 Firstly, I'd like to thank the steering committee for the invitation to come and speak to you today about a new research institute that has been recently set up in the UK, and that is the Research Institute in Secure Hardware and Embedded Systems, or RISE. So RISE is funded by the UK National Cyber Security Centre and the Engineering and Physical Sciences Research Council in the UK, and it's a multi-institutional research centre hosted at the Centre for Secure Information Technologies, or CSET, at Queen's University, Belfast, which is where I am based. So to give a little bit about the background on this research institute, it's one of four multi-institutional research institutes that have been set up in the areas of cybersecurity funded by NCSE, EPSRC, and the Centre for Protection of National Infrastructure, and this is all with the aim of developing the UK cybersecurity capability. So the other three institutes that have already been set up and have been going for some time are RISCs, which is looking at the area of the science of cybersecurity, and that is hosted at UCL. A second institute is investigating trustworthy industrial control systems, RITX, and that's hosted at Imperial, and a third in Rivets, which looks at verified trustworthy software systems, and again also hosted at Imperial. So this is the fourth such research institute that has been established, so set up just in November 2017, and the motivation really behind it is due to the sort of increasing demand for hardware security research innovation, and this is what the growing security needs in embedded devices, networking devices, and also cloud-based services. So key driver for all of this is obviously the internet of things, the predictions for the numbers of billions of connected devices continues to grow, I think the latest is about 30 billion in the next five years, but it's really going forward, it's important to look at addressing security requirements throughout a product's entire life cycle. So right the way from initial design through to its operational environment, and we're seeing this more and more. A multi-layered approach is needed, looking at establishing a trusted computing baseline that anchors trust in tamper-proof hardware, and really we're saying that it's becoming more and more evident that a strong hardware security foundation is absolutely essential for realizing security systems like IoT, and also being able to allow us to offer sort of hardware based security services. So with the volumes of these IoT devices it obviously necessitates the need for machine-to-machine communication, meaning that we'll no longer have direct control over with whom or what our devices are communicating, and the growing presence of these devices also enables new attack methods and attack services for criminals and hackers to exploit, and we've obviously seen numerous practical attacks of IoT devices in recent times, not least of all the Mirai botnet attack that compromised over 100,000 IoT devices as part of the DDoS attack against the DIN service back in October 2016. So this graph, we probably can't see it in too much detail, but the colors really show in terms of pink being the largest number of devices outlining the current state of Mirai-infected IoT devices globally. So the pink shows that over 1,000 devices infected back down to lower numbers in the blue, but you can really see the global scale of the problem here. So this and further more detailed analysis of this are really quite interesting, and we find it in capsula.com website. So, you know, in addition to kind of this motivation for setting up a harder institution, we also need to consider sort of the other threats that we're kind of facing in this space, and one of these is counterfeit devices. So what about clone devices? Are we heading for an internet of cloned things, and how do we go about protecting that? And we also have the issue of sort of untrusted supply chains. So in terms of counterfeit devices, this is, it's a really black market industry, and it's because of that, the true extent of it is really unknown. But we really do have a serious lack in adequate detection techniques. So it's really only like defense companies and so on that are putting measures in place to look at detection for counterfeit devices, but it is occurring sort of everywhere. And then in terms of untrusted supply chains, we now have the globalization of the supply chains. And the design and manufacture of today's devices is now distributed worldwide. We have the use of overseas boundaries. We have the use of third-party IP and third-party test facilities. So with so many different untrusted entities now sort of forming part of the design and assembly phases for electronic devices, it's actually becoming increasingly difficult to sort of ensure the integrity and the authenticity of devices. So the supply chain is now considered susceptible to a wide range of hardware-based security threats from IP piracy to IC cloning, IC over-engineering, hardware atrogyns, side-channel attacks that we've seen some talks on, and then also the, obviously, from the cloning and counterfeit devices. So if we're kind of looking at counterfeit devices, we're also looking at the problem that, okay, they're not just counterfeits, so maybe not performing as we would expect, but they can also then host malicious software, firmware, or hardware. And this is possibly what we're going to see more of going forward. So then another motivation factor is the threat of hardware atrogyns. So in recent years, the threat from hardware atrogyns is becoming more and more apparent. So last year a publication showed the first successful real world hardware atrogyn insertion, this was in an FPGA, into a commercial product. And that was against a high secure USB flash drive. So this is where the hardware atrogyn was enabled to manipulate the bitstream of the FPGA, allowing the, basically, the S-box of an AES to be manipulated such that it turned it into a linear function, and then it could be very easily cryptanalyzed. And then very recently we also had a hardware atrogyn reported that could exploit the sort of internal power distribution network in FPGAs. So this is quite an interesting piece of work. Again, it can be co-located with sort of crypto implementations on an FPGA and sort of internally sort of look at monitoring power consumption and thus being able to sort of uncover secret keys and so on. And then I couldn't have a presentation without also mentioning these two guys. We also, major motivation factor is obviously all of the recent vulnerabilities that have been affecting hardware devices. So we had the ROCA attack towards the end of last year, and then we've had this year the Meltdown and Spectre attack, and you'll hear more about that in the later talk today. So in terms of looking at what are the sort of major research challenges in this space? Well, how do we detect counterfeit devices? How do we detect devices that have manipulated in some way? Is it possible to build attack resilient hardware platforms? And how do we deal with this whole idea of untrusted manufacturing processes and the untrusted supply chain? So these are all kind of looking at the sort of harder research challenges that we need to be addressed in dealing with the hardware security threats. But then looking at the sort of flip side of that, can we kind of look at some of the opportunities that we can sort of take advantage of from sort of hardware roots of trust? So another aspect of this is looking at hardware security sort of use cases. So looking at combining hardware roots of trust with sort of more advanced crypto primitives that could allow sort of sticky policies to be created to protect for protected data. So this is like incorporating different forms of attributes, like user ID, device ID, location and so on. So examples of this are the FIDO, the fast ID online process. So looking at taking advantage of these, can we think about developing novel applications or novel hardware based security services based on or building on hardware roots of trust? So with this motivation we're sort of looking at the research institute in secure hardware embedded systems and our vision for RISE is really to create a global center for research and innovation in this space. And this will be with close engagement with leading industry partners and stakeholders and we wanted to become the go to place for high quality hardware security research. So a particular focus is also on the translation of research output into sort of new product services and business opportunities and building strong networks of national and international collaborations and research partnerships. So key to the vision for this center is sort of the sort of four key elements kind of depicted by the diagram here. So in terms of the institute we have a £5 million budget over an initial period of five years and initial tranche of that has gone into funding for research projects that I'll briefly detail later in the talk. So these will be the academic researchers sort of working on these research challenges. Then another key element is an industry and stakeholder advisory board that we're currently putting together and that sort of across the remit of the of the whole process in terms of a sort of a product life cycle. So manufacturers, product designers down to user communities and we also have representation from investors to encourage sort of startups and so on coming out of research institutions. And then the further elements include an institute manager with a business development function helping with the development and pull through aspect of the research output and we'll also have engineers working alongside academics helping to build proof of concepts and so on. So echoing the sort of major research challenges in area, the specific research challenges to be addressed by RISE are firstly looking at understanding the technologies underpinning hardware security. So this is looking at things like hardware security primitives, looking at tool sets techniques and trying to build attack resilient hardware platforms. The second key challenge is looking at maintaining confidence in security throughout a product life cycle. So this would be looking at understanding supply chain confidence, looking at modeling of hardware security and so on. The third key research challenges will be addressing is looking at this the sort of opportunities that we can gain from this. So looking at novel hardware based security services and use cases. And these will be looking at things like looking at novel secure document of your approaches, novel BYOD approaches and so on. And then the final challenges in the development and pull through process. So easing development, easing the best ways to leverage security in devices and so on and understanding the barriers to adoption of hardware based security solutions. So in terms of as I said the five million pound budget and initial launch was spent on sort of funding for component research projects. And those are as listed here and I'll go through these in detail for the latter half of the talk. So the component projects were went to Bristol, Cambridge, Birmingham and a project with myself at Queens. So the first of these the PIs are Mark Ryan, Flavio Garcia and David Oswald and they're going to be looking at user controlled hardware security anchors and looking at the evaluation design of these. So the their objectives really are to perform thorough security evaluations on a variety of hardware security anchors, such as Intel's SGX, ARMS trust zone and so on. And looking at enhancing those security mechanisms for user centric applications. And they also particular focus and kind of authentication for IoT. And they want to contribute really directly to securing sort of cost devices. And from that they'll be building demonstrators and so on. Their industry importers are HP and UPICO. The second project then is with Robert Watson, Simon Moore and Theo Marquettis at Cambridge and their project looks at the protection and memory safety for IO security. So their overall aim is to look at re-architecting current computer IO systems with security as a first class design constraint. So they'll be building an open source FPGA platform to enable security evaluation of IO devices, including pluggable devices like Thunderbolt 3 and USB-C. They'll want to evaluate current access control mechanisms for a range of commodity hardware and operating systems. And they want to explore the current use of IO memory management units and their ability to prevent attacks. And also obviously, as we've heard very often here already, that there's sort of the trade-off between adding security and the effect on performance. And then their final objectives are to explore new message-based IO architectures that basically avoid exposing memory to peripheral devices, thus mitigating current security vulnerabilities and improving the performance. And finally, exploring some forms of distributed memory protection technologies to avoid sort of a centralized bottleneck that the IONMU provides. And their industry partner is ARM. So the third project is then looking at a side channel hardened risk five platform. And this is with Dampage at Bristol. So his overarching objective is to try and harness risk five to explore and provide an open, secure, flexible drop-in hardware platform for high assurance use cases. So the idea is that this could be used sort of to form part of a wider layered solution for security aware design of hardware, software, and algorithms all play a role together. So the collaborators are Subarus and Talis. So in terms of the main work packages in this work is really to produce hardened implementations of the risk five designs with side channel resilience as one of the key sort of design metrics. And then to explore additions or alterations to the design to better equip it to support the current and next generations of crypto implementations. And finally to deliver a platform that democratizes side channel evaluation by sort of offering a lab free acquisition and analysis workflow for this. And then the final project is my own project which looks at applying sort of deep learning approaches to hardware security. So the sort of longer term goal for this project is to look at investigating the use of deep learning for security verification and EDA tools. So this is specifically in relation to looking at sort of detection of hard retrogens but also side channel analysis with the sort of overall goal of allowing non-security experts to see feedback on how to improve security of designs prior to fabrication. So in much the same way as design tools now currently give feedback on the performance or how fast or slow a design is or the area that there'll be a security metric involved in that. And the industry partners are Rambus, Rescue and BAE systems. So the research objectives here are really to examine and compare the application of machine learning and deep learning techniques in SCA against both unprotected and protected crypto implementations on a variety of platforms. And then to look at sort of investigating this idea of attack defender strategies inside channel analysis. So this initially to look at improving the attack efficiency of the deep learning models but then subsequently to look at using these kind of attack defender approaches to help to help improve side channel resistant designs. And we also want to look at conducting one of the first comprehensive evaluations of the application of machine learning and deep learning approaches for hard retrogens detection. And finally the to sort of meeting the overall goal evaluating the approaches to see how they might be utilized in security framework for EDA tools to provide feedback to a designer. So to finish I want to talk about really the next steps for the institute it was launched just in November. So our next steps really include an open call for participation in the advisory board. So member companies and stakeholders really have an opportunity to first of all challenge the researchers to try and really push the boundaries of future research in hardware security and try and address some of the major challenges and also help to help with the exploitation approaches and sort of maximize the exploitation potential from the research output. So they'll be able to engage with the research projects and offer perhaps commercialization opportunities and with the budget that we have we'll be holding future project and calls so it'll help us to inform future calls related to the research challenges. So we're a range of companies that have already signed up and we're looking for more sort of in this space with an interest in hardware security. We're going to be hosting events to bring primarily to the hardware security community in the UK initially together and then wider field and this will be starting with a spring school at the end of March this year to be located at the Trinity Hall University of Cambridge and more details of this as the talks and things are developed will be available at our website. And then going forward we want to develop international linkages and research partnerships more broadly and then as I said with the funding we'll have further targeted project calls throughout the lifetime of the project. So our next call for proposals will probably likely be the summer of this year. That's me. Thank you. All right thanks. Do we have any questions? We have some time now. Hi thank you for your talk. In 2013 there were a few Dutch researchers who were sued by Volkswagen in the UK and their publication on vulnerabilities in the Volkswagen cars Thales chips was prevented through the courts. How does RISE fit into this development? I mean I see you work together with Thales and are you looking for a coordinated vulnerability role? Are you working with researchers to make sure that such things go more smoothly? Are you actually enthusiastic about courts preventing such publications? What's your take on this? Yeah so Birmingham obviously have the guys at Birmingham have been involved in this and they I think from that point of view the individual research projects will have responsible disclosure approaches or strategies at each of the institutions as I said it's a multi-institutional sort of organization. For us I guess we'll have to ensure that there's a broader responsible disclosure approach for the whole of the research institute but primarily it will be the responsibility of each institution with their individual projects because they have responsibility for their own projects. So our role is more to facilitate the engagement between companies and research projects and then to sort of facilitate that at a high level and then any sort of further collaboration that takes place will be on an individual basis. So yes this is something we need to take into account and I guess we'll be holding sort of awareness and education sort of events around that for all of the institutions that will be involved so it's not just those four institutions that will be involved in RAIS it will be all institutions particularly in the UK and with an interest in this space. So our role is more as an education building awareness and so on but yes it's absolutely something that we need to take take account of thank you. Hi it's very nice to see actually public investment in this kind of stuff. How are you getting run was it if I understood correctly you have a five million for RAIS. Yes. How big is this kind of if you take all the different institutions that are set up and are they run more as kind of an actual government run research institute or more like a kind of a standing project pump which is kind of funneling money into projects but not as kind of an organization of its own. So I think it's what I make of it I guess. So I think there will be some initial funding provided sort of set the whole thing up but ultimately it's meant to be sustainable in the long term without sort of with sort of a baseline of government funding to kind of keep the operational approaches sort of there but it ultimately will be an organization to try and bring the community together both industry and researchers and to facilitate I guess in the future sort of more collaboration between the two parties. So it's not just about funding and funneling that into projects that will be kind of initially to get the thing set up but longer term is to kind of as a stand alone sort of institute to facilitate sort of the community. Yeah I wish you good luck we got something like that axed in Finland last year. Okay I should maybe learn from that. Just make sure we have any questions from above. I think so. Okay so then let's say Mayra again and we have coffee break now.