Technology:Unpacked

This week, Chris and Martin speak to David Koppe, Director of Information Strategy at MongoDB. The discussion covers document-orientated databases, or simply document databases, and their appeal to the enterprise.

Document databases differ from traditional relational databases in that the data is stored within each “document” as a series of key-value pairs. In this instance a document is not to be confused with a Word or PDF file. Unlike relational databases, document records don’t have to be “symmetrical” and aren’t required to contain every field (or a null value for empty fields). This makes the development of applications much more flexible than traditional relational schema.

The team dig down and look at further differences, including exactly how documents are stored. In the case of MongoDB, this is JSON or on disk as BSON. Document databases aren’t required to offer ACID compliance, but this may be a requirement for the enterprise, where transactional applications expect strong consistency (think CAP theorem and PACELC).

Finally, Dave provides some background on the types of applications developed with document-orientated databases. For more information on MongoDB, check out https://www.mongodb.com/.

Elapsed Time: 00:43:34

Timeline

* 00:00:00 – Intros* 00:02:00 – What is a document-orientated database?* 00:05:10 – How are document databases related to key-value stores?* 00:07:10 – How do document databases compare to relational or hierarchical databases?* 00:09:30 – Document databases are like sparse relational databases* 00:13:10 – We are in a polyglot database world* 00:15:35 – NoSQL means “Not only SQL”* 00:18:15 – Diving Deeper – What are JSON and markup languages?* 00:19:25 – MongoDB uses BSON to enforce type fidelity* 00:20:55 – Do document databases offer ACID compliance?* 00:25:00 – CAP Theorem and PACELC Theorem* 00:26:20 – Eventual vs strong consistency* 00:30:00 – How does indexing and searching work with sparse records?* 00:32:40 – How do storage engines work for databases?* 00:36:35 – What tools are used to interact with document databases?* 00:39:00 – How are companies using document databases?* 00:42:10 – Wrap Up

Related Podcasts & Blogs

* #98 – Shared Storage for Scale-out Databases with Walt Hinton* Databases are the next battleground for Public Cloud* Database Replication is Hard

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #AF6G.

This is the second of our predictions shows for 2020 that anticipates what we can expect in enterprise storage for the year ahead. This episode covers solutions and vendors. We open with a discussion on cloud-native storage, which saw a lot of traction in 2019. Will the future be one of these companies or as Chris says, should we just rely on VMware to underpin our container environments?

The discussion moves on to object storage and in particular a look at a company called MinIO. As an open-source solution with S3 compatibility, MinIO could be the future of general object storage while the other vendors focus specifically on market segments.

Next we look at vendors. Whatever happened to Dell EMC Midrange.NEXT? An announcement was due at the end of January and that looks to be delayed. As the market for enterprise storage consolidates, could HPE need to go on a buying spree?

Finally we look at the data management market. There’s the chance of consolidation or acquisitions here. As it turns out, Veeam has been acquired since we recorded (7th Jan, with Veeam announcement on the 9th). Well done to Chris M for his prescient opinion!

Elapsed Time: 00:31:56

Timeline

* 00:00:00 – The Briefest of Intros* 00:00:30 – Solutions – it’s all about solutions!* 00:00:35 – How mature is cloud-native storage?* 00:02:10 – Should we just go with vSphere?* 00:04:00 – Data persistence is “late” in containerisation* 00:06:56 – Cloud-native storage is in a “wait and see” phase* 00:07:30 – Where is the Object storage market headed?* 00:08:30 – Min-aye-oh or Min-ee-oh?* 00:11:54 – Do Microsoft want S3 compatibility (on Azure)?* 00:13:15 – VMware seems to be doing OK* 00:13:40 – Whatever happened to Dell EMC Midrange.NEXT?* 00:15:30 – Introducing launch countdown shirts* 00:16:22 – Captain Canada on a high wire!* 00:17:50 – What further vendor consolidations will happen in 2020? * 00:18:49 – Chris loses focus – does anyone read show notes?* 00:19:50 – What is HPE’s storage partnership strategy?* 00:21:30 – HPE to buy everyone!* 00:24:00 – How could we forget Whiptail?* 00:24:52 – Is the data management market overcrowded?* 00:27:18 – Is data management too vague a term to align companies?* 00:30:15 – How good was our Veeam foresight?* 00:31:17 – Wrap up

Timeline

Related Podcasts & Blogs

* #138 – Storage Predictions for 2020 (Part I)* #82 – Storage Predictions for 2019* Understanding HPE’s Storage Product Portfolio* Thoughts on transitioning to Dell EMC Midrange.NEXT

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

This week Chris and Martin are joined by Chris Mellor for what is turning into an annual look at storage technology for the year ahead. We should point out that this is only for enterprise storage and not a forecast on the world in general!

We start with a discussion on media. At the macro level, capacities continue to grow for SSDs and HDDs year on year. This trend is expected to continue, but what micro-level advancements are being made?

QLC increases in layers, while PLC (penta-level cell) flash is being mooted as increasingly more practical. Hard drives are at 20TB and growing, but suffering the challenges of throughput. Will 2020 be the year we see multi-actuator take off?

Does old media continue into decline as the new tape? Could punched cards make a comeback? We think it’s unlikely.

The final part of the discussion turns to systems and composable systems in particular. Composability enables servers to be dynamically built from their constituent parts. The technology is in the early days and maybe not ready yet for mass adoption. Similarly, ARM processors look to be on the cusp of data centre relevance.

Part II of this recording will follow next week.

Elapsed Time: 00:38:00

Timeline

* 00:00:00 – Intros * 00:01:40 – Storage predictions only please! * 00:02:30 – Kioxia fire in the clean room * 00:04:00 – NAND prices will go up in 2020 * 00:05:15 – 128-layer flash on the way * 00:06:55 – Will PLC flat be “a thing”? * 00:10:40 – Enmotus have developed a hybrid SSD * 00:11:40 – 20TB+ hard drive capacity, what’s next? * 00:13:20 – Are hard drives the new tape? * 00:14:35 – Is the physical reliability of tape good or bad? * 00:16:55 – Has much happened in 12 months – 10 years? * 00:19:30 – Migration to SSDs wasn’t a big architectural change * 00:20:40 – Uh oh!  Gartner Hype Cycles! * 00:22:00 – Bring back punched cards! * 00:23:20 – What has happened with composable systems? * 00:26:34 – Liqid could be an interesting composable solution * 00:27:35 – Obligatory mainframe reference! * 00:30:00 – ARM servers could be popular in 10 years’ time. * 00:31:10 – Containerisation – moving from x86 to ARM * 00:34:05 – Is hyperscaler IT diverging from enterprise IT? * 00:37:40 – Wrap Up

Transcript

Related Podcasts & Blogs

* #136 – The End of the Year Show 2019* #82 – Storage Predictions for 2019* #37 – State of the Storage Union with Chris Mellor* Are ARM Processors Ready for Data Centre Primetime?* Liqid’s PCIe Fabric is the Key to Composable Infrastructure

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

In the first podcast episode of 2020, Chris meets with Jonathan Martin, CMO at Hitachi Vantara to reflect on experiences at Hitachi NEXT 2019. This episode was recorded live onsite at the event in October 2019.

Hitachi Vantara and Hitachi Consulting are planned to merge in 2020 (press release). This is another step in the evolution towards making Hitachi Vantara a solutions company, or as Jonathan describes it, an intellectual property creation organisation.

Will Hitachi continue to sell infrastructure? The simple answer is – yes. Storage, CI and HCI platforms will continue to form the basis of solutions for customers. “as a service” offerings will bring a more cloud-like experience (and we’ll cover that later in the year).

A recording of the NEXT Day 2 general session is available here (registration required).

Elapsed Time: 00:17:25

Timeline

* 00:00:00 – Intros * 00:00:45 – Vegas T-Shirt cannons! * 00:02:00 – Hitachi Vantara is merging with Hitachi Consulting * 00:05:25 – Disney uses Lumada for ride analysis * 00:07:30 – Edge is evolving because of data creation * 00:10:15 – Hitachi will continue to sell hardware solutions! * 00:15:00 – How does Hitachi bring to halves of the business together? * 00:16:45 – Wrap Up

Related Podcasts & Blogs

* #131 – Hitachi Next 2019: Content Intelligence Update* #127 – Hitachi Next 2019: CI & HCI Solutions Update* #125 – Hitachi Next 2019: Ops Center* #123 – Hitachi Next 2019: VSP 5000

Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #8C86.

In this final, final show for 2019, Martin and Chris review both the year and decade just passed. How did predictions work for 2019? What technology has come and gone over the decade and what can we expect in the next 10 years?

Downloads have increased exponentially again over the year. Thank you to everyone who continues to download and listen to us. The most popular posts continue to be those focused on vendors and products. Top of the list is #102 – May Midrange Madness, with the VAST Data dilogy (#105 & #106) close behind.

Our rather generic predictions of 2019 proved to be mostly accurate. QLC NAND and flash in general is taking over. This a theme of 2019 (for QLC) and for flash over the decade. Looking back to 2009/2010, vendors like Xsigo, Nirvanix and Ocarina were presenting at Tech Field Days. 1TB drives were state of the art – a volume of data that can now be purchased on a single microSD card from Micron.

What can we expect in the next decade? Probably more of the same challenges around data management and compliance, with increasing focus on features delivered through software. It’s going to be another interesting decade!

Elapsed Time: 00:30:00

Timeline

* 00:00:00 – Intros* 00:01:00 – Has the decade ended?* 00:01:48 – A big thank you* 00:03:08 – How accurate was the predictions show?* 00:04:00 – QLC, SCM, 3D-XPoint and NVMe-oF* 00:07:30 – How many storage products are not problem-focused?* 00:09:20 – What did we talk about in 2019? What was most popular?* 00:12:39 – What were the technologies of 2009/2010?* 00:15:00 – From TFD – Xsigo, Nirvanix & Ocarina all gone* 00:16:26 – 2009 – 1TB drives – today 16, 18, 20TB* 00:18:00 – SDS has been a success via the back door* 00:20:00 – The 2010s as seen the rise of flash and NVMe* 00:21:14 – Storage administration is no longer a desired skill* 00:23:10 – What can we expect over the next decade?* 00:25:48 – Data management will be the biggest ongoing challenge* 00:28:11 – Wrap Up

Related Podcasts & Blogs

* #102 – May Midrange Madness with Chris Mellor* #105 – Introduction to VAST Data (Part I) with Howard Marks* #106 – Introduction to VAST Data (Part II) with Howard Marks* The Biggest Storage Trends of 2019

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #xn9z.

This episode was recorded live at the Datrium offices in Silicon Valley. Chris chats to fellow Brit Simon Long about the announcement of DRaaS Connect and Instant RTO. Both are new features/products in the data mobility and DR space and complement Datrium’s existing storage and virtualisation offerings.

Enterprises traditionally implement DR failover for virtual environments using techniques like storage replication and SRM. Datrium has offered DVX, Cloud DVX and Automatrix as tools to abstract and deliver DR capabilities to virtual machines. For customers that can’t immediately move to DVX, DRaaS Connect provides the capability to move applications into cloud-based VMware solutions like VMC (VMware Cloud on AWS).

The interesting aspect of this announcement is that Datrium is providing capabilities to build “on-demand” DR, with the option to move to DVX in the future. Many businesses can’t justify running an entire copy of their business, so VMC with a DR capability enables DR when required. This vastly reduces costs and eliminates rarely used equipment.

Datrium also announced the availability of Instant RTO. This enables customers to bring workloads back much more quickly, especially when using “pilot light” VMC environments.

You can find more on Datrium at datrium.com (or check out our previous podcast recordings from the links below). You can follow Simon on Twitter or check out his blog at http://www.simonlong.co.uk/blog/

Elapsed Time: 00:34:33

Timeline

* 00:00:00 – Intro* 00:02:15 – What is a double VCDX?* 00:02:50 – How do traditional enterprises do DR in virtual enterprises?* 00:04:50 – What tools & techniques do enterprises use?* 00:05:55 – Physical LUNs don’t align well to logical applications* 00:07:17 – Failback frequently gets forgotten* 00:08:20 – VMC is a great solution for DR/failover* 00:10:30 – DR should be assigned by SLAs* 00:12:20 – With VMC, you pay for DR only when you use DR* 00:13:50 – Datrium DRaaS is a set of complementary services* 00:16:34 – Recovery into VMC requires no VM conversion* 00:20:25 – DRaaS Connect – DR without the need for DVX* 00:23:10 – DRaaS Connect has no dependency on source storage* 00:26:30 – DRaaS Connect could be a tool for cloning environments* 00:30:00 – Instant RTO is now GA 00:33:30 – Wrap Up

Related Podcasts & Blogs

* #101 – Datrium Automatrix with Brian Biles and Tim Page (Sponsored)* #63 – Datrium CloudShift* #48 – Introduction to Datrium DVX (Sponsored)* Datrium Delivered Data Mobility with Automatrix

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #bfh4.

This week Martin and Chris catch up again with Hal Woods, CTO at Datera. In this sponsored episode, we look in detail at the Datera Data Services Platform and how it differs from traditional storage. DSP is a software-defined solution, which is purchased either as a software licence or in conjunction with partners.

Software-defined storage has been around for some time, but hasn’t really offered the disaggregation from hardware performance that applications need in the enterprise. DSP enables administrators to set performance and availability policies that are then implemented by DSP, based on the hardware available.

This disaggregation distinction is important because over time, hardware gets refreshed but application performance needs to remain consistent. DSP delivers a “perpetual storage infrastructure” from which hardware is added or removed as part of ongoing deployment.

You can learn more about DSP at the Datera Website, including blogs and review the presentations at the recent SDDC Symposium, hosted by Datera.

Elapsed Time: 00:42:28

Timeline

* 00:00:00 – Intros* 00:02:00 – What is the Datera Data Services Platform?* 00:03:20 – Why package hardware with software for SDS?* 00:04:10 – DSP is a disaggregated solution, not HCI* 00:05:45 – How has the need for persistent storage changed?* 00:08:10 – Telemetry and dashboards are key SDS features* 00:10:15 – There’s no greater problem than when storage fails* 00:12:35 – Deeper dive – what features does DSP offer?* 00:13:45 – What is does real API-driven mean?* 00:15:00 – Performance and resiliency is driven by policy* 00:19:55 – DSP is multi-tenant with role-based access* 00:22:35 – Developers see a resource to use, admins to manage* 00:24:50 – Data will outlive any specific cluster configuration* 00:26:15 – How does DSP implement data protection?* 00:29:10 – Customers struggle to know their own requirements * 00:33:35 – How is DSP licensed?* 00:34:25 – How will DSP develop in the future?  New features?* 00:38:50 – how has the momentum of the company increased?* 00:41:05 – Wrap Up

Related Podcasts & Blogs

* #126 – Software-Defined Storage in Review with Hal Woods* Storage Field Day 10 Preview: Datera

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #58ai.

This episode was recorded live at Pure Storage offices in Silicon Valley. Chris talks to Pete Kirkpatrick, Chief Hardware Architect at Pure Storage about the FlashArray//C platform. FlashArray//C is a new solution announced at Pure Accelerate 2019 that offers a lower-cost all-flash solution for less latency-sensitive workloads.

FlashArray//C uses QLC technology, a cheaper, but slower flash alternative to SLC, MLC or even TLC that is being used in all-flash arrays today. Developing a solution using QLC requires more flash management, however using DirectFlash compared to standard SSDs offers Pure Storage cost and management advantages.

How will customers adopt the new technology? Introducing a new storage tier that sits below existing all-flash could be seen as an opportunity for customers to move away from FlashArray//X. Pete doesn’t see this as a threat, because the wider benefits for all-flash solutions are creating a new market.

Elapsed Time: 00:37:48

Timeline

* 00:00:00 – Intros* 00:01:05 – What are the differences/similarities of FlashArray//C & //X?* 00:03:30 – Pure Storage developed flash modules rather than using SSDs* 00:05:20 – QLC and TLC will co-exist in the enterprise* 00:05:50 – Why is QLC much harder to deal with?* 00:07:20 – 3x better endurance with DirectFlash, 10x compared to native SSD* 00:10:55 – Endurance saves cost – but also impacts and reduces risk* 00:12:00 – Latency & cost are the differentiators for //X and //C* 00:14:00 – Tiering solutions were always reactive and not proactive * 00:17:00 – How are FlashArray//C & //X differentiated in price?* 00:19:00 – Bigger drives means more metadata – means more DRAM * 00:19:55 – How does TCO affect pricing decisions?* 00:22:00 – How else will //C & //X be differentiated?* 00:23:40 – Will customers not simply move to //C from //X?* 00:25:45 – Running two arrays allows for real data protection* 00:26:30 – Is there a reason for holding back QLC to 2020?* 00:28:00 – How does Evergreen affect pricing with both //X & //C? * 00:29:25 – Could DirectMemory be used within //C?* 00:32:30 – Will there ever be an SCM-based FlashArray?* 00:33:35 – Is there any practical use to PLC flash?* 00:36:45 – Wrap Up

Related Podcasts & Blogs

* #118 – Pure Accelerate 2019 with Patrick Smith* Pure Storage FlashArray//C – First Impressions* Pure Storage Cloud Block Store Goes GA

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #h6s7.

This week Chris and Martin talk to Eran Brown, EMEA CTO at Infinidat about the challenges of ransomware and expediting fast recovery. Ransomware is a big challenge for IT organisations as hackers attempt to extort money from a range of businesses that include everything from large enterprises to healthcare organisations. However, is the traditional route of “restore from backup” actually practical or do IT organisations need a more comprehensive set of data protection and recovery techniques?

Eran sets the scene by describing the impacts of ransomware attacks on businesses. Companies with customer-facing websites risk lost revenue and customers who will go elsewhere if their services are down. It’s interesting to note that neither Chris nor Martin have seen a full “restore from backup” scenario in an IT organisation. In many respects, ransomware is a subset of a range of data loss scenarios. The aim for any organisation is to bring systems back into operation as fast as possible, and that means using a range of techniques.

Infinidat provides a set of data protection capabilities that includes snapshots, data replication and InfiniGuard, a purpose-built data protection platform. InfiniGuard uses the temporal location capabilities of InfiniBox to ensure that data recovery from disk is achieved at high speed, the time when recovery is most important.

Eran explains how the Infinidat portfolio can be used to reduce recovery times and bring data back within service level objectives. This includes proactive monitoring, which in the future will be key in ensuring IT organisations identify and quickly work to resolve ransomware attacks.

You can find more about InfiniGuard at the Infinidat website, including the white paper Eran references in the discussion.

Elapsed Time: 00:28:03

Timeline

* 00:00:00 – Intros * 00:01:25 – Ransomware 101 * 00:03:05 – Backup systems can be a weak link * 00:03:45 – What is the financial impact of ransomware? * 00:07:00 – Ransomware is a subset of other data loss scenarios * 00:08:30 – Snapshots are a great initial recovery solution * 00:10:35 – Recovery uses a mix of solutions to achieve * 00:11:30 – InfiniGuard – a recovery-focused data protection solution * 00:14:00 – We should be focusing on the performance of recovery not backup * 00:15:00 – InfiniGuard uses InfiniBox and temporal location to accelerate restores * 00:18:00 – InfiniGuard pre-fetches data into cache to speed up performance * 00:20:00 – Customers can use primary and secondary platform snapshots * 00:21:30 – Increased data volumes in snapshots can indicate ransomware * 00:27:05 – Wrap Up

Transcript

Related Podcasts & Blogs

* #86 – Eran Brown Discusses Storage, Security & Multi-Cloud* Managing Legacy Backups

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #3e6a.

In this podcast episode, we continue the discussions from Hitachi NEXT 2019. Chris talks to Scott Baker, Senior Director, Emerging Business and Data Intelligence at Hitachi Vantara. Scott covers the HCP object storage platform and has responsibility for data protection solutions. This conversation looks at updates to the Hitachi HCP portfolio and associated tools and products.

The discussion starts with a look at Data Ops. This is a relatively new term in the industry that we discussed on a podcast earlier this year with CMO Jonathan Martin and John Magee (episode #103). As businesses mature in their data management lifecycle, the curation and ongoing publishing of data sources becomes a critical requirement, the Data Ops concept.

The conversation continues the discussion to the additional products around the Lumada brand (specifically Lumada Data Lake) and the move from data from being purely an archive to an intelligent business asset.

To understand the reference to military helicopters, listen to Scott’s podcast (episode #68) from last year, recorded outside in San Diego at Next 2018. It was a challenging environment to say the least…

Elapsed Time: 00:24:38

Timeline

* 00:00:00 – Intros* 00:02:15 – What is Data Ops?* 00:05:30 – Data management workflows are being pushed to the Edge* 00:08:03 – Workflow policies can be managed in core, pushed to the edge* 00:09:17 – How do we protect against “malware” data injection?* 00:11:43 – User can rate data to validate the quality, model and lineage* 00:13:20 – New – improvements to content intelligence product – ML/AI* 00:16:20 – Data ingest of structured data – conversion to XML* 00:18:00 – ML/AI to build visual dictionaries for non text data* 00:20:00 – Data becomes more “intelligent” and not just an archive* 00:21:50 – Instance Director doesn’t change the format of data* 00:23:23 – Wrap up

Related Podcasts & Blogs

* #103 – Data Management and Data Ops with Hitachi Vantara* #127 – Hitachi Next 2019: CI & HCI Solutions Update* #125 – Hitachi Next 2019: Ops Center* #123 – Hitachi Next 2019: VSP 5000* Hitachi Vantara Introduces VSP 5000 Series Storage Arrays

Copyright (c) 2016-2019 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #75ID.

Disclaimer: Chris was invited to NEXT 2019, with Hitachi covering travel and accommodation costs. There is no obligation to blog or produce podcasts from the event.