Today, Hitachi Vantara announced the VSP E990, a midrange all-NVMe storage appliance. The new solution, available immediately, offers up to 5.8 million IOPS, I/O latency as low as 64µs with up to 64PB effective capacity. In this episode, Chris Evans chats remotely to Colin Gallagher about the new platform.
VSP E990 extends the existing solutions in the VSP family, from the F350 up to the 5000 series enterprise-class solutions. The “E” denotes NVMe – other platform models have “F” (flash) and “G” (general) designations. The quoted 64µs latency figure is lower than the VSP 5000 announced at NEXT last year (see the previous podcast recording on this). As a dual-controller array, E990 provides slightly better performance, shaving 6µs off the 5100 latency figures.
VSP E990 is based on the SVOS storage operating system used across all Hitachi VSP storage solutions. This provides consistency and standardisation for inter-family replication and management. VSP E990 incorporates Intelligent Data Reduction with a 4:1 “sight unseen” guarantee. The solution is also “SCM Ready”.
Chris talks to Colin about the evolution of the midrange market. As enterprise requirements change, mid-market solutions are becoming more important than ever. With this announcement, Hitachi Vantara is introducing EverFlex, a range of consumption models from purchase to “as a service”.
You can read more on the technical aspects of the E990 announcement on Colin’s blog. Here’s a link to the VSP E990 data sheet.
Elapsed Time: 00:35:41
Timeline
- 00:00:00 – Intros
- 00:57:00 – What is the VSP E990?
- 00:03:49 – E990 is NVMe only
- 00:04:36 – E990 is 64µs compared to 70µs with VSP 5000
- 00:07:30 – SVOS runs across all platforms for consistent management
- 00:09:00 – What is Intelligent Data Reduction?
- 00:14:05 – What is “SCM Ready”?
- 00:18:40 – What protocol support will there be?
- 00:21:10 – How does the E990 fit into the portfolio?
- 00:25:25 – Is the midrange market the growth storage market?
- 00:28:35 – Hitachi OpsCenter 10.1 supports VSP E990
- 00:31:36 – EverFlex from Hitachi Vantara introduces new consumption models
- 00:34:16 – Wrap Up
Transcript
[bg_collapse view=”button-orange” color=”#72777c” icon=”arrow” expand_text=”Show More” collapse_text=”Show Less” ]Speaker 1:
This is Storage Unpacked. Subscribe at storageunpacked.com.
Chris Evans:
Hi, this is Chris Evans recording another Storage Unpacked Podcast. I’m here with Colin Gallagher from Hitachi Vantara. Hi, Colin.
Colin Gallagher:
Hey, Chris.
Chris Evans:
How are you doing?
Colin Gallagher:
I’m doing well. I’m doing well. Starved for human company, so great to spend some time talking with you.
Chris Evans:
That’s very kind of you to say. The last time we saw each other was at NEXT and we were face to face talking about the VSP 5000 launch. Sadly, no longer face to face. We’re remote because of the circumstances. But it is good to catch up.
Colin Gallagher:
Likewise. Totally appreciate it.
Chris Evans:
So you’ve just announced a new platform. It’s called the E990, and we’re here to talk about that today and understand exactly what that is and what it does. So why don’t we just get straight into it and talk about what the platform is, what’s being announced and give us some detail.
Colin Gallagher:
Excellent. Yeah. I’m very pleased to announce a new addition to the VSP family, the new VSP E990. It’s a high performance, all-flash NVMe array that really helps our mid-size customers deliver predictable and consistent performance for their business critical workloads.
Chris Evans:
Brilliant. So let’s just talk about some of the detail because if it’s NVMe, I imagine it’s going to be high performance. It’s going to be a powerful box. What speeds and feeds are you causing with this new system?
Colin Gallagher:
It delivers up to 5.8 million IOPS and latency is as low as 64 microseconds.
Chris Evans:
That’s pretty good.
Colin Gallagher:
Yeah, we’re really impressed with it. It’s a really powerful platform and it has the capability to accelerate your applications today and provide a lot of head room for future growth, should you need it.
Chris Evans:
Okay. What terms of in terms of capacity, is this box going to be able to go up to?
Colin Gallagher:
It scales from as low as six terabytes of raw capacity to an impressive 16 petabytes of effective capacity.
Chris Evans:
Okay. I’m guessing that’s not in the same box though, is it?
Colin Gallagher:
No. To get that full 16, you actually have to use external storage. And again, one of the great powers of the VSP architecture is the ability to virtualize external storage. So this box can support up to 16 petabytes of internal and external storage.
Chris Evans:
Okay. Brilliant. Let’s go ahead and talk about a bit more detail about what it might look like, because I think people will be interested because you’ve gone for NVMe to get a rough idea of what the architecture looks like. So for sake of argument, what is a base configuration going to look like? Is it going to be expandable? What size drives are you going to put into the system, that sort of thing.
Colin Gallagher:
Yeah, absolutely. And let me take a step back and say when we were designing this system, we really looked at how we could deliver NVMe in a smaller package. As you know, Chris, when we last met face to face, we talked about introducing the VSP 5000 and interviewing me for that. We actually took that capability and that code, because remember, the entire VSP family runs a common S VOS operating system. And so we took that code and optimizations we designed for the VSP 5000 series around NVMe, and we brought those down into this dual-controller platform.
Colin Gallagher:
So that was one of the key innovations that we did. Obviously, we also have a pretty significant powerful dual-controller platform as it stands, with up to 56 processor cores. And the cores that we use in this product are faster, on average of 40% faster than what we delivered in our previous top of the mid-range platform, which was the F900. So it’s really about synthesizing that innovation and what we’ve designed for the 5000 series and what we’ve also had in market already for mid-range platforms.
Chris Evans:
Okay. Will this be a purely NVMe device or will it be like the 5000 where there was a lot of SaaS in that device, wasn’t there?
Colin Gallagher:
Yeah, no, the 5000 series supports a very wide variety of media types. This box today only supports NVMe, but we are announcing that it is going to be SEM ready and SEM upgradable in the future.
Chris Evans:
Okay. We’ll come back to that point in a moment because I want to dig into what that really means. But NVMe device, potentially expandable or literally just a single chassis?
Colin Gallagher:
No, we do have expansion controllers for this as well. The is E990 is highly expandable. We start as low as eight drives and expand all the way up to 96.
Chris Evans:
Are these FMDs at all or are these just traditional drives?
Colin Gallagher:
These are traditional drives. We’re not supporting FMDs in this platform.
Chris Evans:
Okay, brilliant. So I think that gives us a good bit of background about what the platform looks like, the sort of speeds and feeds we can expect. But let’s go back and talk about that because 64 microseconds is an interesting number. First of all, I remember looking back because I actually went and checked, that the VSP that we talked about the last time we met was looking at 70 microseconds as its response time. And of course, that was using SaaS technology. I don’t think there was any NVMe in that platform at the time. It was quite interesting that you were getting that level of performance, but now you’re talking about 64 microseconds, an even faster response time. And the first thing that strikes me there is that a lot of traditional NVMe SSDs probably struggle to get down to that level of performance in the first place. So how are you doing that? How can customers sort of understand that 64 microsecond number?
Colin Gallagher:
Chris, first and foremost, this is why I love talking to you. You’re the first person to pick up on that 64 versus 70 microsecond difference and it’s really important because that actually shows the power of our architecture and you hit it on correctly that the 70 microsecond latency number that we published for the VSP 5000 series could be SaaS. It can be NVMe as well. And it’s really about the power of that architecture that we can deliver that level of response time irrespective of the underlying media.
Colin Gallagher:
As I mentioned, we took that code that we built for the VSP 5000 series, which is a multi-controller architecture, and we brought it down to the VSP E990 which is a dual-controller architecture. And it does seem a little contrary to conventional wisdom that a dual control architecture would get slightly better performance. But again, if you think about it, when you have a multi-controller architecture, there’s a teeny bit of overhead that you need to have to manage the multiple controller halves. And by eliminating that, we can actually even squeak out a little bit more response time. And that’s what takes us from 70 to 64.
Chris Evans:
Okay. We’ll come onto product positioning in a second, but I think that’s an interesting point for people to understand that modern architectures are incredibly reliable. And even though you said dual-controller architecture, a lot of technologies out there are still dual-controller architecture. And those architectures are very reliable in modern platforms, but the 5000 was very much the high-end super reliable type platform and also had mainframe support anyway, which is a slightly different market.
Colin Gallagher:
Yeah, totally. I mean, it is a different platform designed for a different customer type at a different capacity and to be honest, price point. Right? The mid-range market is pretty broad. There are a wide variety of needs in the market and as you said, we’ll talk about the rest of the portfolio hopefully shortly, but we understand that the VSP E990 and NVMe in particular suits a particular portion of that market. We have other members of the VSP family that meet other capacity and performance points if customers aren’t looking to go NVMe.
Chris Evans:
Okay. Now, you’ve said that you’re bringing this platform technology down from the 5000 series. Is this the S VOS technology that we know exists in that platform already today?
Colin Gallagher:
Exactly. It’s the same S VOS we run in the VSP 5000 series that’s been adapted to run on a dual-controller architecture. And that’s been our strategy for a long time, is to have one storage operating system across our entire storage family from small to large that allows customers to take advantage of common data services, inter-family replication and same management no matter what size, what capacity, what point in their storage life cycle they are.
Chris Evans:
So that’s really quite, I think, an important factor there because ultimately features are delivered in software and hardware is there for things like resiliency and cost and scalability and other aspects like that. But when it comes down to features, they’re very much in the software side of things and being able to, as you just mentioned there, replicate in family and do things like that, that becomes an interesting and quite useful requirement if you are a large customer that might want a mixture of different platforms in their portfolio, I guess.
Colin Gallagher:
Yeah, you don’t have to have to learn different management tools. You don’t have to learn different best practices for different platforms. You get to leverage a standard runbook across your entire data center. A lot of customers run multiple storage arrays. And if you have that common and consistency across the family, it makes it simpler and easier for customers to manage, removes that friction and overhead and to be honest, that chance for human error.
Chris Evans:
Yeah. Let’s dig down then into a little bit more detail about some of the announcements that are part of this platform. One of the ones I was quite interested in and picked up on was something called intelligent data reduction. Now this is your technology for doing dedup compression. It’s that effectively dead reduction technology that many vendors have in their platforms. But what interests me here was that you’re doing what you’ve called a sight unseen four to one guarantee. I guess by sight unseen, what you’re saying is you’re not putting any caveats or prerequisites on the type of data that you’re going to offer this on.
Colin Gallagher:
Slightly different. We are guaranteeing you four to one with… As a marketer, I would love to come out and say no fine print. Unfortunately, we can’t say that. There is a small amount of fine print. But, Chris, it’s really honest and upfront and sane. Obviously, we exclude from that four to one guarantee data that’s already been pre-compressed or encrypted where we can’t access the data, for example. It’s a sane upfront guarantee but it covers a wide variety of data types. But we exclude things where we obviously know we’re not going to be able to get that, and we would expect any reasonable person wouldn’t be able to expect that as well.
Chris Evans:
That’s the reason why I was asking the question because as you said, things like encrypted data where we know it’s effectively randomized, you’re not just not going to get anything out of that. And I don’t think anybody would expect you to.
Colin Gallagher:
Yes. But so it is, there are a few data types that are eliminated but they are a few. But what we mean by sight unseen is that you don’t need to run any tools or have an assessment done on your data for us to guarantee this. We introduced this ADR, this AI power data reduction technology in the VSP 5000 series and this is another one of those capabilities that we’ve leveraged from the common S VOS operating system and brought into the E990, and we’ve seen tremendous success with that and the VSP 5000 series. We know and understand what it’s going to do for other workloads as well.
Colin Gallagher:
We firmly stand behind this guarantee and I think we offer one of the best guarantees in the industry. When you craft these guarantees, it’s always about the fine print and it’s always about the remediation. We promise that we will remediate up to the full capacity promised. If you look at what else some other vendors are doing in the industry, they have caps on the amount of capacity, which really doesn’t make it enough true effective guarantee or they even will compensate you with capacity. They’ll compensate you with other financial things, instruments, sorry, like service guarantees or service credits or other types of things like that.
Chris Evans:
Yeah, I mean we’ve had this discussion before on a different podcast where we talked about exactly what vendors would guarantee and what they wouldn’t. And clearly, a lot of the time, the most obvious thing people would want is hard cash in their hand, but nobody’s going to give them hard cash. But what necessarily you don’t want is some sort of fairly weak guarantee that says, “Okay, we’ll give you service credits,” because that doesn’t solve the problem of the customer having to put their hand in their pocket to put more equipment on site if what they thought they were going to get didn’t work. So I think people look at it and will want to understand that the remediation, if there is a problem, it’s practical enough for them to get out of a hole if it does cause a problem for them.
Colin Gallagher:
Yeah, we want to be upfront and honest with our customers. We aren’t trying to trick or delude them into buying something from us. So if there is some bizarre instance where we can’t get that four to one or better, because we are seeing even better, right, we’re guaranteeing a minimum of four to one, we’ll make them whole. We’ll give them the credit for the capacity to get them up to that four to one level.
Chris Evans:
You said this is part of the S VOS platform. So is this technology that’s been worked on specifically for this platform or has this been an evolution of the technology?
Colin Gallagher:
Yeah, Chris, this was first introduced as part of the VSP 5000 series and we’re leveraging that same common S VOS operating system to bring this capability into the E990. Really, what’s really from a technical point of view interesting about our ADR technology is that it processes the IO in multiple stages.
Colin Gallagher:
So the first, we read data into the cache. So we’ve got it. We’ve kept it. The data is safe. But then we have an adaptive AI engine that analyze the data and can actually switch the processing between inline deduplication, offload deduplication or post-processing deduplication to make sure that we get not only the most efficient deduplication possible, but we also, that process of deduplication minimizes the effect on the workload itself.
Colin Gallagher:
For example, a highly random IO pattern that’s being pushed to the box. Once it hits cache, we’ll actually handle that in post-process so we don’t slow down system performance at all.
Chris Evans:
That’s really quite an important factor, I think, because if you look at the fact that you’re quoting 64 microseconds, and that was an area I was tempted to dig into and be a bit more awkward about, to be honest. And that’s to say, well, 64 microseconds, that could be a hero number. How accurate is that really going to be? And one of the challenges there is as soon as you start putting data services on top of your platform, you can see those sort of response times in the throughput go down dramatically. So it sounds to me like you’ve built something that’s adapting to that to make sure that doesn’t happen.
Colin Gallagher:
Yeah, I mean obviously to your point, data services do consume processing power and they can interrupt the IO path, but we’ve done everything possible to minimize that. Again, by leveraging this AI capability and adaptively adjusting how we do deduplication, by the way, compression is always in line, but deduplication is done adaptively, that really allows us to get the best performance out of the architecture.
Chris Evans:
Okay. So let’s go and talk then about SEM, what some people might call assistant memory. By that we mean effectively Optane, that sort of thing, 3D XPoint, technology where we can write to a device that could be either plugged in as memory or could be plugged in as an SSD or as an NVMe device, but we can write to it at a byte level rather than the block level.
Chris Evans:
Now you say SEM ready and I’ve seen that term used by other vendors, but I’m really interested to understand exactly what you mean by that and dig into how you think you’re going to use the SEM within your platform.
Colin Gallagher:
Yeah. So we understand that customers are looking to deploy SEM at some point in the future. To be honest, particularly in the mid-range, customers are interested in it, but we haven’t seen great demand for it yet. But when we do talk to customers and we were doing our sort of NDA briefings with them, they wanted to know that a year or two in the future, as they deploy them, we would be able to support them so they wouldn’t have to buy new architecture, new infrastructure as well. So we are announcing that this, the E990 is SEM ready and that shortly into the lifecycle of this product, we’ll deliver a non-disruptive upgrade to add SEM capacity to the box.
Colin Gallagher:
What’s important about this and the distinguishing factor is that unlike some others, we’re not using SEM as cache. We intend to treat SEM as a first class storage citizen within the array and use it for persistent data storage, should customers want to do that.
Chris Evans:
Okay. Well, do you do it as cache as well, or is it literally going to be that tiering feature?
Colin Gallagher:
It’s going to be that tiering feature. Yeah. We’ll continue to leverage [crosstalk 00:15:36] for cache as we do today and we’ll use it as a tiering feature within the array.
Chris Evans:
Do you have any idea how much that’s likely to change performance, what that’s likely to cost in terms of the technology?
Colin Gallagher:
Costs are still TBD. Obviously, we’ll talk about that as we get closer to releasing it. Because obviously, the costs are driven by a lot of factors, the market demand, et cetera. But again, on average, let’s just take a step back about costs and we’ll talk about the NVMe costs. One of the things that we’re proud to talk about with the E990 is that we believe it’s highly cost-effective NVMe technology. So we aren’t pricing NVMe at a crazy premium like we’ve seen someplace in the markets and we’re actually really focused on delivering this NVMe performance at really affordable prices. We can deliver per IOPS performance for as low as three cents an IOP or IOPS. Is it still plural? Yes, it is.
Chris Evans:
I’m not sure, yeah. [crosstalk 00:16:30].
Colin Gallagher:
Yeah. IOPS per second. IOPS per second. It’s per second. It’s not a plural, so yeah, it’s correct. It’s as low as three cents per IOPS. Yes.
Chris Evans:
Okay. With SEM then, it will come in as another tier. I think it’s always interesting to see how that balance works because as a customer, you’re always trying to work out how you’re going to move that data around to make use of the SEM. And something we haven’t talked about because I haven’t seen that as a feature in this particular release, but I’m assuming that you have the same technologies you’ve always had in the platform that allows me to auto-tier onto that SEM to make effective use of it because otherwise it becomes very difficult to make practical use for at that point.
Colin Gallagher:
Exactly. Chris, I mean this is if we go back to 2009, 2010 when flash drives first started appearing and enterprise storage arrays, this was the same challenge, right? Putting a whole data volume on a large flash pool was prohibited, costs expensive and you really needed to have that tiering capability. So that’s sort of what we’re focused on is delivering the ability to tier between SEM and NVMe media here. By the way, if you recall, when we introduced the 5000 series, we said that that will be SEM ready as well. And there, you have even more tiering options because as we discussed, it supports a wider range of media. But again, you could do SEM, NVMe, SaaS, SSD, spinning media, et cetera. But it’s really about delivering a full set of data services, dedup, compression, tiering, all those things that really made flash valuable. It took several years for flash to come to market to get them all. Delivering all those same data service capabilities on top of SEM as well.
Chris Evans:
Okay. So I look forward to seeing that one because I think obviously in that, we’ve seen sort of limited size capacity devices, but I think we’re on that cusp really with the SEM market where we’re trying to see whether cost is going to come down, whether the capacities will increase and we need a little bit of driving of that market. So I think we are at the cusp of potentially seeing some really good adoption there.
Colin Gallagher:
Yeah, no. I’m excited about that, too. I said it’s once these data services and other capabilities take advantage, it will really expose a whole new level of performance for customers, much like flash did back in the day when it was introduced and exposed performance that a 10K or 15K drive could never deliver, right? So we’re still looking for that same thing around SEM as well.
Chris Evans:
So one final thing on the detail as we’ve gone down a bit deeper and that’s protocol support because clearly now we’re seeing a lot of NVMe over fabrics both on Fibre Channel and on ethernet. What’s the strategy there in terms of how this is going to develop?
Colin Gallagher:
Yeah, Chris, the VSP 990 like the 5000 series is NVMe fabrics ready as well. And that’s something that we’ll deliver again through a non-disruptive lifecycle upgrade not too far down the future. Today, it supports all of your traditional ethernet and Fibre Channel connectivity protocols.
Chris Evans:
Do you see customers being super ready for NVMe over fabrics yet? I mean, it sort of strikes me as it’s like a bit like SEM. It’s almost there and it’s probably got some very good niche cases, but customers going wholesale into it yet might be a bit early. So I think you’re almost in a position where we’ve talked about it for a couple of years and I mean the market has but it hasn’t really seen that adoption yet. Possibly because, A, customers haven’t got the need for that level of super performance, but also some of the challenges around the protocol were still being… the features were being filled in. So are you seeing a huge adoption? Are you seeing just people asking for it? What’s your view of the market?
Colin Gallagher:
I think you hit it right on the head, Chris, that there is a lot of interest in it, but some of those implementation details are just coming together now. As customers are looking to deploy this in the future in general… I mean there are a few places where we’ve seen customers are deploying it actively today, but that’s rare. That’s rare. What we are seeing is that as with any network transition, that has to be planned and managed. It involves reaching out and beyond the storage team. So those tend to delay those types of implementations as well. But we are seeing active interest in it. We are seeing customers looking at it and being sure that what they buy today can non-disruptively support that as they choose to move to that in the future.
Chris Evans:
Yeah. I think that’s a fair point because if you’re building out an infrastructure today and you’re about to be on the cusp of moving to an infrastructure where you want to choose to eliminate, say a Fibre Channel or something like that, if you bought a whole lot of technology and you think it’s not going to be supported, I think that would be a bit of a mistake to make at this point. So even if you don’t intend to use it in the next six months, 12 months, you might be buying technology you’re keeping on the floor these days for maybe five years. So you do need to make sure that supports that.
Colin Gallagher:
Yeah, exactly. And things like network decisions and fabric decisions tend to involve more than just the storage team. So they may be looking to solve a storage problem today, but by upgrading to NVMe performance and then working with their application teams, their server teams and their network teams, doing a larger replacement of their fabrics two, three years down the road or sooner if they’re ready for that.
Chris Evans:
Okay. Let’s go on and try and help people understand where this new platform fits into your portfolio. So we previously talked about the 5000 series, 5100. Now that is pretty much high-end, what people might have thought was the device that was connected to the mainframe. So you know probably the top of the range, if the better way to describe it. But could you put together a little view for me about the portfolio itself and just help me understand where everything fits?
Colin Gallagher:
Yeah. As I mentioned, so we are talking today a lot about the new VSP E990, the shiny new toy, so to speak, though I would never call any of my products a toy, but the shiny new capabilities. But it sits within a larger virtual storage platform portfolio. This is one storage family that can scale from the micro cloud all the way up to a large data center. We start that with the VSP F350 and 370 which are compact all-flash appliances in a two form factor.
Colin Gallagher:
We then scale up to the VSP F700 and F900 which deliver high performance, simplified storage but at broader performance or broader capacity points. Then we have the VSP E990 which is, we’ve talked about a lot today. And on top of that, we have the VSP 5000 series with the 5100 and the 5500, as you point out, delivering true scale up, scale out capabilities, mainframe support, basically the most powerful and flexible storage platform available.
Chris Evans:
And across all of that, it’s the same software. Again, we just want to reemphasize that because I think that’s quite important from my perspective to understand.
Colin Gallagher:
No, totally. And we’ve put not just the same software and the common data services, but the same management console approaches with Hitachi Ops Center, the same a hundred percent data availability guarantee across the family and some the same purchasing programs as well, which I’m sure we’ll talk about in a little bit.
Chris Evans:
Yeah, sure. So E then. E for NVMe?
Colin Gallagher:
Yes, E for NVMe. Yes. I love talking about the [vagraties 00:23:13] of naming because it’s one of the interesting parts of my job. But let’s start back. So we have the VSP. We’ve talked about the F lot and the 350 and 370, et cetera. The F stands for flash, which is pretty self-explanatory. We also have variants of those that are our G models, the G350, 370, et cetera. And those are general purpose. Those are hybrid arrays. Then we also have our NAS platforms which are begin with an N. So unfortunately when we sort of were planning a dedicated NVMe only array, N was already taken, so we had to use another distinguishing letter, and since we had F and G and E was close and E was in the NVMe name, we went with E.
Chris Evans:
That makes sense.
–
Chris Evans:
Yeah, you’ve got to choose something. And at the end of the day, it’s not always straight forward how you choose something, but I guess that makes sense.
Colin Gallagher:
Yeah. But the other interesting point about this is these naming conventions were sort of forced on the storage industry. If you look at it across other companies, everybody sort of had to have different flash models for their all-flash systems. Because originally when all-flash systems came out, that’s how analysts were evaluating them. If you didn’t have a specific model for all-flash, they would not consider it in that category.
Chris Evans:
There was a bit of debate around that, around the need to keep a separate product line that particularly couldn’t take traditional drives just so that… I think there’s a bit of an accounting thing going on in there, but it helps for people to understand that.
Colin Gallagher:
Luckily though, most analytics have sort of backed off. They come to their senses over the last decade, and they’re sort of collapsing and have a more holistic view of things now. That’s why, by the way, when we went with the VSP 5000 series, we didn’t feel that we had to create different model numbers for flash or hybrid or NVMe, et cetera, because that is more modern thinking going forward.
Chris Evans:
Let’s talk about positioning then and the market because one of the things I find really interesting about the way that the market’s developing is that if you look back 10 or 20 years ago, if we were talking about the enterprise level, we would have been talking about the 5000 series has been the typical device that a large enterprise would be deploying. Now since then, the technology has become much more reliable across the board. And we’ve got much more scalable and reliable and very fast products. So it seems to me that the mid-range market might be a bit more appealing to customers than it used to be simply because the products are more reliable. What are you seeing in terms of the positioning of this product and the market in general?
Colin Gallagher:
I think you’re totally right on that, Chris. I mean, I think if you go back to a different Chris, Chris Clayton. If you ever read The Innovator’s Dilemma, that’s the whole struggle.
Chris Evans:
I did, yep.
Colin Gallagher:
Yes. Which is that when you serve a premium class product, oftentimes the product that is below it increases in capacity to provide good enough capabilities. Right? I think that happens in every industry and that has happened over time in the storage industry as well. But what’s really nice about our approach is that because, and to harp on this for the millionth time right now, because we leverage a common operating system architecture, we can actually port those capabilities down. So it’s about more about refactoring the IP for the particular customer needs, than building new products are being displaced by products that are lower down in the industry.
Colin Gallagher:
I mean, the mid-range in general was always a broad market with a broad set of needs and capability requirements. We’ve always lumped it as mid-range and high-end. But it’s been very hard to define mid-range. There’s a lot of different storage types, a lot of different applications, workloads, everything from a small mom and pop shop to a very large several thousand person company or even larger, all fit within that. And they all have different needs. Right? So it’s really about how do you deliver different capabilities for each of those needs in the most effective way possible?
Chris Evans:
I would say that if you look at the market today, we’ve moved to a situation where a large number of features are being pushed into the application maybe for recovery or we see virtualization being very strong. So replication might be done at an application or a virtualization level, so maybe people don’t always need the super high, super expensive synchronous rep and all the sort of things that go with it for a portion of their applications. The way we deploy applications in the market in general is changing. So having a reliable mid-range platform actually could be very attractive for many customers.
Colin Gallagher:
Yeah. I think you hit it on the head, Chris, that a lot of those… some of those advanced capabilities may not move down market well. But things like performance, availability, particularly as we move to a much more all-digital business model and certainly most of us are living in all-digital experience right now, those types of things become more important, right? How do you deliver the best class digital experience? And you do that through low latency, high performance, uptime, et cetera. Those are things that used to be 10, 15 years ago considered ultra high-end capabilities and now there’s tremendous demand for six [inaudible 00:28:10] or more architectures in the mid-range because this is what people are running their business on.
Chris Evans:
Yeah. I think that’s going to be inevitable, but it’s interesting to see how that’s evolving and it will be a challenge I think for all manufacturers to try and do somethings that fits this part of the market, because I think it’s going to become a very important part of the market for a lot of vendors like yourselves.
Chris Evans:
Let’s talk a little bit about management. You’ve touched on this quite a bit all the way through the discussion so far, but I just want to go on and sort of reemphasize I guess, Colin, the idea of ops center and the technologies that sit around that because I think it’s important for people just to understand the work that’s gone in at that side of the development of your solutions to make sure that all the pieces you need to support the platform are there, too.
Colin Gallagher:
Yeah, and we take a little bit of different approach to management than some other vendors in the industry. We look at it sort of more as a holistic approach. How do you manage, orchestrate, automate, troubleshoot across the entire data path as opposed to looking at just one particular component like storage? And yes, we’re talking a lot about Hitachi Ops Center as part of this announcement. We have a new version of Hitachi Ops Center 10.1 that supports the VSP E990 as well as delivers some additional capabilities.
Colin Gallagher:
But one of the reasons we’re talking about Ops Center is when we introduced Ops Center alongside the VSP 5000 series, there was this misperception that Ops Center and the AIOps capabilities that provided were only really needed by large scale enterprises. And we firmly believe that AIOps really benefits companies of all sizes, that the days in which you can manage all of your data storage infrastructure just through manual capabilities is fast waning.
Colin Gallagher:
I mean, we have customers who are managing a petabyte or more of storage with just one person. And you can’t do that through all manual process or even with just scripts. You really need to leverage AI, ML and automation to do that at scale.
Chris Evans:
Yeah, I agree. I think the interesting thing that’s happened in the past perhaps though with a lot of management tools… And management tools has been a big challenge for every vendor, I would say. And that’s the fact that when you’re deploying that technology, you pretty much don’t want to become a slave to it. So having to install great suites of servers and software and upgrade it and manage it in order just to manage your storage platform becomes a real burden and very onerous. So you really want that software to be efficiently written, modular and you only need to put in the pieces you want and support the pieces you want and you need a good upgrade process. So I think that’s where I would hope that a lot of your development has gone into make that platform work effectively.
Colin Gallagher:
Yeah. I mean, Hitachi Ops Center is a suite of products that all work together. But if you just want to do simple management, we have Hitachi Ops Center Administrator to administer and manage the VSP platform. If you want to do data protection management, local and remote capabilities, copy data management, et cetera, we have our Hitachi Ops Center Protector, which is new to the suite with this release. We provide Hitachi Ops Center Automated for infrastructure automation. And then for telemetry analytics we have Hitachi Ops Center Analyzer that allows you to look across the entire data path, and by the way automated works across the entire data path as well to analyze and look for troubleshooting problems and do forecast for growth in the future as well.
Chris Evans:
I’ll look forward to having a separate discussion on ops center as part of our discussions in this announcement, but I think that’s probably a good point to move on from there. We’ll cover the details separately. Now, I’m going to be a bit slightly awkward and I’m going to ask you about pricing and you probably don’t want to talk about price specifically, but I would like to talk about something that we had in a discussion even just I think it was a couple of months ago.
Chris Evans:
We had Diane and Gary on the podcast to talk about storage as a service and we were trying to sort of get a handle on how that storage as a service type solution would be implemented. Obviously, you have that within the platform. Now, I see as part of this announcement, you now have something called EverFlex, which I guess is a brand around the offerings that you actually have.
Colin Gallagher:
Yeah, it’s a brand around the offerings that we had previously as well as some enhancements to them. As we talked to you and customers and others about consumption models, it became clear that organizations really need more choice and acquisition alternatives. We are living in uncertain times today, and even more uncertain than right now, and that they need better ways to manage that uncertainty and we need to talk more about it and make it more visible with them to let them know that Hitachi Vantara has ways to support them, no matter how they want to consume, to really reduce that uncertainty from their purchase decisions.
Colin Gallagher:
So today to that point, we’re introducing EverFlex from Hitachi Vantara. It’s a complete comprehensive set of consumption options that goes all the way from purchase to lease to cloud-like paper use in two flavors. As you’ve mentioned, we have a NAS surface option, but we also have a utility option as well.
Chris Evans:
Okay. I saw a slide from you and I thought it was great because it really sort of summarized the four options you have if you’re a customer, from purchase all the way through to as a service. And what I liked about that was it showed things like whose responsibility it would be for certain components and who would do things like management. I think having something like that really does help it come across really clearly to a customer to see what their options are. So it’ll be really interesting to see how that develops and what sort of take-up you have from that. Because I would imagine a lot of customers are very interested in definitely the utility model and possibly even more so the as a service model as they work out how to sort of build their own clouds, if you like.
Colin Gallagher:
Yeah, totally. And I think the key difference between the utility and the as the service model is in the utility model, you’re buying a preset set of infrastructure that you’ve specified such as an E990 or a VSP 5000, et cetera, and you’re paying for that on a monthly basis according to how much you use.
Colin Gallagher:
With the as a service model, you’re actually buying an SLA and paying for that on a monthly and usage model. So we deliver whatever is needed in terms of infrastructure and capabilities to meet that SLA. So really, to quote just how you want to think about your infrastructure, [inaudible 00:34:03] the most speed and convenience with the as a service model, but that may be something that customers aren’t ready to adopt fully yet so they’re looking to move to a more utility-based approach where they’re still looking at buying or owning, quote unquote, discrete assets.
Chris Evans:
Well that was a great conversation, Colin. Really interesting to see all the detailed layer about the announcement. So if people who are listening want to go and find out more as part of the announcement, where should we point them to? Where can they head off to? And when is this platform available?
Colin Gallagher:
So the platform is available today. We have tons more technical information available at hitachivantara.com. I would recommend folks interested in learning more about the product to visit my technical block that’s available there as well, and that we cover many more of the technical improvements than we’re able to talk about here. There’s a bunch of availability improvements, cache management, even rebuild times, really nitty gritty geeky things, but we cover those all in the blog so you can get more information there as well.
Chris Evans:
Fantastic. As usual, in our show notes, I’ll point them to all of that stuff so they can find that out and they can just use our show notes as the guide to get them into that. I’m really interested to see some of that detail myself and go back and read it and write some followup blog stuff on that from my perspective. But for now, Colin, thank you for joining me and look forward to catching up with you next time.
Colin Gallagher:
Thanks, Chris. As always, a pleasure.
Related Podcasts & Blogs
- #123 – Hitachi Next 2019: VSP 5000
- #144 – Introduction to Storage as a Service
- Hitachi Vantara Introduces VSP 5000 Series Storage Arrays
Colin’s Bio
Colin is Vice President for Digital Infrastructure Product Marketing at Hitachi Vantara where he leads product marketing for storage systems, storage software, and converged/hyper-converged solutions. Over his 25-year career he has lead marketing and product management team teams at several major storage companies. Colin has a passion for telling compelling stories about technical products that help customers solve both business and personal pain – and he enjoys the challenge of telling them in creative ways. He holds a bachelor’s degree from Georgetown University and an MBA from Northeastern University. Colin tries to put as many miles on his bike as possible, “hangs out” on twitter as @worldc3, and is team Oxford comma.
Copyright (c) 2016-2020 Storage Unpacked. No reproduction or re-use without permission. Podcast episode #48mo.
Podcast: Play in new window | Download