Today's leading minds talk Data Storage with Host Enrico Signoretti
Stephen Foskett is an active participant in the world of enterprise information technology, currently focusing on enterprise storage, server virtualization, networking, and cloud computing. He organizes the popular Tech Field Day event series for Gestalt IT and runs Foskett Services. A long-time voice in the storage industry, Stephen has authored numerous articles for industry publications and is a popular presenter at industry events. He can be found online at TechFieldDay.com, blog.FoskettS.net and on Twitter at @SFoskett.
Enrico Signoretti: Welcome everybody. This is Voices in Data Storage brought to you by GigaOm. I’m your host Enrico Signoretti and today we will talk about data storage, market, and technology trends as well as IT startups. And among other things, we also discuss what we have seen in 2018 and our expectations for this year.
My guest for this episode is Stephen Foskett, a longtime friend, founder of Gestalt IT and Tech Field Day. The market is changing quickly with the multi-cloud in every conversation, flash memory that are replacing mechanical drives for a larger number of workloads and applications, data management now on top of the priority list for everybody. Kubernetes is rocking the scene when it comes to container orchestration, new infrastructure paradigms like composability and yes, security and data governance with GDPR and similar regulation popping up almost everywhere. But also security aspects that we were underestimating a few years ago, and this is only a premise, and without any further ado let me start this by introducing my guest. Hi Stephen! How are you?
Stephen Foskett: Hello Enrico. Pretty good. We're here in Austin, Texas today for Tech Field Day.
Why great! And, I didn't introduce you properly at the beginning of the show. So, maybe because we’ve known each other since Storage Tech Field Day 3 and I always take for granted that everybody knows your event—maybe you can give us a little bit of an introduction about you, your career and what you do for a living?
Sure. Thanks Enrico. Yeah, I always assume that people don't know who I am, but I appreciate that you're assuming that they do. I started out as a technical person probably like yourself doing systems administration, and storage administration. I moved on to being a consultant and started doing writing and speaking at events, and before long I had a whole group of friends who were also doing the same thing, writing and speaking, blogging, tweeting, and those of us in the space got together as Gestalt IT back in 2006 to write about trends in the broader industry—because I realized I couldn't just be talking about storage, I had to also consider the other aspects of enterprise IT—that's kind of what the Gestalt refers to.
We started blogging together, and pretty soon we decided ‘Hey let's have an event, let's get together with some with some people who are like minded and learn from companies and so on’ and that's where Tech Field Day came from. A group of us were actually at a tech event from HP and said you know I bet we could do our own event. I said yeah bet I could pull that off, and I did and that was 2009. And since then that's become my job. So now Gestalt IT is the company and Tech Field Day is the event and we produced a whole bunch of these events on different topics. Now security is the newest one, but [also] cloud and networking and so on [and] of course storage.
Great. And we will meet again at the end of this month for the 18th edition of Storage Field Day.
Yeah, that's great. We're on number 18.
It's unbelievable the success that you had with this event. It's a great thing for the entire industry. So at the beginning of the episode, I mentioned a few market and technology trends that are somehow changing the storage field, flash data management, multi-cloud containers, new ways to design infrastructure, new regulations. There is a lot going on. What did you see in 2018 that was different from the previous years for example?
Yeah, that's one of the interesting things with Tech Field Day, and I know that you've probably noticed this as well is: we don't have a theme beyond just like storage, except there always seems to be a theme. You go to Tech Field Day. Let's say Storage Field Day 16. And most of the companies are talking about NVMe over fabrics—and somebody could be forgiven for thinking that that was the theme of the event (that we intentionally invited that) but no that's just what the industry is doing.
And so in 2018 that was really what we started seeing the first half of the year. It was all about MVMe and the MVMe over fabrics. You know the second half of the year we started to see that fragment a little bit. So, of course, we're still talking about MVMe. We're still talking about fabrics but we're talking about a little bit more mechanical: How is this going to work and what is this for? And that's what I see continuing into 2019 is companies trying to figure out ‘Is this sort of a ‘flash in the pan’ if you'll excuse the metaphor, or is this really the next generation of storage and specifically what is it for and what does it do for us?’ And that's I think the interesting question going into this year.
In my personal view, I saw a lot of vendors switching their cloaking and now they mention AI and machine learning everywhere especially. They claim that they are using these techniques to improve their products. Okay. I think most of it is marketing fluff. But I'm not even sure that this technology is ready for a broad audience. But anyhow, do you have an opinion about AI and machine learning in the storage industry for 2019?
Definitely. And this is what you say is true across the areas—that's the other interesting thing—because I get to see Mobility field and Security field day and Networking. You know, I hear the same messaging coming out from those companies that I hear from storage which is: ‘Hey AI, we've got AI, we've got ML. And I'm just going to put everybody on warning especially I know a lot of the product vendors listen to this podcast.
I'm going to put you on warning if you say AI and ML to me or to any of us tech field day people. Our first question is going to be “Is this really AI. and ML? Is this really machine learning? Are you really doing deep learning? And it's interesting—there is a nuance here in that some people, well some people are just saying ‘AI’ when it really just means like rules and stuff and that's not AI. Other people are really using machine learning databases but kind of in ‘read-only’ mode if you know what I mean, where they'll have learning that's happened at the home office and then they'll just apply those rules and then as information comes in, whether it's logging whether it's storage configuration or whatever. They'll just apply that to whatever data they see.
Like I said kind of in ‘read only’ mode.
And then there's kind of a third category who are actually doing ML on the machine. So they actually have some kind of ML accelerator. Typically they would, otherwise, it's gonna be awful slow. And they will actually be learning in real time from the machine. And I guess in some cases that could be done in the cloud or at the vendor’s side of things. But it runs the spectrum there from marketing fluff to yes, but passive to ‘hey that's actually ML.’
Wow, which is interesting that that demonstrates that you are still at the beginning of this journey with AI and ML actually. Another interesting aspect that we saw during 2018 was all the discussion around multi-cloud and or at least IT cloud. Okay. And I think this will remain a strong topic also for 2019. A concept that was also validated by Amazon with its AWS Outpost for example and others. So what do you think about it and how it is affecting the development of new storage products?
Yeah, I think this is one of those ‘where the rubber meets the road’ kind of things and I'm starting to see… Well, hybrid cloud has always been just like the most fluffy ridiculous useless marketing term.
But, I'm starting to see actual uses that sound like hybrid cloud to me, and one of the aspects of this that I think is important to consider: this came from some discussions over the last year especially you know at AWS re:Invent. We were meeting with some of the cloud people and they're pointing out that many of these hybrid cloud solutions where hybrid means ‘somewhere between there and here.’ Many of these hybrid cloud solutions are kind of ‘outside in’ meaning you’re cloud native and you're trying to pull that stuff in to the data center, whereas other stuff is inside out meaning your data center native and you're trying to bridge out to the cloud.
And once somebody told me that last year, I was like ‘oh, of course, this makes so much sense’ because then you start to look at things like the NetApp ONTAP cloud and you're like ‘Of course now I understand what the point of this thing is.’ This is for companies who have data on Net App and are trying to get out to AWS. Whereas if you look at like what Net App is doing in Azure, for example, that's almost more cloud native and allowing people to pull stuff into on-premises stuff.
And so the same can be true for example of Amazon Outposts. That's definitely an ‘outside in’ kind of thing like Amazon. You know you're using Amazon and you want some of it to be closer in. And so you put an outpost locally. It really opens it up, and it does seem to me that we're starting to get to the point now where cloud is not only a relevant architecture for enterprise applications—at least the cloud kind of applications, but we're starting to see some actual implementable solutions here. You know, like HCI.
Right, and those up multi-cloud also means that you have to step up in the data management field somehow because… at least from my point of view. Okay, multi-cloud or not, most of the organizations are not able to consolidate today's data in a few larger depositories. And data keeps growing anyway. So if you can’t contain it, at least let's try to understand what you have. What is its real value? The risks associated with storing it or deleting it and so on.
What's your point of view about data management, because everybody now talks about data managers. Even a data storage company that was not mentioning data management at all last year, now they have some data management message somewhere in their website.
Yeah, yeah. And that is pretty interesting isn't it, that data management is finally becoming… not ‘cool’ but at least relevant. A mutual friend of ours Chris Evans has been writing about this a lot in terms of data management and data protection and multi-cloud and he's got a lot of good points, but basically the summary is exactly what you said: that if your data is outside of traditional locations, then you really do need to ‘get your head around it.’
You need to start thinking more in terms of: ‘Where is this data? What is the risk of putting this data out there? What is the benefit of putting this data out there? If the data is out there on a different platform how am I going to integrate data management technologies and techniques from on-premises to these cloud platforms? Does that even work?’ And again you have to be very very careful because a lot of the solutions that claim cloud data protection, cloud data management… they're really not apples and apples.
So you look at what they're doing on premises and then you look at what they're doing in the cloud and either they're different things or they're the same thing and it's ignoring the nuances of the cloud. And these are all big, big problems. And so I do think that this is going to be an unsexy but necessary step for the industry to start figuring out: How are we going to do data management across platforms that are truly different?
Yeah exactly. And there is not only that. If we really had a data management at large, there are a lot of new nuances that were not in the rule on the storage administration and now with the rule of these people that is changing pretty quickly also because you can't have the same ratio of terabytes under management versus MB now that you had in 2000 for example. So you need to find mechanisms to manage data differently—data storage and data—differently than in the past.
Yeah. And you need to figure out what data means to you. That's been another theme of Storage Field Day, of Cloud Field Day. Especially with folks like Denny Cherry and Karen Lopez and actual data people getting involved. They start pointing out that when you say ‘data,’ you don't actually mean data. You still mean storage. You're just using the word, the D word for it. Because to a true data person, a data expert, it's something entirely different. And I think that's another challenge that the storage industry is facing and trying to tackle. What do we mean when we say ‘data management’?
Right. And let's take a look at the infrastructure. This year we see a lot of announcements around the QLC we’ll see that now is finally reaching the status of a product available to everybody. Memory cloud storage becoming more mature and with form factors that are more standard and usable. And there are protocols to access this thing like in the NVMe over TCP for example, which is slowly reaching this case its first commercial implementations. So, what do we expect to see in 2019 about these things hitting the market?
Well, I think that one of the interesting things that you mentioned there the QLC. Well, see it sounds like a nitty nitpicky little nonsense thing. Like why should I care about how many bits you're storing? I mean that's not nothing, but it actually is really important. So, again I'll call it attention to somebody else.
Jim Handy, the SSD guy wrote a blog post about this talking about the fact that in many ways SSDs are approaching price parity with hard disks, and that's going to change everything. Now approaching—we're not there but eventually, there will come a crossover point at various capacities and in various use cases where SSD is cheaper than hard drives. Jim's point was basically that's at lower capacities. We're not going to see SSD at the same dollar per gigabyte as hard drives at 20 terabytes. But you know as sizes get lower, we're going to start seeing it make more sense to use Flash, and if we're going to use Flash than it maybe makes more sense to use NVMe, and if we're going to use NVMe, then maybe our whole application has to change.
And so even little nitpicky things like the number of bits you're storing in a given amount of space, it can matter. And it's the same with some of these other other aspects. I mean one of the things that's held back NVMe is just sort of the esoteric nature of implementing it. It's hard and weird and suddenly we're starting to see efforts to standardize it—to standardize the interface. We're starting to see it more integrated into operating systems. And pretty soon maybe we'll start seeing this as a new a really a new class of storage above spinning disk, just like spinning disk emerged as a class of above tape.
And that brings us to storage class memory, which could be truly transformative if it actually happens right. The industry is working on that, but we'll see. So, if this stuff pans out, we could see true architectural change of a sort that we haven't seen since the 1980s. If it doesn't—and I mean fundamental architectural change—not just like all my applications running on virtualization. I mean like the entire system is designed differently.
But, we don't know that.
Yeah. And it's also interesting to note that these new products for example and these new technologies like large Flash drives are somehow enabling new ways to design the data center. And there are now many startups working on these new design paradigms like composability, and in case somebody doesn't know what is composability: practically that the data center can be configuring large resource pools and you can access this resource pool to configure your server nodes—what you need to do to run the workload. It could be done very quickly automated and then you can reconfigure everything for another workload.
So, somehow even if it's an oversimplification, we can think about it as a sort of virtualization applied at the output level which is really interesting. And so I know about a lot of startups that are working on this concept of IT composability. And we met a few of them in the past field days. So do you see them reaching and some sort of scaling velocity to exit the niche and reach a larger market this year or is this still a transition year [in] 2019?
I think they're still trying to find a market. As you mentioned, we see a lot of these companies with Field Day and so far, the composable space is not yet really even a small part of the enterprise infrastructure. But, it just makes so much sense. You know you look at the distributed and composable storage concept architecture and it just makes sense like what you're saying. I'd like to think of it as hardware containers basically. I need to run a thing and I'm going to reconfigure my hardware to give me—not a virtual machine, but a physical machine that has just the resources I need in it. And that's exciting!
And that is also crossed over with this concept of distributed because I think that's important as well. You're not talking about in most cases with composable infrastructure. You're also talking about distributed infrastructure and you're talking about distributed applications and you know you combine these things and you can basically build just mega kind of new mainframe things, but where is it? And right now it's nowhere.
So hopefully we'll see this sort of catch on. I think the thing that it's waiting for in the enterprise is they're waiting for one of the big vendors to get on board. And I think that if we saw an acquisition by a Dell or an HP, maybe Cisco in the composable space. I think that'll be very exciting and I think the one that's going to do it frankly is Dell because you know they've already shown with the Power Edge effects and Air Max and all that. They've already shown that this is maybe a direction they can take in the data center. And I would love to see them go ‘all in.’ I'd love to see a power edge composable. I'd love to see HP have like a pro client composable. And then I think it would and would not.
And what are the other areas you think are interesting in 2019? I know about if you start up funded by some very well-known guys in our industry but are still in start mode. But again, you have a very good observatory. So, I can mention many companies that choose your event for their first appearance for example. So maybe you can give us a few hints for 2019?
Well, it's funny, I actually bet that you might have more insight into the smaller startups because the smallest startups… sometimes they'll tell me, but I can't talk about it until they launch because that's my ‘gentlemen's agreement’ with them. I don't do NDAs but I do ‘friend DAs.’ And so I won't say what they're working on until it's time but indeed a lot of these companies do use Field Day as a way to ‘come out’—especially the ones that are more confident in their in their product. And we do in fact at the end of the month.
You know you're going to have a new company launching at a Storage Field Day if they're listed as a secret company right now. I can't tell you what they're doing, but I can say that if you look at some of the other new companies in the space… there's obviously a trend in the industry and it gets right to what you've just been talking about: composability distributed storage, combining resources and managing resources. I mean it's it's kind of like it's the ‘Borg’ of storage. People look at the Borg and you see a monolith but that's no monolith. That's a hive. And I think that that's the direction that the industry is going.
So we have to wait for Storage Field Day 18 to get more about this.
February 27th. You'll know all about this next company. And of course we've also got some other exciting companies there, and it's it's fun too to see not just the newest companies, but also to see how the companies are growing and adapting and changing and learning.
Yeah, you find the list is really interesting. So there are companies like NetApp and Western Digital we do you know beat vendors and on the other side. You have the TARA and other startups that have an interesting product, but actually still [are] looking for a market for a product. So they are very interesting on paper and maybe they still need to define the right use cases for it.
Yeah, and yes so you mentioned the TARA. I think they're in the category of companies that I think have gotten their second wind and are figuring out exactly what it is that they're doing. You know they brought in a new team and they figured out where they're going. You've got Cohesity who clearly knows where they're going.
And then you've got these new guys: StorPool, WEC.io and the new secret company that are emerging with great new ideas. One of those—the StorPool—already has some establishment work has some establishment. But this new company it's all new. And so we'll see if they're able to transform themselves into a Cohesity, or if they're not.
Okay then. So the event is for the 27th of February and I'll be there. So thank you again for having me at the event. And now how many events like this do you organize during the year? Because we talked about the security cloud storage for the Tech Field Day, which is the original one. There are so many now.
Well, there's different topics, but in each topic, we only do it twice a year. So you look at the event calendar like we do internally and there's quite a few on the calendar. But if you look at it—as you know I'm Enrico—there's only two Storage field day events. You look at it from the perspective of the packet pushers guys, there's two networking events. So it's not as many as it maybe might seem if you if you look at the entire list. But, yes.
So we have storage, networking cloud, security mobility and then tech field day and tech fields and cloud field day are both the genre-bending events where you can have almost anyone. I mean we're at Tech Field Day this week and we've got quite a variety of delegates and quite a variety of company conversations here from VMware’s VC-er team to date, TERA to NetApp and SolarWinds. The same thing is true of Cloud Field Day.
So we're pretty excited that ‘Cloud field day’ is kind of you know Tech Field Day: the next generation. And so you may see some familiar companies. You'll probably see some new companies there that wouldn't present at Tech Field Day, but even the familiar companies—you've got Rubric and Cohesity for example. What they're presenting at Cloud field day is totally different than what they would present at Storage field day, because the delegates are different and the audience is different. And so the delegates there are more ‘cloud first’ and want to know about API and integrations and the people at the Storage field day are more likely to ask “How does this thing work on the nuts and bolts level?” And so it's fun to see the same company have a different approach in different areas.
Okay. And and I'm sure that you get this question all the time, so how can somebody become a delegate for a Tech Field Day event?
Well, we would love to have you. We're always looking for new people. One of the things people say is “oh all the delegates they go work for the vendors and then they can't come to tech field day and that's going to ruin you.” It doesn't because over all these years what we do is we're finding new people and the new people are exciting and have new ideas. And so we are constantly looking for them. That's one of the things that I personally spend a lot of time doing -- looking on Twitter, reading their blogs. And also we have the ability for people to submit their name.
You know the easiest way is just to reach out to me. Find me wherever I am. I'm @Foskett on like every social media platform and just let me know that you're interested. If you go to the tech field day website techfieldday.com/delegates/becomefielddaydelegate, there's a menu you can click on it if you want to type the URL. It describes what we're looking for and has a little form that you can [use to] submit your name. And we do actually look at those forms and bring people in from those forms. It's a great way for us to kind of get introduced to new people. But forms are impersonal, [so] just drop me a Twitter message that's easiest at @Foskett.
How can people follow the event at home?
Thanks. Yeah, so the easiest thing to do is just go to a techfieldday.com or on Facebook. If you're a Facebook person go to facebook/techfieldday. YouTube/tech field day—the videos posted there, otherwise live streamed on the website as well as on Facebook. We post product produced them and post them to YouTube afterward, but that's probably the easiest way to follow along. Like I said, this Thursday and Friday we're going to be here with Tech Field Day and then the 27th through the 1st is Storage field day. In between they're actually on the next week is Networking field day, if you're interested in networking topics. That's February 13 through 15.
And you know just go to www.techfieldday.com. Click on the event, watch the livestream, tweet. That's another way that we find new delegates honestly is sometimes we'll see people active on Twitter asking good questions, making good comments. And I'll say “Hey that's somebody we should invite in the future.” Sometimes some people even write blog posts about the presentation that they saw and we'll notice that and say “oh wow this person is great!” and that's the most surefire way to get involved.
Fantastic. Unfortunately, we have to wrap up this episode but again can you give me your social media handles for Tech Field Day and your personal ones so that people can contact you?
Sure. I'm @Foskett on Twitter. That's the best place to find me. And tech field day, it’s just @TechFieldDay. And GestaltIT, it’s @GestaltIT. And you can go to TechFieldDay.com, GestaltIT.com. You've got a blog up at blog.fosketts.net. That's my personal blog. And I'd love to hear from you.
Great. Thank you Steve for joining me today, and see you soon at Storage Field Day.
See you in California.
If you enjoyed this episode of Voices in Data Storage, please check out the other ones. Unstructured Data Management is the focus of a report Enrico wrote for GigaOm research. To find out more about how data storage is evolving in the cloud era, download the single report or subscribe to GigaOm research for future forward advise on the other technologies, operations and business strategies.