At the end of June I had the tremendous privilege of speaking at Storage Field Day 16 in Boston, on behalf of SNIA. It was both a thrill and a challenge to me, and I thought I would express some of my own personal thoughts about the experience for those who may have had fantasies (as I did) of one day standing in front of any Field Day event.
In case you aren’t familiar with the Field Day series of events, GestaltIT puts on a series of focused technology events that typically center around a particular area: networking, storage, data, mobility (wireless), and cloud in particular.
The idea is absolute genius. Take some of the brightest minds in the tech blogging sphere and sit them down in front of companies who have developed new and interesting technology. These could be established companies, such as Cisco or Dell/EMC or HPE or NetApp, or they could be new startups like Zerto, or INFINIDAT, or Big Switch, or Cohesity, etc.
The point is that these companies get a very precious gift – dedicated time – to ‘geek out’ about their technology. Unlike company pitches or product shills, the purpose behind these hour- or two-hour long presentations is to strip away the marketing aspect and dive straight into the tech.
The trick, of course, is to use the time wisely. When else do you get an opportunity to lay bare the cool technology that most people miss? Strip away the GUI, so to speak, and expose what is really interesting to geeks? It’s a rare moment that is truly valuable to companies who normally have to worry a lot about “keeping it high level.”
For the candidates, too, this is a treat. The people who are invited as delegates are independent analysts and bloggers, known for their attention in a particular field. They get the opportunity to ask questions directly to the people who develop the technology, dig deeper than they normally would allow, and call out companies when they would rather substitute obfuscation for explanation.
A year or so ago, in my role as one of the Board of Directors for SNIA, I suggested that SNIA begin using the Field Day events as a way to promote some of the interesting technical work that is going on in the organization.
It was a huge hit. The first time we did this, we had Mark Carlson, who is the Co-Chair for the SNIA Technical Council, to talk about hyperscaler storage. The second time around, we got Rob Peglar, who also sits on the Board with me, to present about Persistent Memory, and Richelle Ahlvers, who has forgotten more about storage management than I’ll ever learn, to talk about Swordfish.
All of these people are intellectual and industry powerhouses, and perfect choices to talk about technologies that SNIA is working on. You know the kind of people I’m talking about – they’re the kinds of people that become their own reference:
“Where did you hear that about Technology A?”
“Oh, Jimbo showed me a tech white paper.”
“Oh, okay, as long as it’s referenced. What about Technology B?”
“Richelle Ahlvers said so.”
“Ah, got it.”
When it came time to do this one, I was asked to present on behalf of the Ethernet Storage Forum, which focuses (despite the name) on networked storage. The new hot topic for networked storage is NVMe over Fabrics, and so my job was to talk about the “Fabrics” part of NVMe over Fabrics.
My big question at hand was – what could I actually talk about? Unlike the previous SNIA presentations, which had technical work going on, the ESF is primarily an educational group. The ESF is not responsible for establishing standards, nor is it responsible for any kind of technological development. So, we could talk about other people’s technologies, but we had nothing of our own to discuss.
At first, I was a pretty strong opponent to having the ESF do a session because I didn’t think we had anything to add to the conversation. Getting up and talking about “other people’s stuff” didn’t seem like a really good way to endear myself (or SNIA) to SFD16 delegates who are “all about the tech.”
However, then I started to realize. ESF was in a perfect position to answer the questions about the comparative approaches to how Fabrics work. With NVMe-oF, there are a lot of different options, each with some pretty significant learning curves if you aren’t already familiar with them: PCIe, Infiniband, RoCE, RoCE v2, iWARP, iSER, Fibre Channel and, coming soon, TCP/IP.
How do you choose? How do you know the right questions to ask in order to choose? That was the approach I decided to take. But I only had 45 minutes.
Suddenly I was afraid I had too much content to talk about.
“Make Their Brains Bleed”
I do a lot of presentations. They vary from high-level, market-oriented topics to extremely deep technical dives. I’ve spoken in front of all kinds of audiences, from highly-intelligent (but storage-ignorant) Cisco networking gurus to industry standards powerhouses.
Rarely have I ever felt the trepidation I felt when preparing for SFD16.
The source of my anxiety was, interestingly enough, coming from the right place, I think. I wasn’t concerned as much of looking like an idiot (I do that quite well even without SFD!), but rather I wanted to make sure that the delegates got their time’s worth.
What was making things difficult was the time necessary to backtrack and discuss what NVMe was, and how it worked, and NVMe-oF was, and how it worked. All of these things needed to be clear as day before I could get to the different Fabrics options. More on this in a second, actually.
I became obsessed trying to find some information about the differences between RoCE, RoCE v2, and iWARP as they related to NVMe. In general, and at the risk of being technical here, from an NVMe perspective there is no difference between them. NVMe binds into the RDMA semantics, and then it’s up to the transport mechanism to deliver it in the best way it can.
But that lent itself to risking promoting one form over another, and I was determined to remain technologically neutral – it’s not my job to promote a particular type of technology. Just explain the differences.
I was beginning to get concerned that I might not be able to answer questions that the delegates would ask that fell outside the scope of what I was allowing myself to talk about – on camera – and they would think I was being evasive.
I began to feel a bit stressed. The last thing I wanted was to be on the business end of a Field Day firing squad.
I confess, as of this writing, I have not watched one second of my SFD16 presentation. I have seen the stills of me speaking, and it’s enough to turn me off completely. I don’t like the way I look, and don’t like the way I sound. So, I’m going with my recollection of being in front of the camera and in front of the delegates, rather than anything that may or may not be on video.
(Note: I know I’ll have to watch it eventually and get over myself, but for now I’ve been completely happy not watching myself.)
When I stood up in front of the crowd to talk about SNIA as an organization, I felt my breathing suddenly constrict. I have given thousands of presentations, but could I suddenly be getting stage fright? No, this wasn’t stage fright. I was still calm.
It was an asthma attack.
I do have asthma, but it’s exercise-induced. And I hadn’t exactly been running a 5k before I stood up in front of the crowd, but nevertheless my focus shifted – as I was speaking – to try to get my breathing under control.
It was a surreal sensation. Here I was, trying to go through the standard SNIA presentation, emotionally excited to be presenting, and yet mentally split in my focus to get my breathing under control. This is one of the reasons why I don’t want to see the video (the SNIA background video was the first one to be recorded), because I don’t want to watch me have an asthma attack on camera.
At that point in time is when I began to get a little nervous. What if it got worse? What if this persisted throughout the entire presentation?
I felt my mind get a bit distracted from what I was had prepared to say. It was one of those moments when I was very happy that I had rehearsed a few times, because that practice kicked in and I think I was still coherent.
Fortunately, by the time I was to talk about the technology, I had managed to get my breathing under control and I was feeling more like myself, and really started to enjoy doing the presentation.
To a person, each delegate was awesome. They have their own chat behind the scenes in a special Slack channel, and I have no idea what they were talking about there, so it’s possible they were putting up a particularly good facade for my (and the camera’s) benefit.
But to me, they were fantastic. It turned out that the experience level was a bit broader than I had expected, with Luigi Danakos and Matt Leib admitting later on that much of this was new to them, while Ray Lucchesi and Howard Marks – the self-professed “greybeards” of storage – were prepared for the low-level questions straight off the bat.
Ray started asking some low-level questions immediately, which was great. What wasn’t so great was trying to slow him down so that the rest of the group would be able to catch up, and I began to be concerned that perhaps he was thinking I was being evasive.
I did, eventually, get to answering his questions as I had promised, but I wasn’t sure if it was too little too late. At a couple of points I wasn’t sure if he had checked out of the presentation because I wasn’t going fast enough (I honestly don’t know; I haven’t spoken with him about it). Once again I felt my pre-event concerns coming to the fore.
Howard, for his part, was uncharacteristically quiet during the presentation, and that only happens under two situations. The first is when he already understands a technology and how it works, and the second is when a speaker has been so opaque that he simply checks out. In this case, he admitted later on that there was nothing ‘new’ to him in my presentation, so he was politely allowing me to educate the remaining delegates in the room.
I have to confess I’m a bit disappointed that I couldn’t offer something new to Howard, as it was a ‘stretch-goal’ of mine. 🙂
The other delegates – and I’m terribly sorry that I didn’t get the chance to spend more time with each of them – were absolutely great. They asked the right questions, they remained engaged, and I hope that they managed to get something out of this. Being that this was the first presentation in the SFD16 roster, and that this was more educational than informative, there has (at the time of this writing) only been one blog written about the presentation (thank’s Matt! Though, there’s no period after my name 🙂 ), so it’s not unexpected that they would be focusing on other companies’ technology rather than mine.
Self-Criticism: Getting Lost in the Moment
All in all, I think I did okay with the presentation. I didn’t hit it out of the park, not by a long shot. To extend that metaphor, I think I got caught in a run-down between second- and third-base trying to stretch it into a triple, but made it safely back to second for a double.
A good speaker can manage to take the subject matter and make it interesting and complete in the time allotted. At one point when I started to realize I was running out of time, I asked the delegates if they would rather go deep on any one particular type of Fabric, or spend the time evenly on all of them.
That was where I got caught in the run-down of my own making. The whole point of this presentation was to focus on the differences in the Fabrics types, and here I was offering to abandon that very premise altogether.
At the time, I wanted to keep resonating with my audience – in this case the delegates – and try to help them with the understanding. I had gotten caught up in the questions about NVMe and how it worked, and had actually lost sight (brief though it may have been) of the raison d’etre of the preso.
Part of me thinks that this is what you have to do at times; you need to adapt to your audience and not bullishly go forward when they are obviously at a different place than where you anticipated them to be.
The other part of me feels like I should have done a better job with 1) time management, and 2) keeping a clearer image of the goal in mind. This was supposed to be a Fabrics conversation, not a NVMe conversation, and perhaps I spent too much time going over the basic mechanics.
And yet, in retrospect, I’m not entirely sure what I could have excised from the preamble. There are subtleties and nuances about how these things work that affect the different Fabrics types, and eventually (especially given Ray’s questions during the course of the session) I would have had to go over them. Lego pieces interlock in different ways, even though the basic system is consistent across the board. NVMe-oF bindings are the same way, and once you start getting into the choices available it’s not as simple as following a recipe.
I had not wanted to hop-scotch back and forth, which meant taking a somewhat pedantic and deliberate step-by-step approach where the knowledge was cumulative. I suppose that if that meant not being able to dive directly to the specific point without context, that’s something that has to give.
Having said all that, the only reason why I take such self-criticism to heart is because I care(d) about this so much. The Tech Field Day experience is a gift to me as well as the industry, and the service it provides is immeasurable.
I’m not engaging in hyperbole here. There is no other place where you can get companies to provide technical deep dives on their products and be challenged in real-time on their claims. Hitting it out of the park at a Field Day event can make you a legend; failing at a Field Day event can make you a pariah. Reputations are made and broken at one of the Field Days, and you best keep that in mind when presenting your crown jewels there.
I have harbored fantasies about presenting at SFD for a while now (granted, it was originally for some of the cool storage technology that I had developed for Cisco, but that didn’t pan out and never saw the light of day, so… oh well). I love talking about storage tech and bouncing ideas off of people, and yes, there’s a bit of an ego boost when you can get people ‘in the know’ to raise an eyebrow in surprise.
My advice to anyone who will be speaking at one of the Field Day events is simple: Up your game. I’m a good presenter – hell, sometimes I’m a great presenter – but examining my style pre- and post-presentation has shown me that there are some things that are still worth improving. I should have been able to be clearer and more concise while remaining on-topic, even in 45 minutes.
Even so, given that bit of auto-chastisement, I’ve gotten really good feedback so far. It appears people have gotten something good and instructive from the session, and that, to me, is a win.
You can see for yourself and judge, if you are so inclined.
Here is my presentation about the SNIA organization itself:
Here is my presentation about NVMe over Fabrics:
I want to take a quick moment and call out Stephen Foskett (founder of GestaltIT) and Ben Gage, who not only made my presentation logistics smooth as silk, but also warmly welcomed me to the event. They truly made my life easy and comfortable both before, during, and after the event. Also, for Michael Meleedy and Diane Marsili from SNIA, you made kept all the gruntwork off my plate for this. Thank you, everyone.