00:19 Eric Burgener: Hi, this is Eric Burgener again. Well, we're about ready to wrap up the event, and I wanted to make some closing comments about solid-state technologies and how you might consider using those in your own data center and as you build your infrastructure.
So, first off, I want to talk a little bit about workload placement. What's the appropriate location for each workload that I have in my portfolio? We've done some research on this topic just earlier this year, and clearly what's driving the decision about whether workload should be placed in traditional IT or left in traditional IT, placed in a private cloud infrastructure or moved to a public cloud environment. So, workload requirements are driving that, what are the performance, the availability, the ease of use, the flexibility and agility requirements for a particular workload, and then obviously, cost is another issue.
01:14 EB: Now, with a public cloud, there are a lot of services available in the public cloud that enterprises may not have access to in on-premise infrastructure. So, another potential reason to move things to the cloud is if you want to access, for example, particular compute services, CPU-based compute services for artificial intelligence, machine learning workloads. Maybe you want to try things out on the public cloud, see how this might contribute to your own business before you decide to invest in it and move it on prem, so where you place workloads is clearly a key issue. And one of the things to keep in mind as you're building on-premise infrastructure is what are the opportunities for consolidation. Certainly, as you move forward, technologies become more efficient, they get better performance, increased density, but there's also an opportunity to consolidate workloads onto fewer platforms.
02:08 EB: I made a few comments about that in my session, so I just want to leave you with, please keep that top of mind, the new technologies that are available on the storage side in particular make it a lot easier to be able to combine workloads on the same platform, a wide range of workloads with different performance and availability requirements. And now with so many of the systems being software-defined, you have the multi-tenant management flexibility to be able to configure particular storage pools to meet very particular application requirements. So, that's going to be a very good way to leverage some much better economics is if you can consolidate workloads while still meeting your own service-level objectives.
02:54 EB: OK, the second thing I wanted to talk about is NVMe, and thinking about how you'll leverage the NVMe technologies, persistent memory, storage class memory, NVMe over Fabrics, in your own environment. You may not need that type of performance yet. You may be putting workloads in place in the next couple of years that will require that kind of performance. Certainly, one of the key things as you buy new storage infrastructure today is, "Can I move to NVMe in the future nondisruptively, even if I might not need that performance capability today?"
I think I had commented in my session that a number of the established storage vendors basically are providing their NVMe-based systems at the same price point that they sell a comparably configured SCSI system. So, in many cases, there is no price premium to NVMe if you buy at the system level, and certainly being ready for those newer workloads that may demand the kind of performance that you can only get from, for example, storage class memory, you'll be ready for that if you bought a system that uses NVMe.
04:03 EB: The other thing I would mention that's related to this is with the NVMe over Fabrics networking and how that's developing. This really opens up the opportunity for some new architectures that we talked about, and basically to remove data locality as a concern as you build out storage platforms, and this really opens up a lot of flexibility. So, that may not be something that you're ready to go to now, you may not need that kind of performance, but it's something to think about as you're buying storage infrastructure, because again, two, three years down the line, well within the storage lifecycle of the systems that you're buying today, you may need that kind of performance. And it would be great if you could just basically do a software upgrade and take advantage of that kind of performance.
04:52 EB: The final area is to think more holistically about what you could potentially do differently if you had the kind of performance that NVMe brings to the table. What new applications could you build? Could you increase the volatility of certain processes that might drive higher commission revenue, better customer satisfaction, better customer support if you had this ability to run things at a higher level of performance? So, it's not just a matter of taking the same applications, moving them into these environments, but also thinking about how those applications might be able to be evolved over time to contribute in a better way to your business, or in fact, even to maybe open up new business opportunities. OK.
05:42 EB: So, that pretty much brings everything to a close. I just want to say that IDC has a booth at the show, it's a virtual booth. If you're interested in chatting about any of the content that you saw in our presentations today, or you'd like to talk to IDC about other issues that might be top of mind, we'd invite you to drop by. We'll have that booth staffed on November 11th, that's today, and November 12th. So, please drop by if you'd like to chat. We'd love to talk with you. Thanks again for coming to our event. And we'll sign off now.