Computer Engineers of VMIT

म्यूचुअल फंड्स कैसे खरीदें ऑफलाइन : म्यूचुअल फंड में ऑफलाइन निवेश करने के लिए आप किसी financial intermediary यानी वित्तीय मध्यस्थ की सेवा का उपयोग कर सकते है...

For anyone visiting the San Francisco Bay Area, a visit to the Computer History Museum in Mountain View is an important stop. The Bay Area is home to Silicon Valley, center of the universe for many of the past 30 years’ worth of technological breakthroughs.
It’s difficult to “see” Silicon Valley as you drive around, but the Computer History Museum gives a unique opportunity to learn about the people and technologies that have an ever-increasing impact on our lives.
_______________
BEYOND BABBAGE:
_______________

The Computer History Museum focuses on both the technical artifacts that make up our computational past as well as the human stories of the challenges, risks, and failures that led to the breakthroughs that define our field. One of the earliest artifacts in the museum is the Babbage difference engine that computes polynomials using mechanical computation:

I love the Babbage Engine because it’s a model of the way great innovation works. Charles Babbage was a brilliant mathematician, working on a problem that involved books of log tables filled with typographical errors. All the calculations had to be done by hand, transcribed by hand, typeset by hand—the margin for error was phenomenal. Babbage dreamed of a special-purpose, extremely sophisticated calculator that would take all the guesswork and human intervention out of generating these tables. It was exactly the right approach to exactly the right problem. The difficulty was that like many entrepreneurs, Babbage had a brilliant idea but couldn’t get it funded. He never got further than building a piece of it, but he created the blueprints for the full engine in fantastic detail.

Although Babbage designed his difference engine in great detail, he never knew if it would actually work. In 1985, some 150 years after its conception, engineers and specialists at the Science Museum in London undertook building a difference engine based on Babbage’s original plans. They completed the first difference engine in 2002, and you can view it in London:

A second engine was built, so the first one is in London, but the second one is here at the museum in Mountain View, California. It’s the complete Babbage Engine, because Babbage not only envisioned this enormously complicated two-ton machine with 8,000 moving parts, he also envisioned a printer and a stamper that would create the plate from which the book could be printed. Babbage dreamed of warehouses full of these machines being driven by steam engines so they would be constantly calculating. Our machine, which we crank every day at 1 p.m. Pacific time, now works exactly as Babbage envisioned. As brilliant as he was, he died embittered, wondering what the world might have been if he had been able to create it. Generations later, entrepreneurs worked from those plans and brought them to life.

In the 1930s and 1940s, computing moved from mechanical to electronic devices, triggered initially by technological breakthroughs followed by rapid innovation as well as problems and conflicts:
I love our recreation of the Atanasoff-Berry Computer [ABC]. First, it’s the only one that exists. It was built exactly according to the plans that John Vincent Atanasoff and Clifford Berry devised at Iowa State in the 1930s. As you might recall, it was the subject of a very famous patent dispute that destroyed all of Eckert and Mauchly’s original patents based on the ENIAC and the birth of electronic computing. As Gordon Bell says, “It was the dis-invention of the computer.” You can stand in that gallery here, which is called “the birth of the computer,” and you can see the ABC in all of its simplicity standing right next to the JOHNNIAC, [John] von Neumann’s famous computer and one of the first Williams-Kilburn tubes and the Engima encoding machine, right next to a film about code breaking in World War II. So much of history comes together in the space of about 1,000 square feet with the real machines sitting there.
___________________
THE RISE OF CIRCUITS:
___________________

In our field, amazing breakthroughs and rapid innovation are always followed by the scaling up and practical applications of the new technologies. By the 1960s, IBM was designing the 360 architecture that’s still in wide use today:

The IBM 360 is a great story, not only the system itself but also because it was a “bet the company” decision by Tom Watson. In the 1960s, he spent what today would be the equivalent of about $60 billion in research and development money on a single project that succeeded and charted IBM’s course for the next half-century. If it had failed, it would have sunk the company.
The electronic computing revolution went from vacuum tubes in the 1930s to transistors in the late 1940s to the invention of the integrated circuit in the late 1950s. Sixty years later, our current technology is still based on increasingly powerful and sophisticated integrated circuits. The invention of integrated circuits both created and gave the name to Silicon Valley:

We have a replica of Jack Kilby’s notebook open to the page from September 1958 where he reports for the first time his whole design for his version of the integrated circuit. Sitting there in Texas Instruments, he’s told by his bosses in the middle of a Texas summer, “Don’t work on this idea, Jack—we all know you’re in love with it, but you have another job.” They all leave on vacation, and what does Kilby do but spend the month working on the integrated circuit. Everyone comes back in September, he hooks it up to an oscilloscope, and it works exactly as he predicted. At the very same moment, out in Silicon Valley at Fairchild Semiconductor. Noyce, Moore, Hoerni, and Last are working on the same idea, and within a period of months, literally the “big bang” in computing has happened, and everything changes.

Once we had the integrated circuit in hand, we quickly went from invention to engineering. The goal was to make cheaper and faster integrated circuits. Gordon Moore coined Moore’s law, which states that the number of transistors we can put onto an integrated circuit will double about every two years. He predicted the trend in a paper in 1965, and it has been remarkably accurate even to the present day:

The semiconductor gallery is a really special one—we have wonderful examples of wafers as small as a nickel and as large as a dinner plate, and you can see the evolution over time of how complicated it is to make a microprocessor and how difficult the science is.
As is always the case, once a scientific or engineering breakthrough happens, the real money is in applying these new technologies to solve human problems. The creation of increasingly powerful and inexpensive integrated circuitry unleashed a torrent of new applications for computing technologies:
When you go down what I call “application alley” at the museum, you see in very quick succession how we went from artificial intelligence to robotics to graphics to music to art. And then [you move on to] Xerox PARC, the birthplace of the graphical interface, the mouse, and the chorded keyboard that Doug Engelbart invented. [You see] the first Pong machine—the birth of Atari and all the gaming companies that came after it—as well as the personal computer, the Apple/IBM PC battles, and a million different flavors of PCs that flourished. You go through this in very quick succession after you tour the integrated circuit, and you come to a world that we’re experiencing in a very personal way now, but once you’ve walked through it, you know that we’re only still at the beginning.
As we’ve moved from the era of mechanical computing to cell phones that are more powerful than the massive supercomputers of the past, it’s important for all of us in computer science to understand what it took to create our current technologies, so that we can best imagine how we might make our own contributions and continue to evolve our field. For more information about the museum, visit www.computerhistory.org.

Safety-Critical Systems: The Next Generation
----------------------------------------------------

Safety-critical computer-based systems are woven into the fabric of our lives. These systems
must work adequately given user behavior, system interactions, changing environments
and expectations, organizational turbulence, regulatory caution, routine component and operator
failure, the complexity of international projects, and adaptation and refurbishment as well as
security-related issues such as intentional malicious attacks and supply-chain risks.

Safety and Security
----------------------
Apart from the fact that safety-critical systems is an important topic in its own right, IEEE Security
& Privacy is addressing these issues for two other reasons. First, the magazine’s remit is
much broader than “security and privacy.” Its tagline of “Building Dependability, Reliability,
and Trust” reflects that we are partially owned by the IEEE Reliability Society and have a broad
interest in trust and dependability. The other reason is that safety will be an increasingly relevant
application area for security and privacy specialists. These days, air gaps and isolation are
seldom credible arguments for security—the US Department of Homeland Security found, on
average, 11 connections between SCADA and enterprise systems.1 Thus, we can’t consider any
computer-based safety system to be truly safe unless we also address its security.
Both safety and security aim to protect something. Broadly speaking, safety is concerned
with protecting the environment from the system, whereas security is concerned with protecting
the system from the environment. The issue is how to ensure that the protection is adequate.
A classic view of safety is that it’s concerned with preventing accidents by identifying potential
weaknesses, initiating events, internal hazards, and potentially hazardous states, and then identifying
and applying appropriate mitigations to reduce the risks to a tolerable level. Security is concerned with protecting assets against internal and
external threats and vulnerabilities that compromise
them using controls that reduce the risk of compromise
to an acceptable level.
The next generation of safety-critical systems
includes not only the rather obvious application areas
such as air traffic management, nuclear power plant control,
and military systems but also networked patient
care, driverless cars, autonomous air vehicles, and personal
apps. Undoubtedly, there will be new technologies
for building and assuring these systems as well as the
adaptation and evolution of tried and tested approaches.

source@ieee

NEXT GENERATION MEMORY--
**********************


The memory technologies in today's computing systems all emerged in the early 1970s at the dawn of the semiconductor industry. Solid-state memory—static RAM (SRAM), dynamic RAM (DRAM), and flash (initially EPROM)—is based on electron storage in transistors, while mechanical memory—tape and the hard disk drive (HDD)—relies on magnetic storage. These two storage media have had an amazingly long life, roughly doubling in density and halving in cost every two years according to Moore's law. The scalability of these technologies explains why they have been a key enabler in the emergence of increasingly complex computing devices.
Memory is a defining component in many of today's portable devices that themselves are becoming indispensable to our lives. In a smartphone or tablet, memory is typically second, and often equal, to the display as the largest cost component of the system, well above the CPU. For high-end computing systems such as servers, memory defines system performance and power more than any other component. As the demand for large amounts of data instantly accessible to millions of users continues to increase, memory technology becomes both a solution and a bottleneck, spurring the industry to redefine how these systems use memory. One of the best examples of this is the emergence of solid-state drives (SSDs) across the range of computing devices.
The rapid density growth and cost reduction of memory has been enabled by scaling the underlying technologies to ever-smaller feature sizes and thus smaller memory-bit sizes. Unfortunately, this scaling is starting to approach the limitations of the storage physics, making the task of maintaining the cost/density value curve increasingly difficult. In a state-of-the-art 20-nm NAND multilevel cell (MLC) memory unit, the memory state is stored using only a few tens of electrons. Increasing management of the memory bits is required for the system to be able to continue to use them. Fortunately, the ability of modern manufacturing tools to manipulate materials at the atomic level, combined with an improved understanding of and ability to model fundamental storage physics, is opening the door for new advances in memory. Emerging memory devices could prove to be even more scalable than current devices, ensuring continuation of the cost/density value curve.
IN THIS ISSUE
Back to Top
It is generally agreed that today's memory technologies, such as NAND flash, are still scalable for several generations, but it is equally clear that physical limitations will soon slow their pace of scaling and further reduce their functionality. Increasing levels of memory management will therefore be needed to prolong the life of these technologies. “NAND Flash Memory: Challenges and Opportunities,” by Yan Li and Khandker N. Quader at SanDisk, examines the scaling direction for NAND and the management techniques that will be required.
New memory technologies generally do not function like today's memory technologies and have different trade-offs for performance, power, and cost. Similar to the way SSDs are changing the hierarchy of memory usage, emerging technologies are likely to have a transformational effect on memory usage and integration in computer systems. In “What Lies Ahead for Resistance-Based Memory Technologies?,” Yoon-Jong Song and his coauthors at Samsung Electronics highlight and discuss the cutting-edge physics and features that new memory technologies enable.
There is a reason why existing memory technologies have been so successful and no significant new ones have been introduced in 40 years. Introducing a new memory type is complex: the storage physics are not well understood, designing large-density storage devices and state-of-the-art lithography techniques is challenging, and manufacturability at high volumes and yields is unproven. Consequently, it can take up to a decade for a new memory technology to evolve from a basic concept demonstration to a finished commercial product.
New memory types likely will first supplement existing memory technology to help overcome the latter's scaling deficiencies. They might also find an entry point in applications that leverage their unique set of features. Phase-change memory (PCM), magnetic memory, and ferroelectric memory have all emerged at various levels of density and lithography. Of these, only PCM has been demonstrated at Gbyte-level densities and “near state of the art” lithography. PCM combines some of the properties of DRAM and NAND, providing a new set of features. Although PCM is the first, several other technologies under active development promise similar features.
The industry now faces two challenges: developing novel memory types to enable further density scaling, and integrating these into computer systems. These challenges are compounded by the high likelihood that the new memory technologies will have different functionality than the existing ones. In “The Nonvolatile Memory Transformation of Client Storage,” Intel's Amber Huffman and Dale Juenemann examine the transformation that SSDs enable in client systems and the potential for emerging memory types beyond the SSD.
As new memory technologies are introduced into computer systems, so changes the memory hierarchy, which has been defined by the evolution of memory-type capabilities over the past four decades. How computer hardware and software deal with memory has been defined by its access latency, access granularity, volatility, power, and cost. The memory hierarchy has evolved into a multitier system with various levels of caching in SRAM, main memory in DRAM, “fast” storage in the HDD and now SSD, and “slow” storage in tape.
The memory hierarchy, and the resulting hardware and software wrapped around it, are as much defined by the perceived deficiencies of the memory types as by their advantages. For example, DRAM, the most common memory for main code and data storage, will “forget” what is stored when the power is removed. Worse, it will forget even when the power is on, resulting in the requirement to periodically refresh the data. The refresh consumes time, power, and computer resources to manage it. Certainly these are not desirable properties, but because DRAM could be built for a reasonable cost and was scalable, system designers learned how to deal with its inadequacies. Similar observations can be made about other memory technologies, none of which are ideal, and system infrastructure has been built up around extracting value out of each of them.
Hence, new memory types will significantly impact software design, which takes advantage of the memory hierarchy. In “How Persistent Memory Will Change Software Systems,” Anirudh Badam of Microsoft Research addresses the potential exploitation and effect of emerging memory technologies on the OS. And in “Refactor, Reduce, Recycle: Restructuring the I/O Stack for the Future of Storage,” Steven Swanson and Adrian M. Caulfield of the University of California, San Diego, focus on the software-to-hardware interface, evaluating the potential repartitioning of functions across this boundary enabled by such technologies.
Many new memory types are in various stages of commercial development. PCM, commercialized by Micron Technology and Samsung, is on hold until market conditions support mass production. The industry faces significant challenges over the next 10 years as it determines how these emerging memory technologies will evolve from both a manufacturing and memory hierarchy point of view. The impact of these developments will be extremely long lasting.

The problem of making programming both accessible and exciting has its roots in early schooling and is international. In an interview with The Guardian on 10 January 2012, the UK's Secretary of State for Education, Michael Gove, was highly critical of the current school system and suggested that "instead of children bored out of their minds being taught how to use Word and Excel by bored teachers, we could have 11-year-olds able to write simple 2D computer animations." But how exciting is programming to students? Asked in the context of a typical computing course, one middle-school student summarized her perception of programming as "hard and boring," which doesn't suggest a workable tradeoff but instead a heartbreaking lose-lose proposition.
Various government agencies and private organizations have launched numerous efforts to broaden student participation in computer science, and after many years of failed attempts, general interest is finally beginning to grow. For instance, a YouTube video from code.org that quotes Steve Jobs on the relevance of programming and features interviews with famous computer scientists managed to get an unprecedented 10 million views in a short amount of time. It's clear that something should be done to make programming accessible and exciting, but the question is how.

Project Goals
**********
has the ambitious goal of revolutionizing computer science education in public schools by introducing students to computer science through a combination of game design and science, technology, engineering, and math (STEM) simulation creation integrated into the middle-school curriculum. To date, more than 10,000 students from inner-city, remote rural, and Native American schools have participated in the US's largest middle-school computer science education study. Researchers at the University of Colorado systematically developed and evaluated this education strategy based on four core principles:
• Exposure. Broaden participation and reach every student by injecting an easy-to-teach one-week game design module into existing mandatory keyboarding or MS Office "computing" classes.

• Motivation. Motivate students by carefully balancing challenges and skill levels through game design activities in an SGD curriculum that ranges from simple Frogger-like games to advanced Sims-like games.

• Education. Build instruments that analyze student projects for critical STEM-skill acquisition so that learning outcomes can be measured objectively. A latent semantic-analysis–inspired approach helps determine computational thinking and skill transfer between game design and simulation creation.

• Pedagogy. Investigate the interaction of pedagogical approaches and motivational levels across genders and ethnicities. With an optimal pedagogical approach, 35 hours of careful instruction is enough to train teachers to teach SGD curricula in a gender-friendly way (45 percent of our participants were female).

Project data suggest that the SGD strategy works extremely well: 74 percent of male participants and 64 percent of female participants wanted to continue with similar courses as electives.
Making Programming Exciting
Back to Top
Simply put, creativity and ownership are the keys to making programming exciting. Most students aren't interested in the act of programming itself—instead, they want to create animations, make stories, or build games. As early as 1991, with our work on AgentSheets, we found that giving students the ability to create their own shapes significantly increased their motivation. Instead of simply replicating an existing project from a tutorial, they took ownership of the creative process by drawing their own shapes and creating their own worlds. Instead of having cars hitting a frog in a Frogger-like game, they might replace the cars with dogs and the frog with a cat to create their own game. Once students create and own their shapes and worlds, they're much more interested in the idea of using programming as a truly empowering process to bring their creations to life.
If ownership is essential to motivation, how can we advance to more sophisticated levels of creativity, such as the creation of 3D shapes? Most students are intrigued with 3D but only from the consumer viewpoint. Typically, it's quite difficult to create 3D shapes from scratch, so 3D creativity presents a direct challenge to student ownership and motivation. Existing 3D programming tools generally offer limited options—such as selecting prebuilt 3D shapes from a collection or importing 3D shapes from the Web. We believe that using such prebuilt shapes actually reduces ownership and creativity, so we foster student ownership through a casual design tool that lets users create their own 3D shapes.
The goal of casual 3D design is not to create tools for Pixar-level animators but to reach students with no background in 3D modeling whatsoever. With AgentCubes, students create basic 3D models that represent recognizable 3D shapes such as people, animals, and other objects and use them to construct 3D worlds. Inflatable icons as depicted in Figure 1 instantly engage students by helping them create their first 3D shape in about a minute.




With AgentCubes, students can use sophisticated spatial reasoning to build complex 3D worlds that include layers as well as portals to other worlds. Programming includes camera control in first-person and bird's-eye views. Figure 2 depicts the level of 3D design complexity of a game built by a middle-school student. Projects can be run and authored as desktop applications or as HTML5-based Web applications that can run in desktop and mobile browsers without Flash or Java.

Making Programming Accessible

A dangerously trivializing perception implies that drag-and-drop programming, which we helped pioneer with AgentSheets, makes programming easy. It does not. To be sure, eliminating the syntactic challenges of traditional programming languages can remove some devastating programming frustrations, but these gains are commonly overestimated. No one seriously claims that spell-checkers, which are syntactic tools for natural languages, empower people to write best-selling novels. It's just as absurd to harbor similar hopes for syntactic programming tools. Through our decades of training teachers and students around the world to make games and simulations, our SGD project team has found instead that programming's semantic challenges are considerable and generally not well supported by programming environments. Almost anyone can learn how to write a program, but few learn how to debug it.
Semantic tools help users create programs that actually work. Contemporary computers are enormously powerful and can do much more than just provide syntax support. They can proactively help students comprehend the meaning of a program and visualize the difference between the program they want and the program they have. In AgentCubes, this computational power is employed to have the computer run the user's program one step into the future to visualize possible consequences. Figure 3 depicts how AgentCubes annotates the program of the selected object. The red/green/neutral annotations reveal which rules will be tested, which one will fire and if the conditions tested are true or false. In the depicted situation the car will be moving up. These proactive annotations help both with program construction and by letting users identify and fix bugs before they turn into actual problems.

Our data shows that the SGD approach works—even in some of the toughest, poorest, and most isolated schools in the nation. Our results consistently exceed Gove's vision of 11-year-old students building simple 2D computer animations. SGD students not only make simple 2D animations but also create playable games based on sophisticated concepts that include advanced mathematics and artificial intelligence. Significantly younger children at the elementary-school level were able to create 2D and even 3D games. Perhaps most important, many of our SGD students have advanced beyond making games to actually find pleasure in building scientific simulations.
Source@ieee

Software-Defined Networking: On the Verge of a Breakthrough?
***********************************************

As developments such as cloud computing, virtualization, and massive datacenters make networking more complex, network managers are looking for smarter systems and new and better ways to control and manage them.

Enter software-defined networks.

SDNs separate the network's control plane from the data plane, noted Clark DeHaven, senior director of corporate strategy at networking vendor LGS Innovations, a subsidiary of Alcatel-Lucent that serves US government agencies.

Software on centralized general-purpose servers, rather than on individual networking devices such as switches, runs the control plane. As Figure 1 shows, SDNs have APIs that enable administrators to centrally program and manage network resources via applications and services, DeHaven explained.

Figure 1. Software-defined networks transfer network control from individual routers and switches to a control layer that runs on a centralized server. This layer has APIs that let organizations use applications to centrally manage network services.
figure 1

This enables network administrators to program the entire control plane via a common API, easing the process and giving them more fine-grained, flexible control over traffic flows and the policies networks use to manage them. In the past, administrators have had to program every network device to make systemwide changes.

In essence, SDNs transfer network control from various pieces of hardware that must be managed individually to a single programmable platform.

Nonetheless, said Douglas Gourlay, vice president of marketing at Arista Networks, SDNs haven't taken off as quickly as some predicted because they represent a new technology and they carry some risk. For example, he noted, network control could be lost if the single server hosting the control plane goes down.

However, some industry watchers believe SDN is now ready to take off.

In fact, said Andrew Harding Sr., product marketing manager at SDN vendor Big Switch Networks, an increasing number of organizations are already purchasing SDN products.

Under the Hood

In the late 1990s, the rise of the Internet disrupted typical phone-network operations. Numerous new companies tried to take advantage of this by replacing the phone networks' hard-wired switches with an IP fabric that separated the control and data planes.

Stanford University researchers explored SDN-related concepts in their Ethane project (http://yuba.stanford.edu/ethane), whose first deployment was in 2006.

SDN implementations by major companies such as Cisco Systems and Juniper Networks are proprietary and thus don't work with equipment from other networking vendors. This has limited organizations' ability to engineer and manage traffic across equipment from multiple vendors.

According to Big Switch Networks' Harding, SDNs bring together three key elements: the switch, the network controller, and network applications.

In the past, both the packet forwarding and high-level routing decisions occurred on the same router or switch. The SDN switch separates the two functions. The forwarding still occurs on the switch, but the routing decisions take place on a network controller, usually a server.

The switches use either a proprietary technology or the Open Network Foundation's (ONF's) OpenFlow protocol. OpenFlow manages traffic in a way that lets the control-plane server tell the switches where to send packets, thereby moving this functionality from the data plane.

The controller runs the various policy, security, traffic-engineering, and other applications that control network elements via APIs, explained ONF executive director Dan Pitt.

These APIs make it much easier to add new functionality quickly, noted Harding.


Source@IEEE