SVG Sit-Down: CEO Jamie Lerner on Quantum’s Future and Leveraging M&E Prowess for Other Industries
It has been nearly six months since Jamie Lerner came aboard as president/CEO of Quantum, and the veteran tech exec is well on his way in plotting a new course for the video- and data-storage outfit, which has struggled recently. After cycling through five CEOs in less than a year, Quantum is looking for Lerner, along with recently named CFO Michael Dodson, to establish stability at the top of the ranks.
Lerner, who previously led a turnaround of three business units in Cisco’s Cloud and Systems Technology Group and also served in top executive roles at Seagate and Pivot3, aims to expand to new market verticals by leveraging the leading-edge technology that Quantum has developed for media and entertainment (M&E) and exporting it to a variety of other industries. As the cloud and virtualization become more than just buzzwords, he seeks to embrace next-generation workflows and evolve Quantum’s product portfolio to serve media companies’ rapidly evolving needs.
SVG sat down with Lerner for an in-depth discussion to address his roadmap for the company, including expanding to markets beyond M&E, what to expect in Quantum’s next generation of products, the rise of AI and machine learning, how the company is adapting to the demand for virtualized infrastructure, the power of NVMe for storage, the enduring viability of LTO tape, and, perhaps most interestingly, his proposal for a subscription-based, managed service for storage hardware.
What have you learned in your first few months as CEO at Quantum?
I’ve spent the past few months learning what our customers do with our technology. What I’ve come away with is that Quantum is woven so deeply into the fabric of our society that you interact with it almost daily. It’s pretty hard to watch television or online content without seeing something produced on our gear: whether it’s films, [episodic] shows, news, sports, or anything else.
The work that we do in media and entertainment, especially in sports, is informing everything that we’re going to do in this company across every [vertical].
Our analysis shows that, by 2025, 80% of the world’s data will be video. About 50% of that is media and entertainment, and about 50% of that is going to be surveillance data. We believe we are the world’s best-positioned company to handle this deluge of new data. All we think about is rich media and video. Traditional IT storage technology is totally inappropriate for video. If you’re in the video world, we are going to build this from the ground just for you.
So, how does Quantum plan to use the video-production tools developed for M&E in other verticals?
We are starting to realize that, in terms of video needs, police departments look exactly like a television studio. A hospital looks like a television studio. And a manufacturing plant looks like a television studio. When a police officer comes off a beat, his department is going ingest his video from his car and his body camera, take the bits that are of interest and distribute them to lawyers, district attorneys, and courts. And they’re going to archive it for 30 or 40 years in case there’s an appeal. There’s no sound editing or colorizing, but that certainly sounds similar to our experience in media and entertainment.
But, you remain committed to the M&E market?
Yes, absolutely. That is our core use case that all these other use cases are being built from. Our strength in M&E will continue to propel us for the foreseeable future.
What enhancements do you plan for Quantum’s core products, such as StorNext?
StorNext lends itself to really high-speed–media workflows. The first thing we’re doing with the core file systems is make them even faster by moving to all-NVMe–based architectures. We’re making the absolute world’s fastest hardware, and we will be all NVMe end-to-end.
NVMe is hands-down the world’s fastest editing and postproduction platform, but it’s also really expensive. So creating a top-to-bottom NVMe environment is unviable economically. What makes it work is really intelligent tiering. Most data will be sitting on a disk tier, so you can do the bulk of your storage spend in disk and a very thin sliver of NVMe. The whole architecture comes down to how intelligently you put the right content on the NVMe tier at the right time.
The next thing we’re doing [with our portfolio] is hyper-converging: we want to bring the storage as close as we can to the CPU and GPUs, so we can analyze at extremely high speed. We’re moving all of our storage technology into virtual machines and into containers and hyper-converging.
We’re also starting to bring the analytic software down to the boxes to improve metadata analysis. A lot of the metadata tagging on video files is just not accurate. I see dozens of [loggers] just looking at a piece of video, confirming it is what we think it is, and typing that in. I think it’s obvious that it all should be machine learning. Catalogues are getting so big that human logging — typing what each pitch was in a baseball game, for example — is totally non-viable. If a picture is a thousand words, then a video is a million words, and there’s just no way to type fast enough to capture all that. So our core hardware infrastructure is being designed to bring machine learning to the forefront.
And finally, we’re completely revamping our management interfaces. StorNext is a de facto standard and is incredibly performant, but most people would say it’s very hard to manage. Today, the world’s most modern products have become easy to use. So we’re completely rethinking the management interface and making the product manageable from the web and the cloud vs. on-premises management tools. In StorNext version 6.2, we will have all-web-based, all-cloud-based management for the product.
Do you see any of these next-generation products impacting sports-customers specifically?
We’re making the form factor smaller. For years, StorNext has scaled up to huge [installations]. But now we are looking to the small-scale. I can’t haul a giant StorNext-based appliance out to a remote sporting event; I need it smaller and more ruggedized. We’re putting a lot of time to make the product easier to use and scaled down for a smaller sporting event. Why can’t we bring the technology used for the biggest events in the world and scale it down for just a small regional event produced for YouTube? We have to make it easier to use and be the size you can put into a Pelican case.
How do you see AI and machine learning playing a role in this industry?
We are already building our architectures from the ground up for AI. But the other thing we’re doing with machine learning is helping you decide where to put your content. One of the things that Quantum is known for is tiering. You do tiering to save on budget. NVMe: the world’s most expensive storage and the fastest. SSDs: really fast but a little cheaper. Hard drives: middle of the road. And tapes: slow but the most economical way to store data.
When you have millions of assets, you have to have an analytics [system] that shows that, if a file is being used a lot, it should probably be put on the super-fast tier. Or maybe the [system] noticed that, when everyone touches this file, they always touch these eight others, so make those available immediately. Or, if no one has touched it for a while, send it back down [to the deep archive]. We’re looking at analytics and machine learning for that. It saves a huge amount of money and helps you make decisions like when to put something on tape and when to put it in the cloud.
M&E organizations increasingly look to virtualized infrastructure and cloud-based workflows. How do you see this impacting Quantum’s business and strategy?
We’re excited about the cloud. While I don’t think high-end production is an end-to-end cloud item yet, pieces of the workflow are definitely viable in the cloud — like large render farms — and others still belong on-premises — like sound editing. Our customers are going to be moving pieces of the workflow to the cloud, and, if we don’t help them, we will lose them. So we’re starting to look into developing ready-to-go workflows in the cloud.
One of the first releases was FlexTier, which provides intelligent policy-based tiering into the cloud. We also have a tool called FlexSync: even if you’re editing files on premises, you can synchronize them to the cloud or [vice versa] if you’re doing work in the cloud. We’re focusing on data movement and how you synchronize it. But the next piece for us is to get even more granular and say, OK, you’re in the cloud; how can we help you edit, colorize, annotate, catalogue, and do other functions in the cloud?
We are also a Tier 1 supporter of the world’s biggest cloud providers. Our storage technologies are inside these massively scaled archives, and we help them with various tiers and their backup and archive strategies.
Do you see the cloud offering new business models for Quantum?
Definitely. We’ve been going to media and entertainment groups — like postproduction houses and sports organizations — and saying, What if we allowed you to create all the content and we run the storage and archive infrastructure for you? These organizations do storage management and manage a network because they have to, but really they just want to create great content.
The cloud may not be a real option for them, but I get why they want it: it’s easy, it’s simple, it’s elastic, and it’s subscription-based so you can ramp it up and ramp it down. What if we give you on-premises, full postproduction-workflow capabilities and ran it as if it was Amazon cloud? You write files to it, edit, and archive. If you need more storage, it shows up. You need more speed? We speed it up.
We send you a bill at the end of the month, and it can be on your terms. It can be [based on] terabytes or petabytes, but we can also charge you per movie, per event, per episode, per minute of game time.
It’s fully managed on-premises performance and quality with cloud economics and ease. This is a really good middle step. Everyone we’ve mentioned it to has been extremely interested. As it becomes viable, we’ll start pulling that equipment into a centralized data center. We have very extensive remote-monitoring and -management capabilities, so we can find our equipment and analyze it to see how it’s performing. We can service it from anywhere in the world. That’s been explosive for us.
LTO-tape is still one of Quantum’s biggest businesses. As the technological landscape changes, what role do you see for tape in the future?
Video data is exploding, and the world’s most economical way to store it remains on-premises tape. All the cloud providers use tape, and we’re happy to sell them tape, but they cannot get to the price point of just having your own tape robot in your basement. So our goal is to drastically simplify how to run those systems.
In terms of the data path that goes on the tape, we’re going to compress it, de-duplicate it, and ultimately erasure-code it. So we’re giving you the best economics for data as well.
Our tape business is about driving the very best economics per terabyte [or] per petabyte and putting it in a package that makes it really easy.
How are you looking to grow your market share outside North America?
We have a wonderful [customer base] in North America, but, yes, we are looking beyond that. We’re becoming less coupled to the hardware, so that, if [a customer] wants to use Chinese equipment in China or Indian equipment in India, they can do that. The tariff landscape is changing all the time, so it’s a lot easier to export a virtual machine vs. exporting a crate of hardware. We’re getting more virtualized; we’re hyper-converging so we can ram more pieces of software onto a smaller piece of hardware. That allows us to get to price points that make sense in India, in China, other parts of the world.
In September, Quantum completed its internal investigation and concluded that there were misstatements in previous financial statements. As CEO, how are you looking to lead the company out of this misstep?
Our leadership team feels strongly about running this business not only for technical leadership but also for leadership in terms of implementing best-practice policies, procedures, guardrails and things like that. CFO Mike Dodson joined when I did. He’s reconstructing the financial backend: our policies, the way we sell, the way we run supply chain, the way we account for things, the way our sales team interacts with pricing and discounting. All those fundamental pieces are being revamped to allow us to not only comply and resolve the past issues, but also to run a stronger, healthier business. We are working as fast and hard as possible to get there, but it takes time.
Where do you see Quantum headed in the next 12-18 months, and what are the company’s primary goals for that period?
First of all, within 18 months, we want to complete our turnaround in terms of economics and organization. I believe we will be able to do that.
Second, we have turned on our innovation engine, and I think, 18 months from now, you’ll see six to eight completely new products, from our hyper-converged product to NVMe-based products to different software analytic products.
You will see also see a whole new suite of products that are going to enable us to enter what I call vertical swim lanes. Media and entertainment workflows are informing everything we do. [For] any use case where there’s a huge amount of rich media, we want be the company [called first]. We want to lead the world in rich media. We do that today in media and entertainment, but there’s a broader set of programs that we have to run. I believe, within 24 months, we will be well-positioned as the hands-down leader in complex rich media, analytics, storage, and management.
Jamie Lerner will be on hand to deliver the Welcoming Remarks on Day 2 of The SVG Summit in NYC on Dec. 11. CLICK HERE for more information.