Technology:Unpacked

File systems provide structure and a way to store data on computer systems. From the early days of FAT and NTFS, we now see a proliferation of file systems in the enterprise. Why do we have so many and what benefits do they offer? In this episode, Chris and Martin look at the evolution of the file system and why solutions have been developed to meet specific application requirements.

Much of file system development has been driven from the improvement of hardware technology. Larger HDDs meant supporting more file system metadata and so the evolution of FAT16 & FAT32. ZFS provides caching and other logging techniques by using solid-state media for additional journaling and recovery. ZUFS exploits Storage Class Memory to implement very fast local file systems.

The conversation also discusses FUSE, an open-source feature of Linux that enables anyone to develop a file system in userspace. File systems typically run in the kernel and are difficult to develop as a result. A file system needs to be bulletproof, adding to the development challenge.

Elapsed Time: 00:30:54

Timeline

* 00:00:00 – Intros* 00:02:15 – What is a file system?* 00:05:10 – File systems give scalability & resiliency* 00:05:50 – Are parallel and clustered file systems different? * 00:07:17 – Why are there so many different file systems?* 00:08:35 – Even tapes can have file systems (like LTFS)* 00:11:00 – ZuFS exploits SCM for a host-based file system * 00:13:15 – What do file systems like GPFS offer?* 00:15:00 – What is POSIX, what is a POSIX file system?* 00:17:00 – Many file systems are proprietary to operating systems * 00:17:30 – Why are file systems hard to develop?* 00:18:45 – FUSE – File System in User Space* 00:20:40 – ChubaoFS – one of many new file systems* 00:22:00 – Many file systems are niche – like HFDS (Hadoop)* 00:23:30 – What about new file system features for the user?* 00:25:00 – File systems can offer APIs for management (like backup)* 00:25:45 – WAFL is not a file system – or is it?* 00:27:00 – What are the risks in using new (open source) file systems?* 00:28:45 – Should we be looking at using alternative file systems?* 00:30:00 – Wrap Up

Related Podcasts & Blogs

* #92 – Introduction to Elastifile with Jerome McFarland* #52 – An Introduction to WekaIO Matrix with Liran Zvibel* #116 – Fixing Gaps In Cloud Storage with Andy Watson

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #3BQI.

This week, Chris and Martin review their thoughts from the previous discussion on COVID-19 and what we have learned since then. In what is an incredibly fast-moving situation, the way in which we viewed the impacts of the global epidemic were perhaps a little naive. We take another look at how vendors will need to market and sell to their clients and how even virtual events may be impossible to implement with so many staff dispersed to their homes.

The ultimate outcome is that for the foreseeable future, our way of working, selling and messaging will change.

Elapsed Time: 00:35:51

Timeline

* 00:00:00 – Intros* 00:01:20 – It’s all Chris Mellor’s fault (for getting it wrong)* 00:04:45 – Hairdressers will stay open! * 00:07:15 – Everyone’s a WFH expert* 00:09:00 – Social aspects of remote working will be hard to manage * 00:11:00 – Technical aspects will be a challenge too * 00:14:11 – Online tools are struggling to scale * 00:15:30 – Will virtual events happen or are they doomed too? * 00:17:16 – Rest in Peace restrictions * 00:21:00 – Not all physical events have been cancelled (yet) * 00:22:45 – Are sales and marketing roles going to evolve? * 00:24:15 – Is it time for acquisitions?  Which companies look cheap? * 00:27:00 – This is the perfect time for public cloud & technical debt * 00:30:30 – New products might never be launched to the market * 00:33:05 – Wrap Up

Related Podcasts & Blogs

* #146 – Coronavirus and Impacts on the Technology Industry

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #PGHR.

As reported by Chris Mellor, HPE recently announced Q1 2020 figures that show storage down 8% and servers down 15%. In light of these numbers, we look at HPE’s storage products, partners and overall strategy in the market.

HPE primary products are strong contenders, with XP8 for high-end/mainframe, 3PAR & Primera, Nimble and MSA. Outside of the main platforms, HPE has some data protection hardware and of course, tape.

To meet file and object storage needs, HPE partners with the likes of Qumulo, Scality, Weka and others. But finding details on these companies is hard. They’re not listed in any meaningful way through the HPE website.

How do these partnerships work? Does HPE simply introduce vendors to customers? Are products actively branded with HPE? Finally in this discussion, we look at Primera and how the transition from 3PAR has been managed. There’s lots to learn about moving from one platform to another and how that message is communicated to customers.

Elapsed Time: 00:34:37

Timeline

* 00:00:00 – Intros* 00:01:00 – HPE financials show Storage down 8%, servers 15%* 00:02:00 – HPE are a leader in the Gartner Primary Storage MQ * 00:02:40 – HPE’s primary storage portfolio, XP8, 3PAR, Primera, Nimble, MSA* 00:05:35 – HPE has always struggled with products in the NAS space* 00:06:45 – Is HCI a storage product? * 00:08:00 – Has HCI taken business from storage sales? * 00:09:25 – InfoSight is not a chargeable product – yet? * 00:11:00 – HPE has a wide storage partner ecosystem * 00:13:20 – What is the relationship between HPE and storage partners? * 00:15:15 – How are overlapping product solutions explained to customers? * 00:15:50 – What is the HPE partner strategy? * 00:18:55 – Could HPE brand partner products better? * 00:22:00 – Are there any product gaps in the portfolio? * 00:23:00 – HPE should acquire, but how would that process be managed? * 00:25:15 – Jeff Bezos – two products are better than none * 00:26:00 – What’s happening with Primera? * 00:29:55 – Conclusions – what have we learned? * 00:33:25 – Wrap Up

Related Podcasts & Blogs

* #77 – HPE Performance Insights with Ivan Iannaccone* Soundbytes 004: Discussing HPE Storage at HPE Discover* HPE Primera – First impressions on the new storage platform* Why is HPE 3PAR not moving to NVMe SSDs (yet)?* Understanding HPE’s Storage Product Portfolio* Nimble, HPE and 20-Year Old Architectures

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #CM90.

This week, Chris and Martin look at Key-Value stores and in particular, Redis, with Kyle Davis, Head of Developer Advocacy at Redis Labs. Key Value stores are at first glance a lightweight way to store structured data. As it turns out, the implementation of Redis includes significantly more features and functionality as well as multiple complex data types.

Redis runs in-memory and is typically used accelerate traditional databases or store state for web-based applications. The ability to store and retrieve data from memory enables fast response times for features like shopping carts or tracking gaming activity.

One interesting aspect of this conversation is how it highlights the transfer of managing data and state from an application into shared infrastructure. This is something micro-services can use to their advantage by simplifying programming and reducing complexity.

You can try Redis out online by visiting the Redis Labs website.

Elapsed time: 00:35:06

Timeline

* 00:00:00 – Intros* 00:01:35 – What is a Key-Value (KV) Store?* 00:02:10 – VSAM KSDS – mainframe did KV stores first!* 00:03:04 – Key – descriptor – Value – the value (obviously)* 00:04:00 – Redis is natively in memory with persistence to disk* 00:05:59 – What are the use cases for KV stores? * 00:09:00 – What data types can be stored in Redis? * 00:10:50 – KV stores allows complex data structures to be abstracted * 00:12:50 – Redis works well with micro-services * 00:14:20 – Who manages KV store platforms? * 00:16:00 – What is the standard Redis deployment model?* 00:18:34 – How do enterprise and open source versions differ?* 00:21:39 – SCM could vastly improve KV stores* 00:24:07 – Access to persistent memory will be different in public cloud * 00:26:20 – What is the programming model for accessing KV stores? * 00:29:49 – Redis allows data to be automatically expired and deleted * 00:34:20 – Wrap Up

Related Podcasts & Blogs

* #140 – Introduction to Document Orientated Databases* #98 – Shared Storage for Scale-out Databases* Databases are the next battleground for Public Cloud* Stellus delivers scale-out storage with NVMe and KV Tech

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #QFQV.

In this episode we take a look at the impact of COVID-19, commonly known as the Coronavirus. Since the outbreak in China towards the end of 2019, tens of thousands have been infected, thousands of people have died. As the spread of the disease moves outside China, we look at the effect on business, conferences and the positive benefits that may occur with reduced travel.

During the podcast we make some specific references. Here’s a link to the report on supply chain impacts. Here’s a link to the article on the reduction in emissions over China. This one covers the Corona beer story and this the Zoom story. Finally, here’s a link to the Force Majeure article.

Elapsed time: 00:42:38

Timeline

* 00:00:00 – Intros* 00:01:30 – What is the coronavirus?* 00:03:14 – What are the social implications of the virus spreading? * 00:04:00 – Has Chinese New Year masked the impact of the virus? * 00:06:05 – What are the business implications?* 00:07:15 – Travel bans are starting to be put in place* 00:08:45 – Businesses will need continuity plans* 00:10:15 – Will we see more hacking and ransomware occurring? * 00:11:30 – DR plans need to focus on both technical and personnel areas* 00:13:15 – Will we learn anything new about remote working?* 00:13:45 – Vendors are cancelling or postponing conferences* 00:15:15 – No big infrastructure vendors have cancelled – yet* 00:17:45 – Who will cancel an event – the venue organiser or vendor? * 00:18:34 – Could Force Majeure apply to cancellations & contracts? * 00:21:00 – Which big vendor will blink first?* 00:22:15 – Will see a change in how we share information & knowledge?* 00:26:00 – Chinese emissions are down as a result of the virus lockdown * 00:28:30 – Back to the supply chain – is storage being affected? * 00:30:50 – Could we see companies looking to be more efficient with existing resources? * 00:34:00 – What can we do to help – other than washing hands….? * 00:37:45 – Disaster recover is more than just recovering technology * 00:39:00 – Shares in Zoom are going up – but not the conference software people! * 00:39:30 – And Corona beer sales are down…. * 00:41:30 – Wrap Up

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #BR31.

This week Chris and Martin discuss the announcement of Google Cloud partners offering Anthos Ready Storage. Anthos is Google’s on-premises cloud infrastructure running Kubernetes-based containers. Platform users can now deploy locally in their data centre, on local hardware, while using the GCP management plane.

What is the benefit of having storage certified for Anthos? The discussion looks initially at why containers need persistent storage, moving on to examine the profile of the first ARS certified storage companies. Is this a move simply to gain more access to enterprise customers? There’s lots questions in this discussion, so if you’re in a position to provide the answers, drop us a note to get involved.

Elapsed Time: 00:31:34

Timeline

* 00:00:00 – Intros* 00:00:55 – What is Anthos?* 00:03:00 – How does this solution compare to AWS Outposts?* 00:04:40 – Containers need persistent storage* 00:07:00 – Persistence is not futile!* 00:07:30 – What is Anthos Ready Storage (ARS)?* 00:09:45 – Who are the ARS vendors?* 00:11:50 – What are the ARS criteria?* 00:12:35 – Vendors need to support CSI* 00:15:40 – Are the qualification criteria that rigid?* 00:16:40 – Kurian & Kurian – are they working together?* 00:20:30 – Is ARS a beachhead to get into the enterprise?* 00:21:55 – Containerise your workloads!* 00:24:00 – How is Anthos demonstrating cost savings?* 00:28:10 – Storage is the Switzerland of your data centre* 00:30:15 – Wrap Up

Related Podcasts & Blogs

* What is the Container Storage Interface (CSI)?* Storage Management and DevOps* #53 – Persistent Storage and Kubernetes with Evan Powell* #143 – Storage Adoption Myths & Realities

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #94KC.

This week, Chris and Martin chat to Gary Breder (Director of Cloud and Services Product Marketing) and Diane Clay (Sr Manager of Cloud and Services Product Marketing) from Hitachi Vantara about storage as a service. While we’ve had different consumption models in the past, does StaaS offer anything different?

Perhaps the key takeaway from this conversation is the use of service-based offerings as a way to financially engineer more attractive cost structures to businesses. The vendor shoulders the risk with the customer by putting resources on the floor with an expectation of usage at some point.

What’s also clear is that much of the responsibilities have to be set out in contract form. This sees the transfer of management tasks covering fault diagnosis and resolution as well as capacity planning being shifted to the vendor. All of these processes are wrapped in detailed service level agreements.

Storage, data protection or any other offering “as a Service” isn’t for every business. However, as new projects are implemented, this provides a decision point to look at whether service offerings can replace the traditional management practices most common today.

To find out more about Hitachi Vantara’s service offerings, log onto the Hitachi Vantara website and follow the links for Service Offerings and Products.

Elapsed Time: 00:38:46

Timeline

* 00:00:00 – Intros* 00:01:30 – “as a service” models are gaining traction* 00:02:25 – Buy, lease (capex) and now opex* 00:05:00 – Traditional enterprises continue to self-manage* 00:06:30 – “as a service” could be used for new projects* 00:09:45 – However, on-prem and cloud are different consumption models* 00:11:55 – Storage vendors & customers need to share risk* 00:12:00 – Minimum purchases exist – not the same as public cloud * 00:15:00 – Traditional contracts just deferred the costs* 00:17:00 – Hitachi uses Smart Data Center to track usage* 00:18:30 – How does the “as a service” offering differ to purchased storage?* 00:22:00 – What, exactly, are the service level offerings?* 00:23:20 – But, what have I actually gained by on-premises as-a-service?* 00:25:55 – Where’s the dividing line between customer & vendor roles? * 00:28:45 – What happens when things go wrong?* 00:32:00 – What other services could vendors offer – data protection? * 00:35:00 – Is most of the “as a Service” simply financial re-engineering? * 00:36:55 – Wrap Up

Related Podcasts & Blogs

* #137 – Hitachi Next 2019: Wrap Up with Jonathan Martin* Backup as a Service* On-premises infrastructure – as a service

Guest Bios

Gary Breder, Director of Cloud and Services Product Marketing leads the transformation of Hitachi in the development, marketing and go to market of our outcome-based cloud and IoT solutions and managed services. What people like most are the compelling messages he creates, his passion for driving customer value and company efficiency, and his effectiveness as a company evangelist.

Diane Clay, Sr. Manager of Cloud and Services Product Marketing drives the development and introduction of Hitachi Vantara solutions and end-to-end offerings that ...

This week, Chris and Martin chat to Paul Stringfellow, CTO at Gardner Systems, about the realities of technology adoption. We hear lots of talk about Kubernetes, HCI, all-flash and even SCM technologies. But are all organisations deploying these technologies from day one?

In reality, businesses deploy technology in a timeline that matches a normal distribution curve. As detailed in “Crossing the Chasm” by Geoffrey Moore, there’s a gap between early adopters and the mainstream IT organisations.

What causes this inertia? Some of the challenges relate people and process, however there’s also a desire to implement more mature products. Who really wanted to use Microsoft Windows before version 3.0…

You can follow Paul on Twitter at @techstringy and of course, also listen to his excellent podcast, Tech Interviews. Paul also writes a regular blog.

Elapsed Time: 00:34:46

Timeline

* 00:00:00 – Intros* 00:02:00 – Technology adoption myths vs the reality* 00:05:30 – Are enterprises any better at adopting new tech?* 00:06:40 – What do we mean by digital transformation?* 00:10:00 – Who doesn’t like a new car?* 00:11:10 – Does technology adoption allow staff “redistribution”?* 00:12:15 – S&P 500 – 2 week replacement schedule* 00:13:20 – Focus on data and the data model, not just hardware * 00:16:20 – Do the CIO/CTO roles exist in smaller organisations? * 00:17:50 – Crossing the Chasm – why does adoption inertia occur? * 00:19:10 – Always wait for the 3.0 version of the product!* 00:19:50 – Dealing with”day 2” implementation* 00:23:55 – Vendors only turn up at licence renewal time!* 00:26:00 – “Keeping the lights on” is still an issue for IT teams* 00:29:05 – Always ask – does this technology solve a problem?* 00:30:40 – Make white papers free to download – no registration!* 00:32:40 – Wrap Up

Related Podcasts & Blogs

* #104 – Creating a Data Management Strategy with Paul Stringfellow

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #NA5Y.

This week, Chris and Martin discuss the automation of storage. In a follow-on from the Dell EMC presentation at Storage Field Day 19, this conversation looks at the challenges of opening storage provisioning and management to the wider developer community.

In the past, SRM tools have focused on providing the “record of truth” for storage and well as the interface to provision resources to hosts. Dell EMC tried to create a separate data and management plane with ViPR, but the industry didn’t adopt this approach – potentially because Dell/EMC continued to own the technology.

Today, Dell EMC is focusing on automation through tools like Ansible and Python. How will enterprises continue to reduce over-provisioning and protect production systems? How should billing work? What about documentation? These questions and more, answered in this week’s podcast.

Elapsed Time: 00:31:17

Timeline

* 00:00:00 – Intros* 00:00:46 – Dell EMC & DevOps at TFD* 00:01:45 – The Storage Admin is a gatekeeper* 00:03:00 – SRM software has been universally terrible* 00:04:43 – Dell acquires iWave to create ViPR* 00:06:35 – Should we use wrappers or simply script?* 00:08:49 – Storage platforms must have multi-tenancy* 00:09:50 – How is storage billing achieved in the public cloud?* 00:11:00 – Shout out to DFHSM!* 00:13:02 – How efficient are storage system APIs?* 00:15:57 – How are RBAC and credentials managed?* 00:18:10 – What happens to documentation with automation?* 00:20:52 – Storage desired state and actual state* 00:23:27 – Which vendors are delivering best practice automation?* 00:24:09 – Do we simply need better (software-based) storage products? * 00:27:27 – Should Dell be doing ViPR or just simple automation? * 00:28:00 – Could standards’ bodies do more?  * 00:30:50 – Wrap Up

Related Podcasts & Blogs

* Storage Management and DevOps* #129 – Choices for Persistent Container Storage with Niraj Tolia* The Need for APIs in Storage and Data Protection* Dell EMC Presents DevOps at Storage Field Day 19

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #LTVN.

This week Chris and Martin are joined by Erik Kaulberg, Vice President at Infinidat. Erik has appeared on the podcast before and this time is here to talk about how we build storage systems of the future. You can tell from the introductions that we recorded this episode towards the end of 2019 – we haven’t been transported into the future!

With many choices in new media, system builders have a wide choice of persistent storage from which to build new architectures. There’s NAND flash, traditional hard drives and a range of technologies such as SCM and MRAM.

In an ideal world, we’d simply build systems from the best performing media, but cost and resiliency factor into the equation. Building future systems will continue to be a balancing act of choosing the right media for the right data type and matching the two efficiently.

Will there be one platform to rule them all? Probably not, but we may see differences in core and edge designs.

You can follow Erik on Linkedin and find more about Infinidat at https://infinidat.com/

Elapsed Time: 00:36:15

Timeline

* 00:00:00 – Intros* 00:02:54 – What are the future challenges?* 00:05:05 – Is any data truly unstructured?* 00:06:07 – What new media types are we seeing?* 00:08:40 – Application rewrites may be stalling new technology* 00:09:50 – All media sits on a graph combing cost, performance, endurance & capacity* 00:12:50 – Is it practical to put all data onto the fastest media?* 00:15:30 – Will we see more “media defined” storage systems?* 00:17:47 – Is more data moving to all-flash systems over time?* 00:20:26 – Hyper-scalers worked out their storage architectures many years ago* 00:23:31 – How will encryption be built into future storage designs?* 00:27:40 – Should we build one “platform to rule them all”?* 00:31:53 – We’re not going to see an end to new platforms & solutions* 00:33:23 – Will all future storage systems be based on disk or flash?* 00:35:12 – Wrap Up

Transcript

Related Podcasts & Blogs

* #113 – The Expanding Storage Hierarchy with Erik Kaulberg* #130 – Making Money in the Storage Business* #138 – Storage Predictions for 2020 (Part I)* Building a Golden Data Repository

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #I4FE.