 I'm here at the Lonaro Connect Budapest 2017 and who are you? Hi, I'm Andrea Gallo, I'm managing the segment groups at Lonaro. You're managing the segment groups, what does that mean? It means that my role, the title is VP of the segment groups and the role is make sure that the groups are delivering value to the members in each group. So this is the LEG, LNG, LMG, LHG and light. So the groups working on servers, networking, mobile, digital home and of course the internet of things. So you're managing all the groups? So I'm managing the groups. So this means that my role is to make sure that the directors of each group have all the means to work and are delivering value to the members. At the same time, I have a background role in making sure that the engineering teams and the directors and the project managers, they work in an aligned way across them all and so I have a role on operational side making sure that they work in an aligned way according to the way we work in the process that we have at Lonaro. The reporting mechanism, the measurements, the KPIs and I also have another role that is breaking the silence. When you have segment groups with different members and different product perspectives it is too easy to end up in silence. So servers versus mobile or servers versus networking it's too easy to look into your own silence or mobile and digital home. It's easy to end up in your own silence and my role is to break the silence to make sure that there's no waste of effort, no duplication of activity that we are using the resources and the funding from our members at best. So if there's any overlap, it's my duty to go there and remove the overlap and make sure that the overlap turns into synergy. So you're optimizing just like an arm chip, right? You're optimizing the performance and the power consumption which is also money, right? And make everybody happy and things better. Well, I would say it's a lot of background work, lots of... There are phases when I feel that I'm having an impact when I'm helping aligned priorities and the vision of different groups. So that is where I feel that I have an impact. These are the phases that I enjoy the most. There are other phases of my working days or weeks or months. There are other phases where I'm more in a monitoring and reporting phase rather than having a direct impact. Well, those are the... Well, my boss is not hearing but those are more the boring phases when I'm overseeing and the directors are doing such a great job that I don't need to have a direct impact in those days. You need to have a great overview of everything, kind of, right? Yes, so... You've been working with the arm stuff. You've been doing stuff with arm for a while, right? Arm? My very first interaction with arm is even before arm limited existed. When I was a kid, I was writing assembly, small games and those listings that folks who now start having some gray hair may remember is when we were typing listings from magazines. There was no internet. We were just coping from magazines. So I was initially coping and playing games written in assembly for the Sinclair Spectrum. And then I learned from that and I started writing my own games. So I became a contributor to one of these magazines. You were a game developer? No, I would not say that. I had the ambition, but maybe I had not reached that, but I was writing a lot of software even in assembly, even overwriting the ROM codes and extending the basic interpreter of Spectrum. That was the Z80 on Spectrum. And I was writing also routines for graphics. By the way, this week we had a bird of a feather on graphics, embedded graphics. I loved that. I loved preparing for that both, because it just reminded me of when I was 12, 13 years old. So they're talking about graphics on Cortex-M? Cortex-M, which is the same kind of graphics that were back then. It reminded me of what I was playing with when I was a kid. And then through that work, in 1986 there was Acorn in Cambridge and it got acquired by Olivetti. And Olivetti was a PC manufacturer in Italy. Are you coming from that? So when I was a kid, I was 17 years old and I had already been writing articles in assembly, in basic, every month for a magazine in Italy. I applied for some job and Olivetti introduced me to Acorn. So I was so lucky and honoured to be a summer job student in Cambridge in 1987 and then again in 1988. And 1987 is when Acorn introduced and announced to the world the very first Archimedes. It was designed on the very first armchip. It was the Arm2 chip. The Arm1 was a prototype. I think it started in 1984, 1985. And the Arm2 was the first one that went into production. So in 1987 I was there in Cambridge during summer time as a student, as a trainee. Like a good world summer of code? Kind of. Those days there was this wonderful desktop machine, the Archimedes. And the BBC Micro was the most deployed in the UK in schools. And moving to the Archimedes was a big, big change. So as a student, as a trainee, as a student with the ambition of becoming a software engineer, I had the task of testing 10s and 10s of educational software from the BBC on the Archimedes identifying if they would run natively, if they required an emulator, if they were called to be fixed. So that was the work that my task as a trainee in Acorn that was going to become Arm. And as part of that, I'll never forget, I had the honour of meeting Sophie, which you interviewed sometime ago, Sophie Wilson. I can remember exactly what I needed for some of those tests, the compatibility tests that I had to run on. But she volunteered to help me. So I had to go and see her in the secured area because I was working... Secured area. It was a secured area. It was at the old building. A secret R&D center. A secret R&D center. At Acorn in Cambridge, it was in Fulburn Road, and it was very different from the beautiful headquarters that are today. So anyway, I was in more in the public area. I had a badge access, but it was more in the public area. And to go and see Sophie, there was a different level of security, a different badge needed. So she came to the doors. Fingerprint scanner? No, no. Iris? No? No need? No, but she came to see me at the gate and we went to her office. And it was a very narrow but long office in a U-shape. And she had desks on the three sides, and she had five or seven Archimedes. And she was working on all at the same time. And I don't know if I needed a specific floppy drive or something that she could help me with. But as I was there, she gave me a demonstration of some of the advanced technology that she was working with. And that was 1987 or 1988, either one. But she demonstrated to me drag and drop with a mouse on the Archimedes, the ARM2 processor. There was the filer. There were two windows with a list of files and icons. And just with the mouse, she selected multiple files from one folder and moved drag and drop into another one. That was, to me, it was the world first. World first? To me, yes. To me, yes. To me, yes. You had two windows with file, with list. What today is common, right? Two windows coping from one into the other with drag and drop. And it was written in BBC Basic, 1987 or 1988. Now I'm not sure about the exact year. Let's walk over here. Let's go down here. Sorry. Yeah. So did she help you fix everything? And the other thing that she demonstrated was the mailbox application. There was this taskbar with the mail. The icon was the British red mailbox. And when there were incoming emails, the mailbox was getting fatter because there were more mails, envelopes in the box. So that was fantastic. 1987. And you said you had some issues. Was she able to help you with those? Did she fix them? They were extremely minor. And frankly, I can remember what the exact issue is. I have a vivid memory. I can vivid memory of when I was in the office. And she demonstrated, oh, by the way, do you want to have a look at the new things we are working on and demonstrated these things? For that, I have a vivid memory. As if it was five minutes ago. I'll never forget that. But what the issue I had which was not working, sorry, that doesn't matter. Did you have a feeling that ARM was going to take over the world like this? Were you thinking, oh, I'm working on something that's crazy important? At the time it was educational, PC, computers for schools. So for sure important. But nobody in 1987 or 1988 was thinking that the ARM processor was going to rule the world and be everywhere in all embedded devices and all appliances. In 1987 or 1988, the ARM processor was mainly for the educational market. But is it luck or was the team just like crazy brilliant and fantastic and nobody else was as good? They were, Sophie was just visionary and amazingly brilliant for sure. Did something that was so revolutionary for the time and flexible. I think that the fact that then the educational market did not work out and then Acron went in that shape, I think that that was much more than blessing in disguise, much more. My feeling is that if Acron succeeded in schools, I don't know, maybe they would have never looked into other markets while by the fact that the educational market the nationally protected educational market without competition that was collapsing and Acron was getting in financial constraints. That pushed everybody at Acron to look at different opportunities and that's when they started looking into telecom, into the rise of consumer electronics and that's when they changed model and instead of designing the ARM chip for their own exclusive internal consumption they thought about this revolutionary business model that is licensing. I think that that is where things changed because the educational market was not successful. Since that time, were you always working with ARM the whole time? Since 1986, 1977? No, first I had to go back and do all my university studies. I was in Rome, I studied engineering, teleco-engineering and after my graduation I was lucky to go to France to stay microelectronics in France for a few months as a stagère. And after a few months I got hired. So in the end I spent four years in France, in a stay micro in France working on teleco, part of the smartphones at the time they were not smartphones, they were mobile phones and teleco in terms of the eight dial-up models. Dial-up models, we did a lot of work on optimizing software hardware together. Then we started working on ADSL. So I moved from France to Italy to Milan, ST microelectronics in Milan and it was ADSL and we did a lot of great work, great fun. So you've been involved in lots of stuff? So it was on the boundary between teleco and phones and set-to-boxes at the time and then smartphones, application processors. So we have been through the application processor efforts, Symbian. We looked at Linux of course, a lot starting from 2003. We started looking at Linux with the application processors and we worked a lot on trying to have products in MIMO and of course Android. So I did set up a team in Bangalore for ST micro and we achieved the first Android port for ST microelectronics application processor which had one of the first hardware video decoding accelerations in Android. That was, I can't remember anymore, maybe it was 2004. So and then in 2010, let's just move. Far ahead, 2010, I was with ST Ericsson at the very first meeting at the Mobile Congress February 2010 when Rob Coombs and Dave Rusling came to see the execs at ST microelectronics at ST Ericsson about this consortium to foster Linux on ARM. And I was pulled into this meeting, please come, come here. Come and listen. Come and listen and after the meeting we were saying, what is ARM thinking? What is this crazy idea of silicon vendors who are competing fiercely? How can ARM think that we are going to sit in the same room and collaborate? And that was February 2010. Linaro was founded in June 2010 officially and I joined Linaro as the technical representative from ST Ericsson ST micro in the Linaro technical board. And I figured out that their idea was really good and it was working really well. And this is all thanks to the technical leadership of Dave Rusling who has really made competing companies collaborate so well in the Linaro model. So you've always been involved with software and not the hardware chip design, right? What was it related? Back at ST I was on the boundary of the architecture to optimize software and hardware together. That can be ADSL modems, that can be smartphones. So the boundary is between the software algorithms and the hardware accelerators. So the boundary, how to have a hardware architecture that is not designed by the hardware folks. As it was in the old days of the ASICs, that was the most exciting part to me when I was at ST looking at how to provide input to the hardware folks to have the best hardware architecture that allows the software to run in the most optimized way. So do you have this feeling that this is making a big difference, right? Linaro is a big deal. I mean there's Linaro's contribution in everything in the world already, right? It's just there. Everything is Linux on ARM, no? It is and Linaro I think we are extremely successful. I know that sometimes prospective members or members in the renewal phase, they go into a very painful exercise of measuring the ROI. And I did it as well. When I was still at ST and ST had to renew in 2012, I did go through an ROI analysis trying to understand how much ST was paying, putting into Linaro how much it was getting and I did that exercise at the time and I came up with 3.5 or 3.6 times the return compared to what you were investing. But anyway... And it's hard to measure. It's hard to measure. Well, you can... It's a painful process because if you want to be thorough as an accountant, you have to go through all the work items that we log into JIRA and you need to look and have an estimate of the man-month effort and estimate if you had it done internally. But that's an accountant's work. Like an accountant doing the PNL certifying bookkeeping. But there's so much more in Linaro that is the strategic vision. It's the knowledge. Some members used to tell us that even if the return is just one, for them is a huge value because there's so much that they don't measure, they cannot measure, but that is related to having Linaro hire the maintainers. It influences the roadmap, I would guess. For all these companies, the roadmap would be totally different. It enables new technologies. So there's the boring ROI for the accountants. There's the knowledge, there's the maintainers. And there's even more that may not be measurable. Some members now come and tell us why should we renew with Linaro if we are working with SOCs whose software is upstream. We just go upstream and take the software. The only reply to them is, what you are stating is the best evidence that we have succeeded. Because years before, they were investing in us to get that code upstream. And those frameworks, our management frameworks, the DVFS on our... We could spend hours enlisting all the examples. But the fact that now OEMs can go on the internet and download a set of code that runs for some SOC, that's the best evidence that we have succeeded in using the money from these stock vendors at best. Because we have really accomplished what the primary mission is, improve the maturity of Linux on-arm and the availability of upstream. And it's not a one-off, because then there are new SOCs coming and improved frameworks, more complex frameworks. Look at everything that we are doing in all the segment groups. This is so complex and leading edge that we accomplish phases. And then there are new phases that are even more challenging. And so the renewals are all based on looking what we have accomplished but with a strategic view on what they will need and what we will accomplish together in the next years. I understand there is an accounting exercise but we need also this strategic vision and we need to look at what we have accomplished and what we will accomplish together in the coming months. And whenever a member is telling us, I have accomplished it, it's available upstream. I don't need you. I said, no, the fact that it's available upstream, it's the best evidence that we have used your contribution into Linaro at best. Because we have achieved that. If that code was not available upstream and you would still need to be with us for all products, that would be a lock-in. That would not be adding value, that would be locking you in. And this is not what we want. We want to prove that we continue adding value. So we complete a phase. The phase is completed. You may not need Linaro anymore for that phase and we will be successful in the new phase as well. But the industry is moving so fast. There's a huge consolidation. One company is buying another. Yes, and then there are new companies coming up. Things are moving fast. Of course, sometimes companies are making tablets and suddenly there's no more tablets. So it goes very fast. Yes, and then there's something else. There's servers. There's high-performance computing. There's on the networking side. Now we're looking at virtual CP, edge devices. There's IoT. There's gateways. It's huge. 5G is going to be on par. There's no way we can have 5G without on par. It's huge. Everything has to... The keynote by Björn Ekelund here at Konec was so fascinating. I invite you to go and please do post the link in this page, the link to that keynote, because that was so fascinating. I've known Björn for so many years and he's so good and he was so inspiring to the engineers because we have engineers working on RTOS, embedded, and Björn was inspiring them, telling them, look, this is how your work will be used. This is how you impact. And he was talking about things that are changing our society. It's disruptive changes for our society. He was there in the beginning of the NARO, right? Björn was instrumental when the hysterics joined NARO, correct. And so it's open-source software. Is it true or not that ARM is kind of like an open-source CPU thing? Because it's not really open-source, but kind of like... There is some parallel here, right? And the whole thing is just if you have lots of collaboration, lots of companies, everybody innovating in every corner that they want, the market grows much faster. It's an interesting thing that you're sharing. In a way, the fact that the business model that Silicon vendors can license the ARM core for small fee and add their own key differentiating IPs or even license the architecture and design their own implementation, it accelerates the rate of development and innovation. It sounds similar to NARO. So what are you doing? Well, then don't go up to open-source versus open hardware because of course the designing NSOC is so expensive that of course the ARM licenses cannot be free because there's so much investment behind it. So I think we should not go too far, but it's a fascinating part. But you're definitely making a difference in what's going on right now in the world. This is a big deal right now. We needed one trillion chips, right? One trillion ARM chips, and they're all going to run open-source software. Yes, absolutely. You're right. Cool.