Welcome back to Hardware Security. This week we will talk about how hardware designers can protect their design intellectual properties from being misused. We will learn what design intellectual properties are and why they are important in modern hardware design. We will show that in addition to law enforcement, there are several techniques available for you to protect design intellectual, intellectual properties by yourself. These techniques include digital watermarking, digital fingerprinting, integrated circuit metering, and design obfuscation. So, the discussion of this message we will also see how to make the tradeoff amount, security, cost, and system performance. No special background is needed to understand these concepts. However, the examples we will use require some basics of digital logic design. We will also introduce the concept of graph coloring problems and graph partitioning problems. How we use them? According to the Cambridge International Dictionary of English, intellectual property is an original idea, which can be used to earn money. Original reflects the novelty of the intellectual property and that the phrase can be used implies the usability of the intellectual property. Examples of intellectual properties are products, technologies, softwares, and so on. Meanwhile, it also states that's the person or group who is recognized as having the idea can use the law to prevent other people from earning money by copying it. This law enforcement includes, patents, copyrights, trade secrets, and so on. In VLSI design, an intellectual property is a dig, is a design unity that can be reasonably viewed as a stand-alone sub-components of a complete system on-chip design. It can be any innovation and the technology that make designs better. For example, a place design algorithm, a power management technique, or a noble design or testing method. These are all considered as Design IPs. More tangible examples include, microprocessor course, on-chip memories, PISA interface micro, fully routed netlist, Verilog chip descriptions, and so on. These are also as virtual components, blocks, cores, or sometimes called a macros. Based on their performance and the flexibility. Design, design intellectual properties can be categorized into three groups. Hard IPs, they include GDSII files with test the list and high level models, customer physical layout, fully placed in a routed netlist for a given partic, technology library. These hard IPs have predictable performance, and most of the time, optimized. But they are very, very inflexible. On the other hand, soft IPs are delivered in the form of synthesizable HDL code, like Verilog or VHDL programs. They offer great design flexibilities despite the la, the last predictable performance. Examples of Firm IPs are placed RTL sub-blocks and the fully place netlist for, for generic technology library. They can be, they can be considered as a compromise of hard IPs and soft IPs. The driven force behind intellectual properties is the design productive, productivity gap. The gap between what can be built and what can be designed? This figure is available from Wikipedia. It shows the CPU transistor counts in the past of 40 years, following the Moore's law. The Moore's law says, that the number of transistors on a chip is doubled every 18 months. I don't want you to read all the words in small font in this figure. What is important is that back, back in 1971 the Intel 40, 4, 4004 processor has 2300 transistors and they're 10 micrometer technology. Now, technology has advanced to 20 nanometres and more than processors have buildings of transistors. The capacity of si, silicon has increased for almost 1 billion times in the past of 40 years. But designers ability to do meaningful design. With such transistors has not increased at the same pace. This creates the design productivity gap. One of the most effective way to close the gap is reuse-base, reuse-base the design. Starting, starting from the mid 1990s. There are many industry initiatives on IP reuse. In the US, there are Virtual Software Interface Alliance, also known as VSIA, and Virtual, Virtual Component Exchange. At the design and reuse this, organisation in Europe, Europe. An IP-Highway is the one in the Asia. In reuse-base design IPs are developed not only for use, but also for reuse. Design engineers are forced to cooperate and share their data, expertise, and experience. These are details I encourage to be documented and made public to make reuse more convenient. But, at the same time, these efforts has made IP piracy and the infringement much easier than ever. Since the early 1990's litigation cases and semiconductor related to IP infringement have been increased very, very fast causing an estimated annual revenue loss of billions of dollars. The most dangerous threat of IP pi, piracy came in the form of reverse engineering, chip overbuilding, and counterfeiting. Reverse engineering is the process, where a product is taken apart and analyzed in order to improve or duplicate the product. For hardware design IPs, the cost and the schemes required for reverse engineering are much less than those to develop the original IP. This gives a taggers incent, incentive for IP piracy. Chip overbuilding refers to the scenario, where our foundering fabricates more chips than the order. And then sells those extra copies to the market to make profits. Counterfeiting is the case when a low quality product, or IP, is manufactured but is sold as a high quality brand name. The Virtual Software Interface Alliance has listed the following four goals for intellectual property protection. Number one, enable IP providers to protect their IPs against unauthorized use. Number two, protect all types of design data use to produce and deliver IPs. Number three, detect use of IPs. And finally, trace of IPs. [NOISE] In practice, most design, design IPs are secured by deterrents and the protection mag, mechanisms. Commonly used the deterrents, includes patents, copyrights, contracts, trademarks, and the trade secrets. They do not directly provide IP piracy they do not directly prevent IP piracy. But rather discourage the misuse of IPs because the attacker was being caught. They face lawsuit and severe penalty to recover the financial loss of the IP provider. However, it is the IP provider's task to identify and report IP infringement and find the attacker. Protection mechanisms, on the other hand, use means such as encryption, chemicals, obfuscation, and dedicated hardwares to prevent and authorize the access to the IP. Standard encryptions can be applied to protect these IPs. For example Both Xilinx and Altera have offered encryption on their IFPG boards. Chemicals can also be integrated into the chip for a passive protections, mainly from military devices. When the surface of the, the chip is scratched and exposed to the a, to the air it will damage the exposed silicon. And thus, prevents reverse engineering. Meanwhile, obfuscation method have been proposed at various design phases in order to disguise the design and make reverse engineering much more difficult. In addition, industry organizations have been working on establishing standards for IP reuse and IP protection. For example, Virtual Socket Interface Alliance has published two tagging standards to protect hard IPs and soft IPs. Law enforcement is the only means to get back the revenue lost to IP piracy. The encryption and other protection mechanisms make IP piracy more difficult, and more expensive. But only the detection schemes such as digital watermark. digital fingerprinting, and integrated circuit metering can enable IP providers to identify the occurrence of IP piracy. This identification will be the first step towards legal fight against IP infringement. We will focus on this message this week.
