 Thank you very much for joining everyone and yeah, it's it's great to see you guys Good to see you Morning evening and afternoon Exactly, that's what I wanted to say. So we're covering the entire world now Then expanding a bit and yeah, thank you very much for our folk for introducing our panel and and Basically this panel is a follow-up also on a panel that we had a few months ago at state of the source but the questions all always remain the same and Yeah, I would like to Yeah, invite you guys At first to share a bit like what is the current state? Of open source firmware and open hardware then from your perspective and as we had hadn't heard Like at our last panel from bunny and I would like to Yeah, have like an opening statement of bunny and bunny like how are things developing with the open source firmware and hardware and do you feel things are opening up entirely or like Yeah, is it maybe different? Well, I mean it's a constant tension of a lot of players in the market a lot of motivations, right? So I think the open source technology Continues to make really strong strides and I think there are people who are Listening to our message and taking it seriously And then there's people who I think have a motivation to keep it closed, right? But I think I think overall the the the general state of things is pretty good I'm seeing a lot of progress on a lot of fronts. I'm more on the hardware side. So I've been pretty pleased with the Sort of incredible Momentum that's developing on open source silicon actually so people are now Actually battling down and getting the actual transistor level stuff opened up on chips and the physical design kits are getting opened up And people are really excited in that area. So I'm looking forward to seeing a lot of You know action along that front which is sort of the final frontier You know in the on the hardware side of things we always argue about what's inside of our SOC's and we can't see What's on the inside and hopefully that problem will get solved in the next decade or so So I I think there's some control controversial Statement may be behind that but because things are opening up But I think we will talk a little bit later. How are they opening up and who are maybe the actors? Yeah, sure activity and yeah, so then thank you very much and I Would like then go to Jean Marie Verdun Jean Marie so you also had some news in the last few months where you also Had some experience. Could would you like to share what has been happening on your side? Yeah, since we spoke last time so we we've had some tremendous progress at supporting open source somewhere on parlor and family so which is the HP x86 server line, so we started to ship Servers to customers with open source Jamware capabilities Running open VMC and Linux boot So which means that we are getting closer and closer to support open source Jamware We we are doing that currently on private offering but our intent is to expand as much as we can this this product portfolio and We do support systems running in that chips and the MD chips so which which was also critical to us and We open up an integrated continuous integration platform Which is an online tool that anybody around the world can use to validate open source Jamware on parlor We we faced a couple of challenges that we try to address on a daily basis and that's also why this is not Available on a worldwide basis So we need to re-adapt or support process and and enhance the hardware Just to be able to switch back and forward between Proprietary firmware and open source firmware So we do not want to have a black and white world where a customer is locked with open source firmware or locked with proprietary firmware we just want to be able to offer the the widest choice and And this is where it is becoming tricky and and the past to success might be taking Let's say a couple of months is or a couple of years additional years of work Just to be sure that everything is set up properly But we're quite happy about what happened despite the pandemic which slowed down a lot of our defaults during the past few months is And Jean Marie because you're also like talking from the perspective of we right, but so you have been active on on different France let's say so when you're talking about we like which Groups to you for example feel you can represent and like how how do the different perspective? Right now when I'm saying we this is mostly the HP engineering team so and the people who are Designing open source firmware at HPE and that also the open source community is from from open BMC. So we We officially Committed code to the open BMC stream to support the HPE GXP asic so HPE is to designing its own BMC chip and with the security stack associated to it and we upstream that It's now officially available into the the public GitHub repo from open BMC We do have also introduced some support for the DL 360 within the Linux boot repo We we still have a lot of work ahead, but I'm pretty happy about The shift that we are currently operating regarding this We are still facing as HPE some challenges around the support of PCI express devices. So as you may be aware that there is firmware everywhere now and Most of the communities are focused around the BMC and the system BIOS but what about the The nigg drivers and the nigg firmware or the storage controllers and all these kind of devices Which are running proprietary firmware which are getting more and more complex and I must admit in some cases or not Yeah, they're pretty buggy. So and it's pretty tough just to try to fix them And and we still need to rely on our suppliers just to be able to fix that Okay, wow quite quite a lot of news quite a lot of Details and so I understand you're also working a lot from the industry side and Yeah, before I come to run then I would like to Ask Daniel. So Daniel and we see like a lot of progress has been going on on the industry side but you are also more on the like let's say The developer side engaging with like different people also in the community a lot and so what is your impression? Yes, I have I would say Quite diverse view on things. So I'm not just involved in communities, but I'm also tracking lots of different communities and One of them is at large researchers. So You might have seen during the last years now In the news there was every now and then something like data breaches There was progress in attacking hardware on Different layers and different ways, right? So basically when people discovered. Hey You're Ram. That's a clock that 2.4 gigahertz. No, that's also true for Wi-Fi So can we do something to transmit data over that? That's one thing I'm looking at and on the other hand We have such a huge market, especially for consumers Which is a very heterogeneous market. So that's both very very hard for vendors, right? They would need to support lots of different devices which is really kind of the toughest job you can do and On the other hand for consumers, it's very very hard to navigate through that market and you know understand what you can buy So if you want to have a contemporary computer and you want to assemble it yourself Just finding parts that match together is already quite hard and Then still comes the software side, right? And that starts with a firmware and you want to install an operating system to your liking's and And yeah, sometimes you might run into surprises like missing drivers Or maybe you cannot boot at all because the hardware and firmware is actually locked down in some ways So yeah, I've been diving through this and trying to find devices that we can actually repurpose or You know just use a developer development platforms for people who are not working in one of these large companies Providing us with the devices so that we can become a bit more independent And I'm seeing that in some very specialized communities as well There is for example one which is especially looking at cameras the so-called IP cameras Which are attached to networks Very often the internet already and there are actually just a very few vendors behind them But some people figured out what platforms they are running on and they're now building open firmware for it and That's the open IPC project if you've heard about that one Which I find very very interesting and just by coincidence I actually discovered that because I you know just found some cheap cameras. I wanted to look into them So I took a little tool called you route put it onto them And run it successfully and I thought well other people might have tried that as well And it turns out there were lots and I assume that's also true for many many other devices that you can find nowadays Or already have even that actually a body maybe thank you and Yeah, and that was like a perspective I think from the community side and Ron you are based in this so-called Silicon Valley as well and so Yeah, how and have you seen like changes and in the last few months and then yeah I want to come back a bit to what bunny said already Maybe you already have a comment that we see some changes with big actors getting in So I think we could like for example mention that yeah There's some support for example by Google for Like developers who are interested to create chips for example and do you have an opinion on on on that and how things Developing recently from your point of view. So I've been I've been thinking as I've been listening these To to the other three panelists because I learned some things too So there there are two Trends I can see Where people are trying to open up firmware to the extent possible on systems like x86 and so Well, we've done at Google We we've taken Linux boot which means Linux in your firmware and we use you root which Daniel mentioned which is go programs written and go that kind of replace standard Unix commands and embedded Linux boot and you root in firmware and That is that is essentially integrated into UEFI which is the proprietary for files we can't get rid of yet and Google is deployed that at scale In our data centers and I'm not allowed to say what at scale means but people have a rough idea how many computers Google has and Should be able to maybe guess what you know that turns into into the number of computers in our data centers That today now run Linux boot and you root and that's worked out really really well, so What we're doing is part of that project which originally we called nerf and now I forget what we call it We are peeling away All as much as we can of UEFI we're bit by bit just pulling it out of there and Replacing its function with the Linux kernel and go programs and my goal is to drive that as close to zero UEFI as I can that's kind of tough and We know that on many systems. Let's just say those mail from Intel AMD Let's just say that's going to be a long long thing and I don't ever even expect to get there Context here 20 years ago. You can have a full GPO Software I mean firmware stack on an x86 from AMD or Intel and you can no longer do that today So there's context So there's another effort called orbit and that's firmware and rust and We have a number of targets Daniels actually involved in these the first target is in AMD Rome And on AMD Rome because the memory turn on is done in a little arm processor You can imagine having the x86 running completely again GPL firmware written in rust now, so you know We're moving now toward rust because of all the good arguments for rust and we're trying to get a hundred percent GPL code on the at least the x86 CPU Longer term. I don't think the x86 will ever again be as open as it was 20 years ago It's just I don't think that's going to happen. And so people who really want a completely You know free and open source software stack need to look at things like the wrist five and Possibly the arm I but I think the arm Possibly could be more Problematical because a lot of the people who create arm so sees and sell them come with blobs And there's there's no way around that and my belief is if you really want to trust the system You're on you're gonna want that to be you know risk five is an open source by the way risk five is unlicensed So you have to take an extra step with our first five and go to a company that offers it You know open source if you really want that so if you want to complete open source stack all the way from the Transitions up like bunny has been pushing which I think is fantastic It's got a half year risk five because it doesn't come with legal incumbents And then it ought to be I think a GPL thing written in rust because why would you write and see in the third millennium? Anyway, right? You're crazy if you're still reading and see in my opinion and then we go up from there so that's that's kind of my picture of of where we should be going and You know, but we have to bow to the realities of the world Which are that we've all got a gazillion x86 systems and that's gonna kind of change for a long time And those systems are never going to be open again the way they were 20 years ago So you have to accommodate yourself the binary blobs in that world. So it's kind of a nuanced picture, right? Removing away. We'd like to move away from systems that require binary blobs We'd like to move away from see we'd like to move toward not only unlicensed Architectures or architectures free of that kind of licensing. We'd like to be on open source Processors, so that's the work of you know a couple of years coming up, but we can kind of see all the big pieces That that we need to implement to make that happen Thank you very much Ron and You are raising a lot of questions and I always like I also have this tendency to To often say we and then like people always say, who are you talking about? Yeah, and I say yeah Well, because every time I talk about is another we sometimes it's from project perspective Sometimes it's with other team because actually it's very truthful that we are not Operating like in a small like console where we just like work on our will our own, right? Like we are always embedded in teams in projects in societies in our company So I find both truthful and I don't know some people Criticize it, but like I think it's it's it's good, but sometimes need to be defined Who am I talking or where or from what perspective am I talking currently and I see bunny is smiling already? I want to go in a moment to bunny to because like I saw like he's nodding and smiling and Yeah, already speaking with this face And this okay, this is what I love about Like this interaction and video is what we have but I so I see like a number of questions and like their technical questions and their questions in regard to Society and in my view like and we have big companies. They're putting pouring in a lot of money And but there are a lot of questions involved a question that we also have with open source free software today Where we have companies? dominating some parts of this ecosystem and and we have the question of Like one company that hasn't been named yet here, but when you talk about arm You we also of course Think of Google and of Apple now, right? So Apple brought all this chip and one and In this panel everyone said oh, yeah We're going in the right direction things are opening up GPL Free and open programming language and and so on and Before I want to go back before I go back to to bunny and run Is there any comment that you would like to to make it regards to Apple? now Okay, so I think that's also a comment and And so then then bunny what what do you think so far from listening to this panel and listening to run Listen to Jean-Marie and Daniel Do you have any anything to share? me I mean so, you know when I made my my Initial statement. I'd said that they're actors that have incentives to keep things closed despite things Moving in the right direction. I think you know I'll say I think Apple is one of the people who Like to keep things closed. They want to keep people in the ecosystem They want you to go to the genius bar. They want you to buy things for the app store They want to make a nickel every time you play a nickel every time you work You know, that's kind of that's their business model How do you become you know one of the most valuable companies on earth and keep your position when you've kind of run out of ground to Keep running on you got to create a profit line somewhere, right? So there's their incentives that run against Sort of the kind you know creating a competitive landscape and open landscape, right? I mean, I think I think openness is about choice, but it's also about competition and and there's a certain Non-profitability in that I think you know from not having a monopoly position So, you know all that being said that they know Apple's just behaving as expected. What is a public company, right? That's run by what formerly the COO of the company right now Now, you know, you know, they're an operations driven Organization trying to trying to do what they what they all do, right? I think there's I have a friend who's working, you know on on the porting Linux to the to the Apple The Asahi Linux project. I think it's pretty interesting to watch it coming along and and you know sort of Read his exploits and the crazy sort of like WTF moments He has like realizing that the kinds of things that they did in their boot process and they're you know And there's and there's sock and whatnot. I mean, it's you know, but I mean the the other thing is is like I Feel I feel like there's also a bit of you know, if you know I have a thesis that when you when you sort of peel back, you know the number of reasons why things stay closed It's not just also Sort of the old big corporate profit-driven thing there's also just a lot of embarrassing mistakes inside hardware and it's really hard to as Organization that has to support customers and have a reputation for quality hardware It's hard to admit they have a lot of them in your hardware Right, if you actually see the list of Iran that come on like really well documented ships You know if you printed it's a sheath of documentation this thick like thousands and thousands of tiny bugs Things you're working around like a chip isn't So much to you know design from spec. It's you get a chip backing to discover what the spec is after validation You're like, okay, how fast is the Nick run? I don't know. Let's benchmark it. Okay That's that's the number we put in the marketing thing, right? You know, we we wanted to X that performance, but we actually ship that right and that's just how it is because you're not gonna Spend two million dollars on a mass set to fix a tiny bug in an arbitration issue somewhere deep in the XC bus or something like This is just not gonna happen so so so there's also a bit of I Guess you've an engineering hubris, you know around sort of not wanting to admit that we have mistakes And that's I think that almost falls a little bit on the community Expectations we place on ourselves and our inability sort to sort of admit that we have problems But you know, we're getting better about I think as people get more educated About the reality of hardware and the reality of these situations, you know customers are becoming a little more sophisticated They want to sort of see that Texture their granularity the sort of all those interesting things are coming along. So that's that's the hopeful side I would say of of of things going the right direction But like part of the reason we're never gonna see Like you if I and that sort of stuff ever go away is just just so many You know the vendors don't want to publish Why they have that that little sort of bit of firmware is because actually if they didn't publish You know have it there the chip would burn itself out in three seconds because of a problem in the PLL somewhere This is just too embarrassing to you know the stock value of several companies would plummet immediately if they knew that You know if you just did the right bits on boot or something like that You could melt down a half a billion dollars worth of servers or something like that It's just not something they want people to know Wow, that is quite a scary idea that you are putting out here and I'm yeah and I wonder like How much people are really getting into this topic that? Actually and the possibility of hacking these things. It's is it growing or not and I don't know so you're mentioning expectations and community so and a Lot of people say like that the the open hardware community is still relatively small and of course also if you compare it to the like software community of people who really focused on the software layer and So let's talk about the open hardware and open firmware community and who would you count into this? What what's your idea about this community? Is it growing? Is it like a specific set of people? Could you describe it a little bit bunny? No me. Okay back to me. I mentioned it I mean, I don't know I don't I don't feel I don't feel any How do you say? Entitlement to be able to describe the community. It's a big 10 right whoever wants to come in You know come with us out on your screwdriver screwdriver and open mind and now you're part of the community and if you if you want to take yourself and close it, you're not part of the community anymore, right and You know so long so long as I can discover your stuff and I can fork it and I can use it Then I'm I'm a very like I guess traditional open source idealist in that in that sense, right I Think realistically speaking though people have to have day jobs and they have motivations and stuff So generally we get more community as more commercial interest starts to sponsor people to Make it their day job to be in the community. There's very few people who are lucky enough to just As a matter of their choice without any constraint just do what they want to do. That's pretty rare From from a community standpoint and so and so we have a lot of conflict You know with I know everyone's like get hubs on Microsoft. Oh my god Like it's terrible, you know Microsoft is taking over open source and then you know Google is doing this open Titan thing Oh my god, can we trust it? They're getting into like, you know, oh, you know open secure enclaves and stuff. I mean You know, it's a it's a it's a mixed back, right? You need finances to make things happen. You need resources to do it and nothing comes for free So there's a bit of like we have to as a community that the people in this is this is where you end up with you know people You talk about people who are more radicalized than others people who push harder for one thing or the other, right? It's you know, you have to have people who set An aggressive set of goalposts around the idealism type of thing and hold to and remind us where we come from So you don't lose our bearings But then you also need to you know throw gasoline on the fire and get things moving and so there's constant tension that always happens And in any of our communities on on this front It's between like you know progress and scale and motivations and these types of things and I tend to find it's You know as long as we keep the pretty big tent and we have open conversations like this about it Or Frank about we sort of play the game. We know what we're playing. I think we'll sort of keep moving through it I wanted I wanted to surprise you a little about this too About everyone here If you really want to talk about a company that's in the open source firmware game in a big way and in the right way You want to look at IBM? And that should surprise people, but if you look at the power nine all the source code you need to build their firmware has been out there for years and it actually includes Linux and actually I ported Linux and you root to the power nine and The degree the the amount of openness of that power 10 predates in a lot of ways the openness of the wrist five and I think they don't get The credit they deserve for for what they've accomplished there because it's actually this point points out This is a really big deal. You're not opening up your firmware. You open up all your hardware bugs to people when you release this stuff and you know Some very brilliant people at IBM. I know one, you know, gee, I just forgot her last name Elaine Someone remember her last name for me Yeah Palmer. She is just utterly fantastic. Just look her up She's worked very hard on how you do the security side of a system where all the firmware is open source and Lot of lessons and and I'm kind of sorry that you know, let's be blunt It's hard to afford a power nine system, right? You're probably gonna spend a hundred times, you know as to get a working power nine system as you would to get a wrist five board, you know, cheap so That I kind of regret that because IBM has done the right thing here from my point of view. So that 10 should include IBM believe it or not and Intent and open Titan should be in that tent. Actually. I actually Disclaimer on the on the board of low risk or and I've kind of had involved in an open Titan But I know the people and yep, they work for let's just be blunt people consider Google a giant heartless international appropriation I don't know if that's fair or not But the people I know doing open Titan are working very very hard to do the right thing Yeah, and you know, so so you got it you kind of got to look a little bit beyond You know, obviously you have to be a little bit You can't just trust but you can trust and not trust and verify but you have to look a little bit beyond just the surface appearance Oh that person said Google or an open Titan. I don't know if I can trust up and tighten. Yes. Okay. Maybe maybe not but The RTL some get up Yeah, I get nothing get all your RTL for open Titan see if you trust right anyway interrupted money. I'll stop No, no, no, that was great I mean actually like I I know a lot of guys in the open Titan project to and I actually your sentiment I think they're great and this despite sort of the reputation that big companies have I didn't mean to sort of Imply I was disparaging the project. I'm just saying that the generally there's a community sentiment of distrusting these things Yes, and and and you know, but that it doesn't mean that it's all about are all good We have to we have to sort of get into the nuance instead of just putting labels on things and and binning them But I one thing I want to add you said that you know car nines are expensive and said, you know, unfortunately, you know There's an observation I generally have is that projects that Don't subsidize the cost of making something open because open open is is it's actually There's sort of a anyone who's ever worked on Okay, let me let me reset the story here. So so people Often asked me like why is it that I can't get documentation for this leading? mobile phone So see Right like, you know, but I can get documentation for the IMX six or something like this from free skill It's really well documented open source chip, right? It turns it turns out that if you you know, I had a conversation the free skill engineer about they said it cost them by far more To develop this open software stack and document it than it does to make the chip Which is why their chips are always like a generation behind and a process too old But everything's in the main line of litics, right? The thing that actually gates that chip is is that is that and their chips are more expensive and all this sort of stuff It's because they didn't externalize the cost of them the source. It's actually built into that hardware And what what you find is that a lot of times a lot of these cheaper platforms, right? They're they're either subsidized by other things and so you they can't be open or What what it actually is more likely it happens on all these chips as a vendor will say, okay I've got this leading mobile phone sock They'll put the spec sheet out But it's also in the in itself because they don't want to admit they haven't tested all the features yet And then they have then they have a million unit order from somebody you start they start fleshy out They figure what it actually works and then finally they only write the drivers for the things that actually ended being used in the final phone So you'll actually find on the chip a whole bunch of dark silicon for things that didn't actually make it into the final product They didn't had to burn engineer time to support a real-time clock in this mode that no one ever bothered using That's too expensive to just write these these features to be future complete according to documentation, right? So so so what you find is that on these the fastest way to get a product out is actually to build the hardware Make sure it doesn't catch fire Throw it at someone who develops a product that can ship a million units of it It's just enough for the features working qualified for that specific product and leave the rest turned off Right, just don't even turn I don't even admit that you haven't at the end of the day It's and that's a very in because the problem if you want to open source a thing all of a sudden some random guy You know somewhere out in the middle of you know is gonna post in the form Like why is it that when I turn on the real-time clock the system catches fire, right? And and then it's a big big problem all of a sudden right you should even know that this thing even exists, right? So, you know, that's and then but you know when you say the power 9 or power 8 power 9 is expensive It's because I think in part is you know IBM is They're making you pay in the cost of the hardware for all that openness and all that documentation That's actually the price of an engineer built into the into the cost of a chip versus like shows us ford subsidized System and that's a really good point too because we fit this in the coordinate project since inception which is Somebody says man, I really care about it freeing up and source software. I really really really want Corbett on my machine Oh, and by the way, I don't want to pay any penalty and cost performance weight power consumption battery life So I want the high-end features that I associate with the latest bleeding-edge laptop Oh, and by the way, I want well at anytime you kind of come back and say you cannot have those things at the same time The they say okay, well the one thing I'm willing to throw away is freeing up the source software. Yeah, you know every time Yeah, it's true. It's kind of a it's kind of makes you sad, but look I do it, too This I'm you're gonna all kill me, but I'm sitting and talking to a Mac right now So there we are. Yeah. Well, I don't think we have many options Aside, I mean the other thing is I'm using like a Lenovo or something like is that you know, but that it made it better I Okay, so I thought you're using a novena laptop. Okay, so yeah, sure. Yeah, and the novena can't keep up with this video It would collapse Okay, so and I want to sing a lot of point that like I'm afraid people get a bit the wrong Idea that bunny acted like a little bit in the way that people really have to look for opportunities and find a way where Whereas I think the one person in the world who knows how to make business around open hardware It's bunny and we will hear later more about this in your keynote about precursor Which was a very successful crowdfunding campaign and but like in regards to the community and aspect I am Daniel and what would you like to add here? So we hear like different and angles and you know generally I think everyone is happy that things are opening up and but what is your perspective on? Yeah, on the engagement with companies Well That's also a bit of a mix back so You can't even talk about a company at large even so, you know, there are usually lots of lots of different departments in a single company and That's also true for companies who sell any sorts of technical products Let's say HP for example, right? They sell servers. They still they sell their business laptops They sell the ordinary laptops for regular people of which you will find hundreds Just somewhere. I don't know at Walmart or wherever you go They also have printers right it's it's a huge amount of different things and It's always a bit of I Would say first of all a bit of a matter of luck if you just approach a company and you ask them for some information You might just get it right away For example, what run was mentioning the orbit projects? Where we have a target currently which way we are focusing a bit on We were very very lucky in the choice that we made And we found a chip on on the board that we picked for development For which we couldn't find open documentation But it turned out that the company was very approachable who made that chip and they said yeah, sure here you can have that data sheet and Well, we asked again to be sure it can we also publish that so, you know, it's an open project and so on. Yes, of course Go ahead that happens the same time there are also Others where you might write an email you get a response like a week later That might be along the lines of yeah, we would support like custom development if you want to produce or buy lots of devices or chips or whatever and We're talking about numbers that ordinary people, you know Kind of calculate in their regular thinking, right? So it's not like your grocery list. It's more like, you know If you've been in a very large company in a somewhat higher position, that's the numbers. We're talking about here and if You know, we as a community of regular people. So that's the thing. There are different sorts of communities Mostly when we talk about communities in the open The impression is that we are talking about, you know, the no ones that are just walking around everywhere, right? So like you or me, you know, we just want to buy a laptop and we would like to have something which performs well enough Well, it turns out that actually nowadays you can take a laptop from 10 years ago. You can still use big blue button from it I'm doing that right now. This year is a Lenovo x2 30 Old enough, but it still works fine But there are also communities of people who would like to collaborate a bit more Which are actually employees at some of those larger companies and That's especially true for people who are working for example on the linux kernel So there are many people who are contributing to that kernel But you know, they are also their own community and they also have their own conferences and so on where they meet together and that's Kind of an interesting mixture so on the one hand side you have vendors who would actually Prefer to keep things closed a bit but they need to also Support the market. So, you know, they make blobs and header files You know, like there is this linux firmware repository where you know, you find a bunch of blobs and lots of headers for it but there are also some which are, you know, just as open and willing to collaborate and Then again, what bunny was just saying There is also this weird spot where you run into all of these hardware issues and Guess where you will find documentation on those issues in A large and well-known kernel, of course, so if you just look through the linux kernel source code you will find thousands of issues documented in there where there is code which You would never assume to fight there because you would say no, no, this must be wrong But the documentation then actually says well, this is a work around for that and that and so on and you know If things twist around in the right way It works. Otherwise, you know, it might catch fire Yeah Thank you, Daniel and I want to Bring this to Jomari so we have this topic of community of openness and and Yeah companies that have their own commercial interest and you are working for HPE and on your side You shipped your first systems with open source firmware recently they are available to private customers who are making a request and You have extended your CI to integrate Removed API as to ease the validation of end users built on HPE Pro Liant and and the big topic for you is that you are willing to expand security features So do you see that security features make a difference? And do you think that open source firmware and open hardware will really make a difference for customers? because of security features and What are your ideas? Okay, I think the complexity of a company like HPE is that we are selling systems on the worldwide basis and we see a split Regarding the customer demand between North America and Europe for example So in North America people do not want to take care of their security So they are trusting their suppliers and they want their suppliers to be liable for the security If you go to Europe So that's a totally different story they do not trust their suppliers and they want to take care about their security and And they are building up a security stack through a transparency. So which means that when you work at HPE So you are you are designing a systems that you try to ship worldwide With two different with two totally different target and that that is the challenge that we are facing currently Open source firmware currently fit better probably in Europe than in North America Because people can have a look to the to the code can adapt it and and there is far much more transparency By releasing the source code. I Am I am obssumistic for the future of open hardware For many different reasons The main reason why people might be thinking it is slow Regarding the adoption and the shift from companies like HPE is The first one that I just did explain the market is not the same everywhere. So it makes More difficult to take the right decision and then we have to keep to keep in mind that company like HPE have legacy So our engineering team designed a proprietary firmware for the past 20 years So you cannot shift all of these people in a day So this is like a big boat. So when you when you say we have to turn that so it's it's not going to happen super fast So this is not a speedboat. This is like a big long-haul turn and and we're working on this but from what I have seen during the past year a crazy amount of people inside a company where Becoming more and more interested about what's going on into the open source firmware communities and As I said, we started to come in code It could look like something super basic, but within a company like HPE it does require so crazy amount of process so we we need to get approval for many different different business units or IP attorneys or all this kind of stuff that people do not imagine when they are working into an open source community Which are happening into a big corporation like HPE, but all of that has happened during 2020 So which makes me optimistic for 2021 2022 and 2023 So we we are going to move forward I'm Believing that the market is going is going to stay split between North America and Europe I'm always surprised about that because I'm living in North America and I am a European Citizen, so this this is really weird for me just to have a look at how North American company react to open source firmware compared to what's going on in Europe But this is also super interesting because I'm discussing on a daily basis with engineers which are American people and I'm trying I'm trying to explain them if we do not do that that we are going to struggle in this European market and this is still super important for a company like HPE because we are doing 40% of our revenue in Europe and and that is critical just to be successful to follow well our customer demand here and so Jean-Marie I have a question to you because there wasn't very interesting talk of lime proven at foster recently and Lime is also form Like talking about HP right like so and he has this perspective where Until today or like today and all computers have used hard disks and at least two storage levels and and yeah The future might look different. He says the future might look More like the past again and where Yeah, this kind of difference. It's a could become obsolete and this kind of separation And a computer's permanent storage. It could be all right there in the processor's memory map And he says there could be that there is no need for disk controllers or file systems It could all just be run and so if we talk about the future And it could be imagined also like a future where the software and hardware stack could look different What's your idea? Oh That's uh It's not that easy to answer to that questions because um I know the past we are trying to follow but I don't know whether I can disclose publicly So that's the that's the complexity of that questions. Uh, what I think is uh short term the future is going to be more modular and uh, we are we are currently focused at just trying to Uh break down the servers just to offer more opportunities for innovation around how server works um, that is true that, um, that may be needs for, um Revisiting the where the storage of system works and um That's also why I'm pushing extremely hard. Uh, the PCI express card vendor or just to carefully review what they are doing because they are adding a ton of complexity into their Aziz and shellware And I don't think this is the right path to follow so and and there are also our suppliers and um, and they probably have to listen a little bit more to their customers and um, and to what we might be expecting from them just to enhance their product So my believing is um, when the system is becoming too complex It's probably not the right path to follow or at least just to uh, Stay independent from what's going on and being sure that you have a B plan I'm always happy when people tell me, uh, we are approaching difficult questions So this means we come Back to this later definitely in detail and I would like to Ask Ron because Ron you, uh, work on Harvey os You have all this background with the core boot and all boot and what's you take on this? I don't know if you have heard about this talk. So that basically the Yeah pass could be also, uh The future I remember years ago BOS for example, they had a huge kind of They had a very good marketing Yeah, and everyone was so excited But it never took took off and what do you see as the technology stack of the future? I you know, I I I don't think I'm smart enough to have the faintest idea, but The the one thing I have been noticing It takes me on nine months the last time I put a patch of the Linux kernel For that patch to come out the other end of the of the process of managing a patch and I I hope there's a feeling in the community that maybe Um Unending growth of the Linux kernel is not really the right future and it's time to start rethinking The architecture such that when I make a two-line change to something that manages memory technology device driver It doesn't take nine months for that to get through whatever processes exist and That we aren't on the right track when Our our path is to basically suck every new bit of functionality into the Linux kernel itself and So there is I wish it were two weeks from now. There's going to be some pretty neat news about plan nine and about two weeks at last Let's just say becoming free of the restraints now That's an obsolete piece of code as kernels go at sea But it's a good lesson maybe in how you construct your things so that They aren't so massive and monolithic and yet still might have power Um, I actually think if we look at open source processors Uh, you know, you kind of think of okay that news me about about a decade and a half in the performance And maybe I want to look at an operating system that doesn't quite consume So many resources as it uses the machine so um, I feel like I like The move and the work going on and open source processors and I think now we need to think about What how would the world look different? Yeah, if um, We have so much memory and it's persistent memory. How do we rethink doing file systems? I don't I don't think though that you can just say well We don't have file systems anymore because we have this persistent memory that lasts forever You still need some sort of structure. You still need a way to say You know, I have a view of the memory But it's not the view of the memory you have because my view goes back two weeks ago and your view is from yesterday so they're You know, you don't just simply you know Um throughout everything that's ever come with file system lessons from the last 50 years Just because you don't have a rotational media anymore. So I think this is a really complicated question and I um I I think this is actually a really good time in that way because There aren't any ideas that we shouldn't be trying and we might as well try them in an open source community and I don't think We need to worry about whether It's the way things are done today because open source All on its own is the way you break away from the way things are done today Especially in processors and the fact that will be resource constrained Because we'll have an open source processor that's slower and not as dense and not as capable As the bleeding edge of of the nice processors we get from intel amd. I think that's a good thing You know unix originally grew up out of a resource constrained machine You had to think a lot and the reward was for code. You didn't write not for code. You did write And we need to it This is a chance to kind of get back to that to where we've got to really think a little And and emitting code is not the virtue the virtue is showing how you don't need to write code And I I'd love it to to see people from this community kind of try and look at that model So how much code can get rid of it? How much code can you not write because you came up with a good abstraction and then plan nine? It's an old code base, but you know plan nine they really nailed the abstractions. They really got them right and You know, there's just lessons and lessons and lessons in there and my biggest regret I'll stop in a second but My biggest regret when I see a new operating system written in rust is that people implement The unix or the linux abi You know, if you're going to begin anew for god's sakes, don't don't begin anew with a 50 year old design You know, think think a little look at your abstractions. That's where I saw a unix got where it was They they they got the abstractions right And we've really forgotten that lesson Now i'm done Yeah, I really love the way you talk and it's it's something that's very difficult To say as a german it always sounds like a command look a little yeah like commanding people But actually, um, I totally understand where you are coming from and you you said in a very good and Understanding way. So thank you very much Ron and we are really getting to the end of this Panel again and I found it super interesting and and before we close I would like to make a final round Where we have a short outlook a short comments What we should take away from this panel and what you are taken away From this panel and I would like to start again with the first person on my screen Bunny and I think bunny always looks surprised when I address him, but bunny, please What would you like to take us and what would you like to take away yourself from this panel? All right, I uh It was interesting. I have I I have a lot to think about I guess Um, just hear people's different comments and different perspectives on it. Um I was it was actually really good to hear ron's comments about his opinions on sort of You know giving credit for reducing the complexity of things and rethinking things from a very low level um One of the things we're doing in precursors. We're trying to really run in a very minimal environment Then we get a lot of flak for not just running linux on it and and calling it a day Um, you know But uh, you know, you know, I'm gonna I'm going to try try established this year writing even like A file system layer that gets rid of the idea of files But just as a key value store or something like this for a number of reasons like just Reasons right just don't even try to argue with me as to why I should or should not I've had people try to discourage me from doing this to say why are you crazy just file systems are hard Don't do it like yeah, but you know You're like one of the few people who said that to me um But uh, but it's uh, it's um, it's interesting just to sort of I think I think one of the honestly just and this is at this mark a criticism myself just because the whole pandemic and everything I I haven't this is my this is my world like literally you're looking at where I sit You know 18 hours a day or something like that. I don't I haven't gone five kilometers in a year or something like that right so so honestly, I don't get my head out very much in Um, unfortunately the internet's being really bad about delivering me content because everything is now filtered and targeted towards me So it's hard to discover new ideas and to hear fresh opinions and this sort of thing So it's actually really nice to just You know just be put in a panel with a bunch of people who I otherwise wouldn't have had a chance to hear the perspective on and to get some You know some clarification get myself out of my little shell and um and and give me a little bit of things to think about So thanks guys Thank you very much and yeah going over to Daniel. What's the takeaway? I would actually like to uh open up another perspective right now So I've um I've been attending many many different events and conferences over the last years You know until covid started Uh, and at one of these um, I met someone who recommended me to read a book Uh, it's called culture and empire by pita hintons and In that book I was reading that you know technology over time was always Becoming something Um, which was more and more accessible to everyone You know after people who were just rich or large companies. Um Got their hands on them And that's what I'm seeing now On the hardware side as well So if you just think about it, you can already Make your own whatever stuff with uh 3d printers people are doing that a lot They're sharing all of their designs and swan their large platforms for it you can take programs like free cat or key cat And design your own PCBs just order them in I don't know if you're well trained a few minutes I guess Um and likewise with many many other different things and I've seen Some people also talking about designing their own chips now Of course, they will not start right away with a seven or five nanometer process, right? Um, but that's not the point the point is getting started with that and if The law is true and so far it holds on you know all layers Then this will be a thing in just a few years actually so That's another perspective on looking at risk five or arm more than x86 now But there is so much more and I am very very sure There will be another architecture coming up as well So that's my perspective on the future here Thank you very much And uh, Jean-Marie, um, what's the takeaway from your side? I think the good news first is that we are we are still speaking about About speaking about an hardware a year later So it means that there is still some interest for that and um, I do see that the interest is growing because um I am also a big contributor to the free cat project And during the past year we have seen our end user base just growing like crazy I don't know if it was the pandemic effect or if the open hardware communities are growing on that but We we are seeing more and more and users and more and more contributed to that project And um, and we have new ideas about how to expand this And um, I do envision that there is more and more people who are collaborating around the world And that is the positive impact of the pandemic. So we try to Uh, keep working with uh with others and the only one way was to just properly the internet Um, as I am always seeing inside hpe the world is not changing in a day But um, I am super optimistic that within the next couple of years We will see more and more of a great open hardware ideas and designs coming up and they will be, uh Implemented into a big cooperation. I'm optimistic regarding this Even if I know all the pain that we will have to go through just to make it a success, but But I do believe this is this is doable now Thank you very much and Ron The final words are on to you So I was thinking that I was at a DARPA US DARPA sponsored project in 1988 where The head of the computing division announced that uh, they were no longer going to fund any operating system work Unless it was Windows NT because Windows NT had barely won, right? So uh by the by the Mid to late 90s a lot of organizations had given up and were ready to surrender To Windows NT and we all know the story Linux had already won by the time DARPA thought that unix had lost Um in 2010 a lot of us in the open source community had just basically given up on core boot I'll be blunt a lot of us that decided we'd lost we have a lot of Big companies going after us and trying to kill us and we thought they'd succeeded Every chromebook sold today runs core boot. There are more chromebooks than there are macbooks today And and I think the lesson from all this I used to say it and I lost faith in it for a while And I now believe it again open source is always going to win And there are a lot of reasons for that one of the reasons is nobody is smarter than the rest of the world So I can have a smart person at a company that's doing a closed source thing They're not smarter than the rest of the world. So in the end open source wins because nobody's smarter than the rest of the world It's just the way it works So if you ever get discouraged don't open source always wins Thank you very much. This is a wonderful closing word. I see everyone is nodding and Yeah, I also want to add small Like a few small sentences here, which is like, um, we talked about linux a lot Greg core hartman. He will be here on tuesday giving a keynote. Um, I would like to invite you For questions or any comments that you would like to make it's exactly at the same time that we're ending now this session So in two days, um, yeah, it would be nice to have a real Like onsite event because we could go right over And talk to him and yeah at an online event. This is not possible in the same way, but like let's try So and the other thing is like that there are still a lot of topics that we didn't address and daniel already opened up New topic and I think we have like the huge topic of Analog computers that some people see as a huge business opportunity now even and for example in germany There's quite some investment in this area and that we could address maybe in a future Discussion as well and of course like I think like the topic of the technology stack That's definitely not finished and we probably have Yeah, we need more chances to go into detail here. So thank you very much for this Really interesting panel. Um, I appreciate that you all took the time at very different times I know for daniel. It was very early. Um bunny was in a good time zone. It's The experience for people in singapore who always like have to arrange on others And yeah, and run and uh, so thank you very much for making Time with this the challenging time. So um, ah, there's a nice picture Oh, yeah He was here I'm going to step up for just one second and get some water and coffee and I'll be ready to give the next talk So I'll be give me give me like 30 seconds. Of course. I'm an injury. Thanks everybody Thank you A couple minutes more. So thank you very much everyone and thank you for this wonderful panel And we are documenting this and we're releasing it would be releasing that it was also live on youtube So um, I say goodbye and I hope to see you soon again. Thank you everyone. Thank you Thanks a lot for having us take care