flash in one frame.

Virtual Matrix is now Infiniband, while the backend is now SAS.

EMC claims base support for 6 * 9s of availability, and 7 * 9s availability with VPLEX (that’s 5 seconds per year of downtime).

Snapshotting has been refreshed, with SnapVX supporting up to 1024 copies per source. Doesn’t impact I/O, and doesn’t require target configuration.

Finally, read up on EMC ProtectPoint, it’ll be worth your time.