 Welcome everybody to this CUBE Conversation. My name is Dave Vellante and we're here to talk about object storage and the momentum in the space and what Dell Technologies is doing to compete in this market. I'm joined today by Anahad Dhillon who's the product manager for Dell EMC's ECS and new object scale products. Anahad, welcome to the CUBE, good to see you. Thank you so much Dave. You appreciate you having me and Dell here, thanks. That's always a pleasure to have you guys on. We dig into the products, talk about the trends, talk about what customers are doing. Anahad, before the cloud, you know, object was this kind of a niche we've seen and you know, he had simple get put. It was a low cost bit bucket, essentially. But that's changing. Tell us some of the trends in the object storage market that you're observing and how Dell Technology sees this space evolving in the future, please. Absolutely, and you hear it right on, right? Historically, object storage was considered as cheap and deep place, right? Customers would use this for their backup data, archive data, so cheap and deep. No longer the case, right? As you pointed out, the object space is now maturing. It's a mature market now. And we're seeing out there customers using object for their primary data. So for their business critical data. So we're seeing, you know, big data analytics type of use cases. So it's no longer just cheap and deep. Now they're your primary workloads and your business critical workloads being put on an object storage now. Yeah, I mean, go ahead, please. Yeah, I was going to say, you know, there's not only the extent of the workload being put in, we're also seeing changes in how object storage is being deployed. So now we're seeing a tighter integration with new dev models where, you know, object storage or any storage in general is being deployed how applications are being built, right? So customers now want object storage or storage in general being orchestrated like they would orchestrate their customer applications. Those are, you know, few key trends that we're seeing out there today. So I want to dig into this a little bit with you because you're right, it used to be, it was cheap and deep, it was slow and it required sometimes application changes to accommodate. But so you mentioned a few of the trends, devs, you know, everybody's trying to inject AI into their applications. The world has gone, software defined. What are you doing to respond to all these changes and these trends? Absolutely, yeah. So we've been making tweaks to our object offering the ECS elastic cloud storage, you know, for a while. We started off tweaking the software itself, optimizing it for performance use cases. In 2020, early 2020, we actually introduced SSDs to our nodes. So customers were able to go in, you know, leverage these SSDs for metadata caching, improving their performance quite a bit. We use these SSDs for metadata caching. So the impact on the performance improvement was focused on, you know, smaller reads and writes. What we did now is, you know, a game changer we actually went ahead later in 2020, introduced an all flash appliance. So now EXF 900 and ECS all flash appliance, it's all NVMe based. So it's NVMe SSDs and we leverage NVMe over fabric for the back end connector. So we did it the right way. We didn't just go in and qualified an SSD based server and ran object storage on it. We invested time and effort into supporting NVMe over fabric. So we could give you that performance at scale. My object is known for scale. We're not talking, you know, 10, 12 nodes here. We're talking hundreds of nodes. And to provide you that kind of performance, we went ahead. Now you've got an NVMe based offering EXF 900 that you can deploy with confidence, run your primary workloads that require high throughput and low latency. We also are releasing our next gen SDS offering. But this takes the proven ECS code that our, you know, customers are familiar with, either provides the resiliency and then the security that you guys expect from delt. We're replatforming it to run on Kubernetes and be orchestrated by Kubernetes. This is what we announced that VMware 2021. If you guys haven't seen that, you guys go to go on demand for VMware 2021, search for object scale and you'll get a quick demo on that. With object scale now, customers can quickly deploy enterprise grade object storage on their existing environment, their existing infrastructure, things like VMware infrastructure, like VMware and infrastructure like OpenShift. I'll give you an example. So if you were a VMware shop, you've got vSphere clusters in your data center. With object scale, you'll be able to quickly deploy your object enterprise grade object offering from within vSphere. Or if you're an OpenShift customer, right? If you've got OpenShift deployed in your data center, in your red hat shop, you could easily go in, use that same infrastructure that your applications are running on, deploy object scale on top of your OpenShift infrastructure and make available object storage to your customers. So you've got the enterprise grade ECS appliance for your high throughput, low latency use cases at scale, and you've got this software-defined object scale which can deploy on your existing infrastructure whether it's VMware or red hat OpenShift. Okay, I got a lot of follow-up questions, but let me just go back to one of the earlier things you said. So object was going to cheap, deep and slow, but scale. And so your step one was metadata caching. Now, of course, my understanding is with object, the metadata and the data are within the object. So maybe you separated that and made it high performance, but now you've taken the next step to bring in NVMe infrastructure to really blow away all the old sort of scuzzy latency and all that stuff. Maybe you could just educate us a little bit on that if you don't mind. Absolutely, yeah, that was exactly the step approach that we took, even though metadata is tightly integrated with object, in object world, in order to read the actual data, you still got to get to the metadata first, right? So we would cache that metadata into SSDs, reducing that lookup that happens for that metadata, right? And that's why it gave you the performance benefit, because it was just tied to metadata lookups, the performance for larger objects stayed the same because the actual data read was still happening from the hard drives, right? With the new EXF 900, which is all NVMe based, we've optimized our ECS object code, leverage NVMe, data sitting on NVMe drives, the internote connectivity, the communication is NVMe over fabric. So it's through and through NVMe. Now we're talking milliseconds in latency and in thousands and thousands of transactions per second. Got it. Okay, so this is really an inflection point for objects. So these are pretty interesting times at Dell. You got the cloud expanding on-prem, your company's building cloud-like capabilities to connect on-prem to the cloud, across cloud, you're going out to the edge. As it pertains to object storage though, it sounds like you're taking sort of a two-product approach to your strategy. Why is that? And Chin, you talk about the go-to-market strategy in that regard. Absolutely, and yeah, no good observation there. So yes and no. So we continue to invest in ECS. ECS continues to stay a product of choice when customer wants that traditional appliance deployment model. But this is the single hand-to-shape model where everything from your hardware to your software, the object solution software is all provided by Dell. ECS continues to be the product where customers are looking forward that high-performance, fine-tuned appliance use case. Object scale comes into play when the needs are software defined, when you need to deploy the storage solution on top of the same infrastructure that your applications are running. So yes, in the short term, in the interim, it's a two-product approach of both products taking a very distinct use case. However, in the long term, we're merging the two code streams. So in the long term, if you're an ECS customer and you're running ECS, you will have an in-place data upgrade to object scale. So we're not talking about no forklift upgrades. We're not talking about adding additional servers and do a data migration. It's a code upgrade. And then I'll give you an example. Today on ECS, we're at code version 3.6. So if you're a customer running ECS, ECS 3.x in the future, so we've got a roadmap where 3.7 is coming out later on this year. So from 3.x, customers will upgrade the code data in place. Let's call it 4.0, right? And that brings them up to object scale. So there's no notes left behind. There's an in-place code upgrade from ECS to object scale, merging the two code streams. So in the long term, single code, short term, two products are both solving a very distinct use case. Okay, let me follow up on my customer hat. And I'm hearing that you can tell us with confidence that irrespective of whether a customer invested ECS or object scale, you're not going to put me into a dead end. Every customer is going to have a path forward as long as their ECS code is up to date. Is that correct? Absolutely, exactly. And very well put. Yes, no notes left behind. Investment protection, whether you've got ECS today or you want to invest into ECS or object scale in the future. Correct. Talk a little bit more about object scale. I'm interested in kind of what's new there, what's special about this product. Is there unique functionality that you're adding to the product? What differentiates it from other object stores? Absolutely, my pleasure. Yeah, so I'll start by reiterating that object scale. It's built on that proven ECS code, right? It's the enterprise-grade reliability and security that our customers expect from Dell EMC, right? Now we are replatforming ECS to allow object scale to be Kubernetes-native, right? So we're leveraging that microservices-based architecture, leveraging that native orchestration capabilities of Kubernetes, things like resource isolation or seamless self-peeling, I'm sorry, load balancing and things like that, right? So the in-built native capabilities of Kubernetes. Object scale is also built with scale in mind, right? So it delivers limitless scale. So you could start with terabytes and then go up to petabytes and beyond. So unlike other file system-based object offerings, object scale software won't have a limit on your number of object stores, number of buckets, number of objects you store. It's limitless. As long as you can provide the hardware resources under the covers, the software itself is limitless. It allows our customers to start small, so you could start as small as three nodes and grow their environment as your business grows, right? Hundreds of nodes. With object scale, you can deploy workloads at public cloud-like scale or with the reliability and control of a private cloud day. Right, so it's in your own data center. And object scale is S3 compliant, right? So while delivering the enterprise features like global replication, native multi-tenancy, fueling everything from DevTest Sandbox to globally distributed data limits, right? So you've got in-built object scale replication that allows you to place your data anywhere you've got object scale. From edge to core to data center. Okay, so it fits into the Kubernetes world. I call it, you know, Kubernetes compatible. You know, the key there is automation because that's the whole point of containers, right? It allows you to deploy as many apps as you need to, wherever you need to, in as many instances, and then do rolling updates, have the same security, same API, all that level of consistency. So that's really important. That's how modern apps are being developed. We're in a new age here. It's no longer about the machines. It's about infrastructure as code. So once object scale is generally available, which I think is soon, I think it's this year. What should customers do? What's their next step? Reach out to your Dell representatives, right, get an in-depth demo on object scale. Better yet, you get a POC, right? Get a proof of concept, have it, you know, set up in your data center and play with it. You can also download the free, full-featured community edition. We're going to have a community edition that's free up to 30 terabytes of usage. It's full-featured, download that, play with it. If you like it, you can upgrade that free community edition with a licensed paid version. And you said that's full-featured. You're not neutering the community edition. Exactly, absolutely. It's full-featured. Nice, that's a great strategy. We're confident. We're confident in what we're delivering. And we want you guys to play with it without having, you know, your money tied up. Nice, I mean, that's the model today, you know, gone are the days where you got to get customers in a headlock to get them to, you know, they want to try it before they buy. So that's a great little feature. Hanaha, thanks so much for joining us on theCUBE. Sounds like it's been a very busy year and it's going to continue to be so. Look forward to see what's coming out with ECS and object scale and seeing those two worlds come together. Thank you. Yeah, absolutely. It was a pleasure. Thank you so much. All right, and thank you for watching this Cube conversation. This is Dave Vellante. We'll see you next time.