EP08 – Dell EMC PowerMax: NVMe All Flash Arrays, Storage-Class Memory and Machine Learning with Vince Westin



In this 8th episode of the TECHunplugged Podcast, hosts Max Mortillaro (@darkkavenger) and Arjan Timmerman (@arjantim) are joined by Vince Westin (@ApplyITNow),IT Architect and Dell EMC Technical Evangelist working with the PowerMax and VMAX high-end storage arrays.

Show schedule:

  • 00:00 Intros & presentation
  • 01:20 A recap about the Dell EMC PowerMax
  • 02:30 PowerMax was introduced in May 2018, what is the current adoption rate?
  • 04:00 Is there a difference between the two PowerMax models? Are there any reasons why choosing one over the other?
  • 08:22 VMAX has been known for its robust architecture, what about PowerMax, has there been any major improvements?
  • 11:45 Where is PowerMax now in Dell EMC’s product range? Is it displacing the VMAX?
  • 14:30 What about SCM (Storage-class Memory) support and Optane / 3DXPoint? And is 3DXPoint really making a breakthrough in the enterprise storage world?
  • 20:00 Max has a brain fart around NVMe / SCM proximity to the CPU, Vince fixes that and expands on the value that matters, i.e. the backend latency
  • 21:55 Vince & Max discuss current form factors and the future of storage arrays looking at what SCM will bring next (EDSFF format), and mixing different memory / flash drive types to drive down the $/GB cost while delivering outstanding performance
  • 25:15 Is there a huge pricing gap between a VMAX All-Flash and a PowerMax? Is there an incentive to go directly with PowerMax?
  • 26:15 Is CloudIQ going to support PowerMax?
  • 26:48 What about PowerMax Machine Learning engine? ML is a hot topic, but also a very abused buzzword, what is Dell EMC doing with it? Vince discusses about getting the best performance possible out of the array, and avoiding any features that might unnecessarily increase latency (for example not compressing data on SCM to avoid latency penalties)
  • 30:25 Closing the discussion & Thanks

EP07 – The Elastifile Story – Discussing scalable distributed cloud file systems with Jerome McFarland



In this 7th episode of the TECHunplugged Podcast, hosts Max Mortillaro (@darkkavenger) and Arjan Timmerman (@arjantim) are joined by Jerome McFarland (@JeromeMcFarland), VP of Marketing at Elastifile.

Show schedule:

  • 00:00 Intros & presentation
  • 01:15 Jerome McFarland introduces Elastifile
  • 01:46 Arjan Timmerman returns on the changes at Elastifile: new CEO, collaboration with Google Cloud
  • 02:32 Jerome covers the changes and focus of Elastifile following CEO change
  • 03:33 Arjan asks to understand what kind of cloud setups are supported by Elastifile
  • 04:00 Jerome replies and explains the predominantly public cloud orientation of Elastifile
  • 04:55 Arjan asks if that is due to customer base requirements and what about hybrid cloud
  • 05:30 Jerome covers the question and discusses a bit about Elastifile customers
  • 06:45 Arjan asks about the multi-cloud story
  • 07:27 Jerome explains the cloud agnostic approach at Elastifile and the links with Google Cloud
  • 08:05 Max wakes up and asks about whether it makes sense to run Elastifile on multiple public clouds simultaneously (for example for DR / replication)
  • 08:30 Jerome acknowledges that they predominantly see single cloud deployments, but see a trend where customers also design for multi-cloud from the ground up
  • 09:47 Arjan and Jerome talk about data mobility features and how to import data onto Elastifile, covering also the filesystem structure and object format
  • 12:10 Max asks about cost efficiencies when data is stored on an Elastifile cloud file system but not actually used
  • 12:41 Jerome covers the question and explains the data structure
  • 13:26 Arjan asks about the relationship between Elastifile primary scale out file storage solution and the object storage story
  • 14:12 Jerome returns on the question, covering the object storage involvement in Elastifile solution delivery, also covers bursting vs continuous use
  • 15:40 Arjan whether there are any predominant verticals leveraging Elastifile
  • 16:15 Jerome covers the verticals, specifically data-centric use cases, but also Kubernetes support, and Machine Learning
  • 17:15 Max returns on data-centric use cases and asks whether Elastifile can be a good platform for HPC based projects
  • 17:55 Jerome explains that this is definitely an use case and provides more info on already published information with one of their customers
  • 18:30 Arjan asks for Jerome if there is anything we haven’t yet covered in the discussion
  • 18:56 Jerome covers the cloud-native origins of Elastifile, explaining that it isn’t a file system design retrofitted for the cloud but built to leverage cloud-native and cloud-scale features while delivering performance
  • 20:22 Max asks whether Elastifile have the impression of having displaced specific existing use cases or if these are net new use cases
  • 20:55 Jerome explains that it’s a mix of both, new use cases with analytics & containers, and offloading of existing use cases to the cloud
  • 21:49 Arjan asks whether customers are coming to Elastifile out from traditional file-based storage systems
  • 22:40 Jerome says that there is no specific vendor being displaced and that it’s a broad range of on-premises use cases covered
  • 23:30 Wrap up and thanks

EP06 – AIOps: Augmented Intelligence for ITOps and DevOps – with Dominic Wellington



In this 6th episode of the TECHunplugged Podcast, hosts Max Mortillaro (@darkkavenger) and Arjan Timmerman (@arjantim) are joined by Dominic Wellington (@dwellington), Director of Strategic Architecture at Moogsoft.

You can find Dominic’s blog at http://findthethread.postach.io. The top post right now is actually pretty relevant to this conversation, since it talks about augmented intelligence. Permalink: http://findthethread.postach.io/post/the-road-to-augmented-intelligence. Dominic also covered how to think about Ops automation and learning from the Dev side of things, at devops.com: https://devops.com/ops-the-other-devops/.

Show schedule:

  • 0:00: Usual introductions & welcoming
  • 2:02: Dominic explains what Moogsoft does and what AIOps stands for
  • 5:05: Max covers the struggles of IT Operations teams and the issues of “many pains of glass” (and struggles a bit with some echo on the line)
  • 6:17: Dominic explains how to address such issues: reducing the noise, automating routine activities, correlating seemingly unrelated events
  • 09:20: Dominic talks about MTTI (Mean Time To Innocence)
  • 10:10: Max talks about lack of event correlation in traditional IT environments and silo-induced blinkers
  • 10:45: Max asks what goes beyond event correlation in AIOps
  • 11:15: Dominic covers collaboration aspects, identifying root cause, and how machine learning helps identify potential reoccurrences and bring things faster into context
  • 13:40: Arjan and Dominic cover Moogsoft’s place in the enterprise monitoring landscape especially in environments who already use multiple tools
  • 15:31: Dominic explains how AIOps will enable a kind of “Level-0 support”, i.e. AI-based resolution for frequent, well-understood issues, leaving time for humans to focus on what matters
  • 16:57: Max is surprised about this statement and asks Dominic how Moogsoft would trigger this automated resolution of issues, followed by Dominic’ reply
  • 18:21: Arjan brings up the topic of DevOps, sparking a conversation around traditional ITOps having to adapt to DevOps, the use of ChatOps etc. Dominic covers innovations brought into Moogsoft from the DevOps world
  • 23:15: Arjan and Dominic discuss whether AIOps can help bridge the culture & process gap between Dev and Ops
  • 26:55: Arjan asks Max about the customer perspective on DevOps adoption (and relevancy of AIOps) in large corporate organisations, followed by an overview of the reality of Enterprise IT in large orgs and a slower moving pace
  • 29:15: Dominic covers how AIOps adoption begins in organisations
  • 30:50: Max has an epiphany and thinks AIOps is the One Ring to Bind Them [monitoring tools] All, but in a cool way
  • 32:05: Closing thoughts, and a heads-up to Sysadmin Day

EP05 – ARM-based architectures, Object Storage and OpenIO with Enrico Signoretti



In this fifth episode of the TECHunplugged Podcast, hosts Max Mortillaro (@darkkavenger) and Arjan Timmerman (@arjantim) are joined by an old friend & special guest: Enrico Signoretti (@esignoretti), the original founder of TECHunplugged. Enrico left TECHunplugged in early 2017 to join OpenIO as Head of Product Strategy.

In this episode we talk about Enrico’s role in OpenIO, the latest news and their Raspberry Pi kit, how that relates to the rise of ARM-based architectures & their impact on the data center. We also cover OpenIO object storage solution from a capability and performance perspective. Finally, because life can’t be only about work, we briefly cover Enrico’s holiday plans, gelato, and try to invite ourselves for a summer break on his boat.


EP04 – HPC / GPU virtualization on VMware vSphere with Gina Rosenthal and Tom Hollingsworth



In this episode of the TECHunplugged Podcast (recorded live at Dell Technologies World in Las Vegas), host Max Mortillaro (@darkkavenger) is joined by two awesome industry friends: Gina Rosenthal (@gminks), a VMware Product Marketing Manager working on VMware on AWS, and Tom Hollingsworth (@networkingnerd), Organizer at Tech Field Day. Gina, Tom and Max cover the challenges with HPC (High Performance Computing), then how HPC / GPU virtualization support improvements in VMware vSphere are helping research organizations be leaner / better leverage investments thanks to GPU sharing, and finally how well-known virtualization advantages can now also benefit mission-critical HPC systems.


EP03 – Midrange Storage and NVMe with Pierluca Chiodelli and Chin-Fah Heoh



In this episode of the TECHunplugged Podcast recorded live at Dell Technologies World in Las Vegas, host Max Mortillaro (@darkkavenger) is joined by Chin-Fah Heoh (@storagegaga) an independent storage consultant hailing from Malaysia, and Pierluca Chiodelli (@chiodp),Vice President of Product Management & Customer Operation for Midrange Solutions at Dell EMC, to discuss about midrange storage and NVMe.


EP02 – Oracle Ravello: Lift & Shift or gateway into Oracle Cloud Infrastructure Services – a discussion with Abhinav Gupta



In this episode of the TECHunplugged Podcast, hosts Max Mortillaro (@darkkavenger) and Arjan Timmerman (@arjantim) are joined by Abhinav Gupta, Senior Director of Product Management at Oracle Cloud Infrastructure.

Historically, Ravello has been much loved by VMware / Infrastructure / data center community folks (especially among the vExpert subgroup) for the ability to create labs and test environments in the cloud at very affordable prices. After the Oracle acquisition, Ravello’s own HVX hypervisor was engineered to run on bare-metal and on nested virtualization with hardware extension assists, which both allow for considerably better performance and makes Ravello finally production-grade capable.

In this discussion with Abhi, we also set to explore the use cases for Oracle Ravello, their customers and what is the future of Ravello as well as the broader picture with Oracle Cloud Infrastructure as a challenger to well-known public cloud operators.

We hope you’ll enjoy this episode and will survive Max’s French/Italian and Arjan’s Dutch accents 🙂


EP01 – Talking Open Convergence and Business Outcomes with Datrium – André Leibovici, Keith Townsend, James Green, Max Mortillaro



In this inaugural episode of the TECHunplugged Podcast recorded live at TECHunplugged New York 2018 on 22-Mar-18, Max Mortillaro (@darkkavenger) is joined by Datrium’s Field CTO, André Leibovici (@andreleibovici), as well as industry influencers Keith Townsend (@ctoadvisor) and James Green (@jdgreen). During the talk we cover the data center journey from the early converged infrastructure days up to open convergence, about business outcomes and even into container consumption models. We hope you’ll enjoy the format!