Intel – It’s About Getting The Right Kind Of Fast At The Edge

Disclaimer: I recently attended Storage Field Day 22.  Some expenses were paid for by Tech Field Day. 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.

Intel recently presented at Storage Field Day 22. You can see videos of the presentation here, and download my rough notes from here.

 

The Problem

A lot of countries have used lockdowns as a way to combat the community transmission of COVID-19. Apparently, this has led to an uptick in the consumption of streaming media services. If you’re somewhat familiar with streaming media services, you’ll understand that your favourite episode of Hogan’s Heroes isn’t being delivered from a giant storage device sitting in the bowels of your streaming media provider’s data centre. Instead, it’s invariably being delivered to your device from a content delivery network (CDN) device.

 

Content Delivery What?

CDNs are not a new concept. The idea is that you have a bunch of web servers geographically distributed delivering content to users who are also geographically distributed. Think of it as a way to cache things closer to your end users. There are many reasons why this can be a good idea. Your content will load faster for users if it resides on servers in roughly the same area as them. Your bandwidth costs are generally a bit cheaper, as you’re not transmitting as much data from your core all the way out to the end user. Instead, those end users are getting the content from something close to them. You can potentially also deliver more versions of content (in terms of resolution) easily. It can also be beneficial in terms of resiliency and availability – an outage on one part of your network, say in Palo Alto, doesn’t need to necessarily impact end users living in Sydney. Cloudflare does a fair bit with CDNs, and there’s a great overview of the technology here.

 

Isn’t All Content Delivery The Same?

Not really. As Intel covered in its Storage Field Day presentation, there are some differences with the performance requirements of video on demand and live-linear streaming CDN solutions.

Live-Linear Edge Cache

Live-linear video streaming is similar to the broadcast model used in television. It’s basically programming content streamed 24/7, rather than stuff that the user has to search for. Several minutes of content are typically cached to accommodate out-of-sync users and pause / rewind activities. You can read a good explanation of live-linear streaming here.

[image courtesy of Intel]

In the example above, Intel Optane PMem was used to address the needs of live-linear streaming.

  • Live-linear workloads consume a lot of memory capacity to maintain a short-lived video buffer.
  • Intel Optane PMem is less expensive than DRAM.
  • Intel Optane PMem has extremely high endurance, to handle frequent overwrite.
  • Flexible deployment options – Memory Mode or App-Direct, consuming zero drive slots.

With this solution they were able to achieve better channel and stream density per server than with DRAM-based solutions.

Video on Demand (VoD)

VoD providers typically offer a large library of content allowing users to view it at any time (e.g. Netflix and Disney+). VoD servers are a little different to live-linear streaming CDNs. They:

  • Typically require large capacity and drive fanout for performance / failure domains; and
  • Have a read-intensive workload, with typically large IOs.

[image courtesy of Intel]

 

Thoughts and Further Reading

I first encountered the magic of CDNs years ago when working in a data centre that hosted some Akamai infrastructure. Windows Server updates were super zippy, and it actually saved me from having to spend a lot of time standing in the cold aisle. Fast forward about 15 years, and CDNs are being used for all kinds of content delivery on the web. With whatever the heck this is is in terms of the new normal, folks are putting more and more strain on those CDNs by streaming high-quality, high-bandwidth TV and movie titles into their homes (except in backwards places like Australia). As a result, content providers are constantly searching for ways to tweak the throughput of these CDNs to serve more and more customers, and deliver more bandwidth to those users.

I’ve barely skimmed the surface of how CDNs help providers deliver content more effectively to end users. What I did find interesting about this presentation was that it reinforced the idea that different workloads require different infrastructure solutions to deliver the right outcomes. It sounds simple when I say it like this, but I guess I’ve thought about streaming video CDNs as being roughly the same all over the place. Clearly they aren’t, and it’s not just a matter of jamming some SSDs in one RU servers and hoping that your content will be delivered faster to punters. It’s important to understand that Intel Optane PMem and Intel Optane 3D NAND can give you different results depending on what you’re trying to do, with PMem arguably giving you better value for money (per GB) than DRAM. There are some great papers on this topic available on the Intel website. You can read more here and here.

Random Short Take #60

Welcome to Random Short take #60.

  • VMware Cloud Director 10.3 went GA recently, and this post will point you in the right direction when it comes to planning the upgrade process.
  • Speaking of VMware products hitting GA, VMware Cloud Foundation 4.3 became available about a week ago. You can read more about that here.
  • My friend Tony knows a bit about NSX-T, and certificates, so when he bumped into an issue with NSX-T and certificates in his lab, it was no big deal to come up with the fix.
  • Here’s everything you wanted to know about creating an external bootable disk for use with macOS 11 and 12 but were too afraid to ask.
  • I haven’t talked to the good folks at StarWind in a while (I miss you Max!), but this article on the new All-NVMe StarWind Backup Appliance by Paolo made for some interesting reading.
  • I loved this article from Chin-Fah on storage fear, uncertainty, and doubt (FUD). I’ve seen a fair bit of it slung about having been a customer and partner of some big storage vendors over the years.
  • This whitepaper from Preston on some of the challenges with data protection and long-term retention is brilliant and well worth the read.
  • Finally, I don’t know how I came across this article on hacking Playstation 2 machines, but here you go. Worth a read if only for the labels on some of the discs.

Fujifilm Object Archive – Not Your Father’s Tape Library

Disclaimer: I recently attended Storage Field Day 22.  Some expenses were paid for by Tech Field Day. 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.

Fujifilm recently presented at Storage Field Day 22. You can see videos of the presentation here, and download my rough notes from here.

 

Fujifilm Overview

You’ve heard of Fujifilm before, right? They do a whole bunch of interesting stuff – batteries, cameras, copiers. Nami Matsumoto, Director of DMS Marketing and Operations, took us through some of Fujifilm’s portfolio. Fujifilm’s slogan is “Value From Innovation”, and it certainly seems to be looking to extract maximum value from its $1.4B annual spend on research and development. The Recording Media Products Division is focussed on helping “companies future proof their data”.

[image courtesy of Fujifilm]

 

The Problem

The challenge, as always (it seems), is that data growth continues apace while budgets remain flat. As a result, both security and scalability are frequently sacrificed when solutions are deployed in enterprises.

  • Rapid data creation: “More than 59 Zettabytes (ZB) of data will be created, captured, copied, and consumed in the world this year” (IDC 2020)
  • Shift from File to Object Storage
  • Archive Market – 60 – 80%
  • Flat IT budgets
  • Cybersecurity concerns
  • Scalability

 

Enter The Archive

FUJIFILM Object Archive

Chris Kehoe, Director of DMS Sales and Engineering, spent time explaining what exactly FUJIFILM Object Archive was. “Object Archive is an S3 based archival tier designed to reduce cost, increase scale and provide the highest level of security for long-term data retention”. In short, it:

  • Works like Amazon S3 Glacier in your DC
  • Simply integrates with other object storage
  • Scales on tape technology
  • Secure with air gap and full chain of custody
  • Predictable costs and TCO with no API or egress fees

Workloads?

It’s optimised to handle the long-term retention of data, which is useful if you’re doing any of these things:

  • Digital preservation
  • Scientific research
  • Multi-tenant managed services
  • Storage optimisation
  • Active archiving

What Does It Look Like?

There are a few components that go into the solution, including a:

  • Storage Server
  • Smart cache
  • Tape Server

[image courtesy of Fujifilm]

Tape?

That’s right, tape. The tape library supports LTO7, LTO8, TS1160. The data is written using “OTFormat” specification (you can read about that here). The idea is that it packs a bunch of objects together so they get written efficiently.  

[image courtesy of Fujifilm]

Object Storage Too

It uses an “S3-compatible” API – the S3 server is built on Zenko inside (Scality). From an object storage perspective, it works with Cloudian HyperStore, Caringo Swarm, NetApp StorageGRID, Scality Ring. It also has Starfish and Tiger Bridge support.

Other Notes

The product starts at 1PB of licensing. You can read the Solution Brief here. There’s an informative White Paper here. And there’s one of those nice Infographic things here.

Deployment Example

So what does this look like from a deployment perspective? One example was a typical primary storage deployment, with data archived to an on-premises object storage platform (in this case NetApp StorageGRID). When your archive got really “cold”, it would be moved to the Object Archive.

[image courtesy of Fujifilm]

[image courtesy of Fujifilm]

 

Thoughts

Years ago, when a certain deduplication storage appliance company was acquired by a big storage slinger, stickers with “Tape is dead, get over it” were given out to customers. I think I still have one or two in my office somewhere. And I think the sentiment is spot on, at least in terms of the standard tape library deployments I used to see in small to mid to large enterprise. The problem that tape was solving for those organisations at the time has largely been dealt with by various disk-based storage solutions. There are nonetheless plenty of use cases where tape is still considered useful. I’m not going to go into every single reason, but the cost per GB of tape, at a particular scale, is hard to beat. And when you want to safely store files for a long period of time, even offline? Tape, again, is hard to beat. This podcast from Curtis got me thinking about the demise of tape, and I think this presentation from Fujifilm reinforced the thinking that it was far from on life support – at least in very specific circumstances.

Data keeps growing, and we need to keep it somewhere, apparently. We also need to think about keeping it in a way that means we’re not continuing to negatively impact the environment. It doesn’t necessarily make sense to keep really old data permanently online, despite the fact that it has some appeal in terms of instant access to everything ever. Tape is pretty good when it comes to relatively low energy consumption, particularly given the fact that we can’t yet afford to put all this data on All-Flash storage. And you can keep it available in systems that can be relied upon to get the data back, just not straight away. As I said previously, this doesn’t necessarily make sense for the home punter, or even for the small to midsize enterprise (although I’m tempted now to resurrect some of my older tape drives and see what I can store on them). It really works better at large scale (dare I say hyperscale?). Given that we seem determined to store a whole bunch of data with the hyperscalers, and for a ridiculously long time, it makes sense that solutions like this will continue to exist, and evolve. Sure, Fujifilm has sold something like 170 million tapes worldwide. But this isn’t simply a tape library solution. This is a wee bit smarter than that. I’m keen to see how this goes over the next few years.

Storage Field Day 22 – I’ll Be At Storage Field Day 22

Here’s some news that will get you excited. I’ll be virtually heading to the US this week for another Storage Field Day event. If you haven’t heard of the very excellent Tech Field Day events, you should check them out. It’s also worth visiting the Storage Field Day 22 website during the event (August 4-6) as there’ll be video streaming and updated links to additional content. You can also see the list of delegates and event-related articles that have been published.

I think it’s a great line-up of both delegates and presenting companies this time around.

 

I’d like to publicly thank in advance the nice folks from Tech Field Day who’ve seen fit to have me back, as well as my employer for letting me take time off to attend these events. Also big thanks to the companies presenting. It’s going to be a lot of fun. Last time was a little weird doing this virtually, rather than in person, but I think it still worked. As things open back up in the US you’ll start to see a blend of in-person and virtual attendance for these events. I know that Komprise will be filming its segment from the Doubletree. Hopefully we’ll get things squared away and I’ll be allowed to leave the country next year. I’m really looking forward to this, even if it means doing the night shift for a few days. Presentation times are below, and all times are US/Pacific.

Wednesday, Aug 4 8:00-9:30 Infrascale Presents at Storage Field Day 22
Wednesday, Aug 4 11:00-13:30 Intel Presents at Storage Field Day 22
Presenters: Allison GoodmanElsa AsadianKelsey PrantisKristie MannNash KleppanSagi Grimberg
Thursday, Aug 5 8:00-10:00 CTERA Presents at Storage Field Day 22
Presenters: Aron BrandJim CrookLiran Eshel
Thursday, Aug 5 11:00-13:00 Komprise Presents at Storage Field Day 22
Presenters: Krishna SubramanianMike PeercyMohit Dhawan
Friday, Aug 6 8:00-9:00 Fujifilm Presents at Storage Field Day 22
Friday, Aug 6 10:00-11:30 Pure Storage Presents at Storage Field Day 22
Presenters: Ralph RonzioStan Yanitskiy

Random Short Take #58

Welcome to Random Short take #58.

  • One of the many reasons I like Chin-Fah is that he isn’t afraid to voice his opinion on various things. This article on what enterprise storage is (and isn’t) made for some insightful reading.
  • VMware Cloud Director 10.3 is now GA – you can read more about it here.
  • Feeling good about yourself? That’ll be quite enough of that thanks. This article from Tom on Value Added Resellers (VARs) and technical debt goes in a direction you might not expect. (Spoiler: staff are the technical debt). I don’t miss that part of the industry at all.
  • Speaking of work, this article from Preston on being busy was spot on. I’ve worked in many places in my time where it’s simply alarming how much effort gets expended in not achieving anything. It’s funny how people deal with it in different ways too.
  • I’m not done with articles by Preston though. This one on configuring a NetWorker AFTD target with S3 was enlightening. It’s been a long time since I worked with NetWorker, but this definitely wasn’t an option back then.  Most importantly, as Preston points out, “we backup to recover”, and he does a great job of demonstrating the process end to end.
  • I don’t think I talk about data protection nearly enough on this weblog, so here’s another article from a home user’s perspective on backing up data with macOS.
  • Do you have a few Rubrik environments lying around that you need to report on? Frederic has you covered.
  • Finally, the good folks at Backblaze are changing the way they do storage pods. You can read more about that here.

*Bonus Round*

I think this is the 1000th post I’ve published here. Thanks to everyone who continues to read it. I’ll be having a morning tea soon.

StorONE and Seagate Team Up

This news came out a little while ago, but I thought I’d cover it here nonetheless. Seagate and StorONE recently announced that the Seagate Exos AP 5U84 Application Platform would support StorONE’s S1:Enterprise Storage Platform.

 

It’s A Box!

[image courtesy of StorONE]

The Exos 5U84 Dual Node supports:

  • 2x 1.8 GHz CPU (E5-2648L v4)
  • 2x 256GB RAM
  • Storage capacities between 250TB and 1.3PB

 

It’s Software!

Hardware is fun, but it’s the software that really helps here, with support for:

  • Full High Availability
  • Automated Tiering
  • No Write Cache
  • Rapid RAID Rebuilds
  • Unlimited Snapshots
  • Cascading Replication
  • Self Encrypting Drives

It offers support for multiple access protocols, including iSCSI, NFS, SMB, and S3. Note that there is no FC support with this unit.

 

Thoughts and Further Reading

I’ve had positive things to say about StorONE in the past, particularly when it comes to transparent pricing and the ability to run this storage solution on commodity hardware. I’ve been on the fence about whether hybrid storage solutions are really on the way out. It felt like they were, for a while, and then folks kept coming up with tweaks to software that meant you could get even more bang for your buck (per GB). Much like tape, I think it would be premature to say that hybrid storage using spinning disk is dead just yet.

Obviously, the folks at StorONE have skin in this particular game, so they’re going to talk about how hybrid isn’t going anywhere. It’s much the same as Michael Dell telling me that the on-premises server market is hotting up. When a vendor is selling something, it’s in their interest to convince you that a market exists for that thing and it is hot. That said, some of the numbers Crump and the team at StorONE have shown me are indeed compelling. When you couple those numbers with the cost of the solution (you can work out for yourself here) it becomes difficult to dismiss out of hand.

When I look at storage solutions I like to look at the numbers, and the hardware, and how it’s supported. But what’s really important is whether the solution is up to the task of the workload I need to throw at it. I also want to know that someone can fix my problem when the magic smoke escapes said storage solution. After a while in the industry, you start to realise that, regardless of what the brochures look like, there are a few different ways that these kind of things get put together. Invariably, unless the solution is known for being reckless with data integrity, or super slow, there’s going to be a point at which the technical advantages become less of a point of differentiation. It’s at that point where the economics really come into play.

The world is software-defined in a lot of ways, but this doesn’t mean you can run your favourite storage code on any old box and expect a great outcome. It does, however, mean that you no longer have to pay a premium to get good performance, good capacity, and a reliable outcome for your workload. You also get the opportunity to enjoy performance improvements as the code improves, without necessarily needing to update your hardware. Which is kind of neat, particularly if you’ve ever paid a pretty penny for golden screwdriver upgrades from big brand disk slingers in the past. This solution might not be for everyone, particularly if you already have a big arrangement with some of the bigger vendors. But if you’re looking to do something, and can’t stretch the economics to an All-Flash solution, this is worth a look.

Storage Field Day 21 – Wrap-up and Link-o-rama

Disclaimer: I recently attended Storage Field Day 21.  My flights, accommodation and other expenses were paid for by Tech Field Day. 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.

This is a quick post to say thanks once again to Stephen and Ben, and the presenters at Storage Field Day 21. I had a great time. For easy reference, here’s a list of the posts I did covering the events (they may not match the order of the presentations).

Storage Field Day 21 – I’ll Be At Storage Field Day 21

Storage Field Day 21 – (Fairly) Full Disclosure

Back To The Future With Tintri

Hammerspace, Storageless Data, And One Tough Problem

Intel Optane – Challenges and Triumphs

NetApp Keystone – How Do you Want It?

Pliops – Can We Take Fast And Make It Faster?

Nasuni Puts Your Data Where You Need It

MinIO – Cloud, Edge, Everywhere …

Also, here’s a number of links to posts by my fellow delegates (in no particular order). They’re all very smart people, and you should check out their stuff, particularly if you haven’t before. I’ll attempt to keep this updated as more posts are published. But if it gets stale, the Storage Field Day 21 landing page will have updated links.

 

Jason Collier (@BocaNuts)

 

Barry Coombs (@VirtualisedReal)

#SFD21 – Storage Field Day 21 – Tintri

#SFD21 – Storage Field Day 21 – NetApp

#SFD21 – Storage Field Day 21 – Nasuni

#SFD21 – Storage Field Day 21 – MinIO Session

#SFD21 – Storage Field Day 21 – Pliops

#SFD21 – Storage Field Day 21 – Hammerspace

#SFD21 – Storage Field Day 21 – Intel

 

Becky Elliott (@BeckyLElliott)

 

Matthew Leib (@MBLeib)

 

Ray Lucchesi (@RayLucchesi)

The rise of MinIO object storage

Data Science storage with NetApp’s Python Toolkit

Storageless data!?

115-GreyBeards talk database acceleration with Moshe Twitto, CTO&Co-founder, Pliops

 

Andrea Mauro (@Andrea_Mauro)

 

Max Mortillaro (@DarkkAvenger)

Nasuni – Cloud-Scale NAS Without Cloud Worries

Storage Field Day 21 – The TECHunplugged Take on Nasuni

Pliops: Re-Imagining Storage, Crushing Bottlenecks and a Bright Future in the Cloud

 

Keiran Shelden (@Keiran_Shelden)

 

Enrico Signoretti (@esignoretti)

Object Storage Is Heating Up

Storage Options for the Distributed Enterprise

 

Paul Stringfellow (@TechStringy)

Looking ahead with Storage Field Day 21 – Barry Coombs, Jason Collier, Max Mortillaro – Ep 149

Storageless data, really? – Doug Fallstrom – Ep156

 

Frederic Van Haren (@FredericVHaren)

 

On-Premise IT Podcast

Is Storageless Storage Just Someone Else’s Storage?

 

Now please enjoy this group photo.

[image courtesy of Gestalt IT]

MinIO – Cloud, Edge, Everywhere …

Disclaimer: I recently attended Storage Field Day 21.  My flights, accommodation and other expenses were paid for by Tech Field Day. 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.

MinIO recently presented at Storage Field Day 21. You can see videos of the presentation here, and download my rough notes from here.

 

What Is It?

To quote the good folks at MinIO, it is a “high performance, Kubernetes-native object store”. It is designed to be used for large-scale data infrastructure, and was built from scratch to be cloud native.

[image courtesy of MinIO]

Design Principles

MinIO has been built with the following principles in mind:

  • Cloud Native – born in the cloud with “cloud native DNA”
  • Performance Focussed – believe it is the fastest object store in existence
  • Simplicity – designed for simplicity because “simplicity scales”

S3 Compatibility

MinIO is heavily focussed on S3 compatibility. It was first to market with V4 and one of the few vendors to support S3 Select. It has also been strictly consistent from inception.

Put Me In Your Favourite Box

The cloud native part of MinIO was no accident, and as a result more than 62% of MinIO instances run in containers (according to MinIO). 43% of those instances are also managed via Kubernetes. It’s not just about jamming this solution into your favourite container solution though. The lightweight nature of it means you can deploy it pretty much anywhere. As the MinIO folks pointed out during the presentation, MinIO is going everywhere that AWS S3 isn’t.

 

Thoughts And Further Reading

I love object storage. Maybe not in the way I love my family or listening to records or beer, but I do love it. It’s not just useful for storage for the great unwashed of the Internet, but also backup and recovery, disaster recovery, data archives, and analytics. And I’m a big fan of MinIO, primarily because of the S3 compatibility and simplicity of deployment. Like it or not, S3 is the way forward in terms of a standard for object storage for cloud native (and a large number of enterprise) workloads. I’ve written before about other vendors being focussed on this compatibility, and I think it’s great that MinIO has approached this challenge with just as much vigour. There are plenty of problems to be had deploying applications at the best of times, and being able to rely on the storage vendor sticking to the script in terms of S3 compatibility takes one more potential headache away.

The simplicity of deployment is a big part of what intrigues me about MinIO too. I’m old enough to remember some deployments of early generation on-premises object storage systems that involved a bunch of hardware and complicated software interactions for what ultimately wasn’t a great experience. Something like MinIO can be up and running on some pretty tiny footprints in no time at all. A colleague of mine shared some insights into that process here.

And that’s what makes this cool. It’s not that MinIO are trying to take a piece of the AWS pie. Rather, it’s positioning the solution as one that can operate everywhere that the hyperscalers aren’t. Putting object storage solutions in edge locations has historically been a real pain to do. That’s no longer the case. Part of this has to do with the fact that we’ve got access to really small computers and compact storage. But it also has a bit to do with lightweight code that can be up and running in a snap. Like some of the other on-premises object vendors, MinIO has done a great job of turning people on to the possibility of doing cool storage for cloud native workloads outside of the cloud. It seems a bit odd until you think about all of the use cases in enterprise that might work really well in cloud, but aren’t allowed to be hosted in the cloud. It’s my opinion that MinIO has done a great job of filling that gap (and exceeding expectations) when it comes to lightweight, easy to deploy object storage. I’m looking forward to see what’s next for them, particularly as the other vendors start to leverage the solution. For another perspective on MinIO’s growth, check out Ray’s article here.

Nasuni Puts Your Data Where You Need It

Disclaimer: I recently attended Storage Field Day 21.  My flights, accommodation and other expenses were paid for by Tech Field Day. 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.

Nasuni recently presented at Storage Field Day 21. You can see videos of the presentation here, and download my rough notes from here.

 

Nasuni?

The functionality is in the product name. It’s NAS that offers a unified file system across cloud. The key feature is that it’s cloud-native, rather than built on any particular infrastructure solution.

[image courtesy of Nasuni]

The platform is comprised of 5 key components.

UniFS

  • Consolidates files and metadata in cloud storage – “Gold Copy”
  • Ensures durability by storing files as immutable, read-only objects
  • Stores an unlimited version history of every file

Virtual Edge Appliances

  • Caches active files with 99% hit rate
  • 98% smaller footprint vs traditional file server / NAS
  • Scales across all sites, including VDI
  • Supports standard file sharing protocols
  • Built-in web server enables remote file access via web browser (HTTP)

Management Console

  • Administers appliances, volumes, shares and file recovery
  • Automated through central GUI and REST API
  • Provides centralised monitoring, reporting, and alerting

Orchestration Center

  • Multi-site file sync keeps track of versions
  • Advanced version control with Nasuni Global File Lock
  • Multi-region cloud support to ensure performance

Analytics Connector

  • Translates file data into native object storage format
  • Leverage any public cloud services (AI, data analytics, search)
  • Multi-cloud support so you can run any cloud service against your data

 

Thoughts and Further Reading

I’m the first to admit I’ve had a bit of a blind spot for Nasuni for a little while now. Not because I think the company doesn’t do cool stuff – it really does. Rather, my former employer was an investor in the tech and was keen to see how we could use the platform in every opportunity. Even when the opportunity wasn’t appropriate.

Distributed storage for file sharing has been a pain in the rear for enterprises ever since enterprises have been a thing. The real challenge has been doing something sensible about managing data across multiple locations in a cogent fashion. As local becomes global, this becomes even more of an issue, particularly when folks all across the world need to work on the same data. Email isn’t really great for this, and some of those sync and share solutions don’t cope well with the scale that is sometimes required. In the end, file serving is still a solution that can solve a problem for a lot of enterprise use cases.

The advent of public cloud has been great in terms of demonstrating that workloads can be distributed, and you don’t need to have a bunch of tin sitting in the office to get value from infrastructure. Nasuni recognised this over ten years ago, and it has put together a platform that seeks to solve that problem by taking advantage of the distributed nature of cloud, whilst acknowledging that virtualised resources can make for a useful local presence when it comes to having the right data in the right place. One of my favourite things about the solution is that you can also do stuff via the Analytics Connector to derive further value from your unstructured data. This is not a unique feature, but it’s certainly something that gives the impression that Nasuni isn’t just here to serve up your data.

The elegance of the Nasuni solution is in the fact that the complexity is well hidden from the end user. It’s a normal file access experience, but it’s hosted in the cloud. When you contrast that with what you get from the sync solutions of the world or the clumsy web-based document management systems so prevalent in the enterprise, this kind of simplicity is invaluable. It’s my opinion that there is very much a place for this kind of solution in the marketplace. The world is becoming increasingly global, but we still need solutions that can provide data where we need it. We also need those solutions to accommodate the performance and resilience needs of the enterprise.

If you’re after a great discussion on storage options for the distributed enterprise, check out Enrico’s article over at GigaOm.

Pliops – Can We Take Fast And Make It Faster?

Disclaimer: I recently attended Storage Field Day 21.  My flights, accommodation and other expenses were paid for by Tech Field Day. 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.

Pliops recently presented at Storage Field Day 21. You can see videos of the presentation here, and download my rough notes from here.

 

The Problem

You might have heard of solid-state drives (SSDs). You might have one in your computer. You might even have a data centre full of them. They’re quiet and they’re fast. It’s generally accepted that SSDs perform way better than HDDs. The problem, however, is that CPUs haven’t kept up with that performance increase. The folks at Pliops have also pointed out that resiliency technologies such as RAID generally suck, particularly when you’re using SSDs. In essence, you’re wasting a good chunk of your Flash.

 

The Solution?

The solution, according to Pliops, is the Pliops Storage Processor. This is “a hardware-based storage accelerator that enables cloud and enterprise customers to offload and accelerate data-intensive workloads using just a fraction of the computational load and power”. Capabilities include increased performance, capacity expansion, improved endurance, and data protection capabilities.

System Integration

From an integration perspective, the card is a half-height, half-length PCIe device that fits in any standard rackmount server. There’s a Pliops Agent that’s installed, and it supports a variety of different deployment options. There’s even a cloud / as-a-service option available.

[image courtesy of Pliops]

Use Cases

The Pliops SP is targeted primarily at RDBMS, NoSQL and Analytics workloads, due in large part to its ability to reduce read and write amplification on SSDs – something that’s been a problem since Flash became more prevalent in the data centre.

[image courtesy of Pliops]

 

Thoughts and Further Reading

On the surface, the Pliops Storage Processor seems to be solving a fairly specific problem. It’s not a problem that gets a lot of airplay, but that doesn’t mean it’s not an important one to solve. There are scads of solutions in the market that have been developed to address the problem of legacy systems design. For example, the way we addressed resilience previously (i.e. RAID) doesn’t work that well as drive capacities continue to increase. We’ve also fundamentally changed the media we’re working with, but haven’t necessarily developed new ways of taking advantage of that media.

Whenever I see add-in devices like this I worry that it would be a pain to manage at any sort of scale. But then I remember that literally everything related to hardware is a pain to manage at any kind of scale. The Pliops folks tell us that it’s not actually too bad, and any disadvantages related to having super specialised add-in cards deployed in servers is more than made up for by the improved outcomes achieved with those cards.

Ultimately, the value of a solution like the Pliops Storage Processor is absolutely tied to whether you’ve had a problem with this in the past. If you have, you’ll understand that this kind of solution is a reasonably elegant way of addressing the problem. It has the added bonus of taking fast media and eking out even more performance from that media.

Pliops has only been around since 2017, but it recently announced a decent funding round and the product is being ready for mass deployment. I’ll happily admit that I’ve done a fairly poor job of explaining the Pliops Storage Processor and what it does, so I recommend you check out the solution brief on the Pliops website. If you’d like another perspective, be sure to head over to TECHunplugged to read Max’s thoughts on Pliops.