Scality co-founder and CEO Jerome Lecat wants to instill a culture at his company that is cognizant of every stakeholder, including the planet.
As part of his focus on environmental, social and corporate governance (ESG), Lecat is pushing beyond measuring his company's carbon footprint by taking proactive measures to hire more female engineers as well as documenting the company's procurement processes and the sustainability practices of procurement partners. These efforts were recently recognized with a silver medal by global ESG assessor EcoVadis, putting Scality in the top 25% of all companies it assessed this year.
A provider of petabyte-scale software-defined object and file storage, Scality made a name for itself by bending object storage, designed to be massively scalable, into a primary storage offering. Today, its flagship Ring is object storage software that can scale to 100s of petabytes.
In this Q&A, Lecat talks more about ESG as well as object storage's use in generative AI and how storage itself is a solved problem but that problems around storage, such as security, persist.
Editor's note: The following was edited for length and clarity.
Do you think the IT industry in general needs to take a bolder stance on ESG initiatives?
Jerome Lecat: I am going to be a little controversial here. I don't think IT is a problem. If you look at all industries and all countries, relatively speaking, IT is a good global citizen. I've read some articles about the increase of energy consumed by data centers. I think it's still below 2% of all the energy consumed on the planet. At least in the articles I have found, it's true that energy consumed by data centers is increasing, but it's increasing much more slowly than the usage. In other words, IT does more with the energy it consumes.
Included in the things that IT does is, for example, better management of energy grids, by which we can reduce energy consumption significantly. All in all, if you include the positive impact of IT -- and in this I include predictive maintenance, the ability to enable IoT in factories to decrease energy consumption in factories -- it is a net positive.
But what about the technologies storage uses, like hardware? What can IT do to reduce energy consumption there?
Lecat: Earlier this year we compared HDDs to SSDs. There has been a lot of hype lately by successful companies that suggests SSDs or silicon storage should be used for everything. One of the things they say is that silicon storage has less energy consumption than HDDs. When you read it, you're like, 'Yeah, sure, silicon takes less energy than spinning a drive.' It sounds logical.
We did some research looking at the specs from the hard drive manufacturer and from the SSD manufacturer, and it is far from obvious based on the specs today. This may evolve in the future, but today, SSDs consume more energy than HDDs. Once you spin something, such as a disk, the energy necessary to maintain the spinning is very low, especially at 7,000 rotations a minute, which is what we use.
A part of how we [operate] is to put [this kind of] data in front of customers and in front of the ecosystem to make better decisions for stakeholders and for the planet. As a software company, we don't have manufacturing facilities, we don't have as direct of an impact. But I do think we can make a huge impact on how the industry as a whole behaves.
Look at something like generative AI. It's become incredibly popular this year and does consume a lot of energy. It seems to use mainly file systems. But what is object storage's role in AI?
Lecat: All the generative AI projects we've looked at closely do not require a lot of storage. Instead they require fast access to storage. They do require huge amounts of energy to process. There is a potential problem with the energy consumption. So far, gen AI is just a small fraction of IT. They're a huge consumer of energy if you zoom in on them, but they're a small fraction of IT right now.
On storage, they're not a huge consumer of storage. We have quite a few customers that use us for the bulk of the data for AI projects. The data that feeds the NVIDIA GPUs is a subset of the bulk data -- that uses something faster than object storage. But the bulk of the data is too expensive to store on faster storage. When customers have 10s of petabytes of data that are going to be analyzed, object storage comes into play.
Unstructured data will triple by 2028, according to a recent Gartner report. Is object going to take more of a role in primary storage to deal with this growth?
Lecat: For me, primary storage is the storage that is in direct contact with humans. Humans are in direct contact with applications that request storage now as opposed to later. Using this definition, Scality has always been deployed as primary storage -- the first application we served was a large-scale email platform.
I do agree with people who say object storage is going to be used more and more as primary storage. An example would be CCTV cameras. The typical architecture of CCTV cameras is local NAS storage and then long-term storage on another media. Modern CCTV cameras can do TCP/IP to distant object storage.
I see applications where it makes sense to have object storage as primary storage. That being said, I don't think that object storage is the right type of storage for everything, such as relational databases, for example.
Block and file storage are still growing; it just appears that object is growing faster. Is that your perspective?
Lecat: I started Scality in 2010 with the idea that object storage would grow fast. Honestly, the growth was sluggish compared to what I expected in the first 10 years. In the past couple of years, we have seen an acceleration of growth and projects. It's related to the fact that more and more applications can write in the S3 API format.
Now if you take, for example, backup -- it's a natural use case for object storage. Five years ago, pretty much none of the backup software was able to write to object efficiently. Some had an S3 API, but it was not as efficient as their NAS storage. Nowadays, I'd say that Veeam and Commvault have an efficient use the S3 API.
Speaking of backups and data protection, adding security elements to storage is a growing trend. What is Scality doing on that front?
Lecat: In 2010, 80% of our engineers were working on storage, which you would expect for a storage company. Now it's only 25%. Storage is pretty much a solved problem, and now I have a whole team on security.
Security is a multi-dimensional [issue]. It's about enforcing multi-factor authentication everywhere in our products [as well as] directly interacting with the [key management service] of the customer so we don't hold the keys and the customer holds all they keys. It's about reusing the OS, the underlying version of Linux that's deployed with our systems, and making the Linux package as small as possible to be used as the attack surface. It's about scanning for [common vulnerabilities and exposures in] all the open source software we use to make sure we always have the latest version.
We are used as an element of security, and that is something new. It wasn't the case 10 years ago. Now people consider the storage of backup as part of their cybersecurity strategy, and we're used for backup. Backup is about 35% of our revenue.
You said storage is a solved problem. Do you think storage innovation has stalled or that its roadmap of bigger, faster, more secure just lines up with expectations and needs?
Lecat: At Scality, the big projects we have are around ease of use and security -- two things you can infinitely improve on. We are looking at storage temperature -- making it easier to move data from hot to warm to cold storage. We are looking at ways to improve how our storage will help analytical workloads. When I describe these, they are not directly storage projects.
Customers always demand more performance and more reliability. … However, I don't see from a customer demand, today, new problems that are fundamentally different that we haven't solved. … I'm not saying we have no work to do, just that I think storage is a solved problem.
Adam Armstrong is a TechTarget Editorial news writer covering file and block storage hardware, and private clouds. He previously worked at StorageReview.com.