Moore's Law and the evolution of microprocessor technology

Computer Fundamentals

The pace of innovation across the series of technologies that comprise a personal computer continues to accelerate, often leading to product lifecycles that are eighteen months or less. The level of investment in Research & Development (R&D) in each of the areas assessed and analyzed in this paper continue to increase, often leading to higher levels of price/performance than succeeding generations of comparable systems (Dedrick, Kraemer, 2005). The intent of this analysis is to evaluate computer hardware and components, the operating system functions, how computers manage the input/output (I/O) process, in addition to the areas of multiprogramming and concurrent programming. The last two sections of the paper analyze hardware and memory management, and security.

Analysis of Computer Hardware

One of the most critical concepts relating to the innovation and rapid advances in computer hardware is Moore's Law (Larus, 2009). Dr. Gordon Moore is one of the founders of Intel Corporation, and during the initial product generations of the microprocessor he also helped to invent, he noticed that every year up to a 40% increase in the number of transistors per chip unit could be achieved while reducing the cost by 50% or more (Larus, 2009). Moore's Law is what has driven the rapid growth in functionality of microprocessors, video and memory components, and the pervasive use of Ethernet-based chipsets that allow computers to communicate directly over the Internet. Microprocessors' levels of functions supported quickly progressed based on Moore's Law, which became a design objective within Intel Corporation and still is today (Pankratius, Schulte, Keutzer, 2011).

Because of Moore's Law and the continual investment in R&D, Intel has been able to define the microprocessor market by innovating faster than its competitors innovate. The company's latest generation of Core i5 and i7 processors have been designed to support real-time video streaming on laptops and tablet-based systems and devices. In addition, Intel relies on Moore's Law to define their hyper-threading technology, which takes the lessons learned from multi-threaded operating systems of the past including Windows XP and Windows 7. Intel HD Graphics was also designed based on the advances the company made with their graphics accelerator technology within the last five years (Taylor, 2005). Microprocessor and chipset producers have also adopted the concepts of Moore's Law, and this concept dominates the industry today as a result. Form factors of PCs and laptops continually shift based on the advances made at the microprocessor and chipset levels. This allows for greater performance and portability of increasingly high performance systems without having to invest in expensive cabinets, cases or bulky systems. The pace of innovation is quickening as laptops and tablets gain greater computing power and sophistication over time, for the most part predicated on Moore's Law and the lessons learned from decades of development on microprocessors, graphics chipsets, and network chipsets. Intel is in all three of these businesses and continues to experience rapid growth as the i5 and i7 processors are now being adopted by Intel's customers into tablet PCs, laptops and high-end servers in anticipation of advances in operating systems, which is the next section of this paper.

Analyzing I/O and file systems, file structures, naming and disk management

The I/O systems of any computer are defined by the operating system. The goal in designing an I/O system is to ensure the highest level of performance while maintaining the integrity of the data as well. Each of the dominant types of operating systems, from Microsoft Windows, to UNIX and open source including Linux, take a different approach to defining I/O as each has a completely different kernel architecture that defines the entire operating system's functionality (Boudreau, 2010). All however share a common attribute of having I/O that is synchronous and asynchronous, and each operating system can be configured to support networked I/O functions, either ad hoc or structured (Boudreau, 2010).

File systems also vary significantly across operating system and define the taxonomies of how data will be indexed, queried, stored and accessed. File systems and the corresponding file structures are deigned to be optimized for use with the specific kernel aspects of the operating system (Volkel, Haller, 2009). The Microsoft operating systems rely on the File Allocation Table (FAT) file system, which in turn has a direct implication on naming conventions for files and disk management as well (Volkel, Haller, 2009). The inclusion of the High Performance File System (HPFS) within the Windows operating systems is a direct result of security becoming a primary concern of enterprise or large business customers and a design objective as a result (Funell, 2010). All aspects of the I/O systems, file systems, file structures, naming and disk management revolve around optimizing the efficiency of the entire operating systems' performance.

Subtopic 4: Multiprogramming & concurrent programming, history, processes and scheduling; mutual exclusion, synchronization and communication

One of the most important functions of the kernel is to orchestrate the use of memory and the needs of applications to complete tasks, often referred to as processing threads (Boudreau, 2010). Multiprogramming and concurrent programming are used by operating systems to stay synchronized to the microprocessor and provide commands as to which task or thread to complete in which sequence. In this way, multiprogramming and concurrent programming serve as the foundation for optimizing processes and scheduling based on the systems' constraints (Volkel, Haller, 2009). Using multiprogramming and concurrent programming, Microsoft was able to create the first multithreaded, commercially successful operating system in Windows NT (Taylor, 2005). The lessons learned in that operating systems' development are what allow for multithreading of 64-bit applications on mobile devices that require very little power; an engineering design accomplishment not possible just three years ago (Boudreau, 2010). Multiprogramming, synchronization and communication are critically important for making an operating system optimized for its hardware constraints.

Subtopic 5: Hardware and Memory management: static relocation, virtual memory, segmentation, paging, load control, etc.

From the very first operating systems, the need for hardware and memory management was designed into the kernel, often with procedure calls and later with Application programmer Interfaces (APIs) that would allow for virtual memory definition and segmentation (Boudreau, 2010). The use of virtual memory management, paging and the optimization of memory is today managed in the Windows operating system at the processor thread level, and can be configured for pre-emptive vs. collaborative multitasking (Volkel, Haller, 2009). This refers to an operating systems' capability of managing the many interrupts to process requests and requirements while at the same time, and has led to the development of applications that can partition memory requirements to also optimize performance and response time from calculations.

Subtopic 6: Job management and protection; include a serious discussion of security aspects

The most commonly leveled criticism of operating systems is the inherent lack of security they have (Funell, 2010). Defining operating systems to have partitioned memory is just the start, as Microsoft learned with their Windows NT platform. Dedicated memory partitions by user account can be hacked and have been (Funell, 2010). The need for greater levels of user authentication is required, including the use of biometrics for advanced systems that have highly confidential data within them. The reliance on security-based algorithms that also seek to analyze patterns of use to anticipate security threats are increasingly in use today (Volkel, Haller, 2009). This aspect of an operating system can capture the levels of activity and the patterns they exhibit, which can provide insights into when a threat is present or not. The use of predictive security technologies, in addition to biometrics, is changing operating systems rapidly in 2011 (Funell, 2010).