PowerMax – A next Generation “12 Factor” All-Flash Array is born

After we discussed the details about the new PowerMax with our Modern Data Center Specialist Sanel Samardzic, I was very impressed. Even as a “Swiss Systems Engineer” – we tend to find always something to be sceptic about – I have to say that the PowerMax storage array is close to perfect.

I call it a next generation All-Flash array and will describe which factors you should consider when you are in process to decide on your future enterprise storage technology.

PowerMax is architected with a no-compromise multi-controller architecture leveraging industry-standard NVMe technology that will ensure customers are ready to support NVMe-over-Fabrics and high speed, low-latency Storage Class Memory (SCM) drives when they become available in early 2019.

For sure most of you are aware about the 12 factor App design principle to build a modern Cloud Native Application. I think it makes sense to define the 12 factors which in my point of view define a next generation All-Flash array, using latest NVMe and SCM technologies which provide the opportunity to deliver additional performance tiers for the most demanding apps while at the same time, providing flexibility in service level granularity and additional opportunities for cost reduction. So what do we mean by that.

EMC introduced back in 2008 the first SSD drives in an Enterprise High-End array. I remember back the moment as I saw the announcement. After some time with the possibility to reflect it got clear that we should have something like an automated storage tiering. Because only the intelligent data placement functionality would allow us to store very demanding data on fast low latency Flash drives and at the same time using traditional but as well much more economic drives for the rest of the data. A view month later in 2009 we publicly introduces the idea of FAST – Fully automated Storage Tiering. It was a milestone and for years used in our hybrid arrays up to the moment in time Generation 1 All-Flash Arrays entered the market and the industry started to build and use single tier SSD arrays.

Independent of the short term trend, it was absolutely clear that an intelligent data placement functionality would be critical for the design of the next generation All-Flash array. Why?

Because of new technologies like NVMe and especially Storage Class Memory (SCM) which will provide again much lower latencies. Therefore it is essential to have an intelligent and fully automated data placement enabled by advanced Machine Learning capabilities.

This is important mainly for two reasons. First SCM will be much more powerful as well as much more expensive compared to other NVMe NAND technologies for the next couples of years, therefore it is a must to have intelligent data placement enabled by Machine Learning. Secondly, to decide if you would like to apply data efficiency technologies, which even if they are inline and the latency impact was negligible compared to the latency of the medium in the past, in 2019 with SCM could have a measurable impact as the latency of the medium and protocol will get much smaller compared to the software latency as shown on the pictures.

Latency impact of NVMe, NVMeoF and SCM – This is why you need deicide today for a future ready architecture.

Intelligent Data Placement will let your Application profit from SCM and at the same time reduce your TCO

So in fact you will not be able to use the power of NVMe and SCM from an economic perspective without an intelligent data placement enabled by Machine Learning technologies. So let’s have a look at the 12 factors which a next generation all Flash Array needs to address.

12 factors of a next generation All-Flash Array:

  1. Scale-Out Architecture – Scale Performance and/or Capacity based on business needs
  2. End-to-End NVMe support and NVMeoF readiness– results in lowest possible latency
  3. Industry standard dual ported NVMe storage medium – important for availability and economic reasons
  4. Automated intelligent data placement using Applied Machine Learning Engine enabling the use of SCM (and thing after SCM) from an economic perspective
  5. Inline Data Efficiency Technologies – Thin provisioning, Deduplication and Compression
  6. Simple QoS – enables massive consolidation and guaranteed consistent Service Level’s
  7. Self-Service: Powerful Roll Based Access enabling self-service capabilities via GUI or secure automation via REST Full API
  8. Advanced Replication: Powerful Snapshot Technology providing “unlimited” copies without performance or capacity impact. Remote Replication, providing the flexibility to choose between Active-Active (SRDF/Metro), synchronous or asynchronous replication or even a combination to store date in more than two data centers.
  9. Integration into Cloud Framework and Application layers to simplify automation, operation to reduce cost and improve agility
  10. Non-disruptive automated hardware lifecycle management – providing a huge operational cost and risk reduction
  11. Trusted and Secured Platform fulfilling highest D@RE, Secure Audit Log, Roll-Based Access and Erasure Services requirements
  12. Last but not least highest availability, field proven 99.9999%, non-disruptive code upgrades and market leading global customer service support providing proactive intelligent maintenance tasks.

Now the good news, the described 12 Factor next Generation All-Flash Array is not only a vision or a design principle it’s reality! The new PowerMax is planned for general availability in May 2018 in two versions.

If you are interested to learn more about the new PowerMax, stay tuned: We will soon publish a more detailed description of the PowerMax.

In the meantime, please visit our website www.dellemc.com/en-us/storage/powermax.htm#collapse and contact your Dell EMC Account team.

Schreiben Sie einen Kommentar

Ihre E-Mail-Adresse wird nicht veröffentlicht. Erforderliche Felder sind mit * markiert.