Category Archives: Economy

Switching to the cloud? Advantages and disadvantages

Posted on by
1

Cloud computing graphic (courtesy of Sam Johnston, accessed on Wikipedia)With tougher economic times, companies are reducing IT budgets but are still finding the need to replace legacy systems. Because of this, claims a Gartner report released earlier this year, many CIOs have turned to cloud computing as a way to save on IT costs.

Cloud computing, for all of its recent publicity, is not new. In the decades before PCs, companies purchased computing resources on mini-mainframes that were accessed through end-user terminals. Cloud computing’s recent re-emergence is a result of companies again considering the cost savings of having someone outside the company handle IT infrastructure and maintenance.

But the companies that have arguably the most to gain from cloud computing, small and medium sized businesses (SMBs), are still unconvinced. According to a recent Newtek survey, 48% of small businesses owners do not see a switch to the cloud as a cost-reducing move. Continue reading →

HPC can grow US manufacturing

Posted on by
Reply

Photo of an fighter aircraft construction plant in New York, ca. 1944HPCwire recently ran a great article describing in a nutshell the potential that high-performance computing (HPC) can have in increasing the competitiveness of US manufacturing.

Despite what media coverage about the ailing manufacturing sector may lead one to believe, the US still leads the world in manufacturing output. In 2009, as Michael Feldman of HPCwire describes, the US accounted for 20% of the world’s manufacturing output and was 45% more productive in this area than China. Nevertheless, there are challenges ahead; the US is only fourth in manufacturing competitiveness worldwide according to one study.

Why is manufacturing such an important part of the US economy even though it employs only 10% of the national workforce? Feldman writes:

The real value of the US manufacturing sector is that it’s at the heart of much of the science and engineering innovation on which the remainder of the economy rests. Today US manufacturers employ more than a third of the country’s engineers and account for 60 percent of all private sector R&D. As such, it creates products that are used by the more lucrative service industries. Think, for example, of all the myriad services that are dependent on the production of computer chips and other electronic devices. Manufacturing, like agriculture before it, is a foundational activity that acts as a catalyst to other business sectors.

As we’ve been following on this blog, this great potential for reinvigorating the foundational sector of the US economy has been taken up as a cause by various organizations ranging from the federal government to the National Center for Manufacturing Sciences (NCMS) to universities to system integrators like us. Continue reading →

U.S. manufacturing and HPC: A bold plan to equip the “missing middle”

Posted on by
1

Photo of a factoryHPC in the Cloud ran an article last week written by Jon Riley of the National Center for Manufacturing Sciences (NCMS). The article talked about the initiative by NCMS to rejuvenate manufacturing in the United States by connecting manufacturers with supercomputing hardware and software to make them more competitive in the global marketplace.

The leap from tried-and-true methods to newer technologies has historically been a difficult one, but it almost always pays off big in the long run. In the late 1990s, for instance, the Electric Boat Corporation (which has been the builder of submarines for the U.S. Navy ever since they were introduced a hundred years earlier) embarked on designing and building the USS Virginia, the first of a new class of attack submarines. This was the first time a U.S. submarine was going to be designed entirely on a computer before it was to be built.

There was no margin for error. A brand new class of submarine had to go from design to construction to launch in a matter of several years and perform excellently at sea to prove its worth as a weapons platform.

As it turns out, the USS Virginia was launched on schedule in 2003 and operated without a glitch during its first sea trials. Computer-aided design, which had accurately simulated every corner and lever of the submarine, proved an extremely effective way to design submersibles. The CAD software even alerted designers whenever a space was laid out in the blueprint that would prove too cramped for sailors to live and work.

Continue reading →

Worldwide race heats up for HPC leadership

Posted on by
1

Planet earthAn article in Computer World by Patrick Thibodeau reports how David Turek, the vice-president of IBM, spoke last week about the growing emergence of China in the worldwide HPC competition. Turek said, “Within a year, there will be more Top500 systems in China than there are in Europe collectively.”

China currently holds the claim to having built the second-fastest supercomputer in the world. The U.S., according to the article, has 282 supercomputers in the Top500 listing, China contributes 24, and Europe has about 100. Growing their share of HPC four times over in the next year would be an impressive achievement for China.

U.S. supercomputer manufacturers, naturally, are somewhat fearful of this progress abroad. Turek, somewhat alarmed, commented, “You have sovereign nations making material investments of a tremendous magnitude to basically eat our lunch, eat our collective lunch.”

U.S. business and government are taking measures to make the United States even more competitive in the HPC market, lest that homegrown production goes the way of the U.S. automobile industry. While China is building an enormous new supercomputing center in Shenzhen, the challenge for the United States will be not only to continue leading technologically, but convincing more and more sectors of the U.S. economy that HPC solutions can grow their business beyond what their obsolete equipment now allows.

Mircosoft responds to government raids in Russia

Posted on by
1

Image of Lady Justice courtesy of Einar Einarsson Kvaran. Accessed on Wikipedia (http://en.wikipedia.org/wiki/File:JMR-Memphis1.jpg)An article in Computer Reseller News by Kevin McLaughlin describes how Microsoft has been forced to change its software licensing policy because the Russian government has been overzealously prosecuting small organizations for violating piracy laws.

In reality, these Russian NGOs that have had their offices raided by the Russian government were not committing software theft. Rather, they offended the Russian government in some way by either opposing or protesting government policy, and software piracy is a convenient excuse in Russia for law enforcement to bash up political opposition.

Microsoft grants NGOs a certain amount of free copies of their software, and these Russian organizations were within their legal right to use it. Nevertheless, Microsoft has been slow to clear the reputation of these maligned NGOs, as McLaughlin notes.

Perhaps as a result of bad publicity, Microsoft has just started to take aggressive measures to make sure these fake charges of piracy do not happen again. They have changed their NGO license policy to make it explicitly clear that organizations such as the Russian NGOs that were raided by the government are not in committing theft if they use the free copies of Microsoft programs that are allotted to them.

While this may stop the bad publicity and also prevent the Russian government from using Microsoft as a political tool, the root cause of the problem will remain. If corrupt officials in Russia can’t use the software piracy excuse to raid the offices of dissident organizations, they’ll find some other reason.

Will China’s CPUs soon overtake Intel and AMD?

Posted on by
Reply

Photo of an old processor under CC license courtesy of Andrew Dunn (http://www.andrewdunnphoto.com/) and referenced on Wikipedia (http://en.wikipedia.org/wiki/File:Intel_80486DX2_bottom.jpg)Like the antiquated American cars still driving on Cuban roads, China’s economy still runs on Five-Year Plans. With China’s growing wealth, the last few Five-Year Plans have been able to fund projects from the Chinese space program to developing green energy.

The latest Five-Year Plan is named China Next Generation Internet (CNGI). China is seeking to broaden its use of and role in the Internet community. Part of its strategy is to foster a homegrown industry of hardware and software and perhaps to eventually lead the markets in those areas.

One key industry that the Chinese government is focusing on is the processor market. An article in EE Times describes the new generation of Godson CPUs that are being developed under the leadership of the Beijing Institute of Computing Technology. The next generation is due out in 2011, but will still be a bit behind the market currently dominated by Intel and AMD.

The Godson 3B microchip will have eight cores and 65-nm parts, be 64-bit, and will incorporate vector processing (read more about the difference between scalar processing and other forms such as probability processing). Although most current processors on the market use less-robust scalar processing technology, China’s new CPUs will still be behind in the number of processor cores when they hit the market in 2011.

But the future for Chinese computing looks brighter. After 2011, the next generation of Godson CPUs will skip over the current standard of 45nm/32nm process technology and use 28nm instead. But some are skeptical that China can reach this benchmark. And once it does, perhaps Intel and AMD will be there too.

Nevertheless, China is an emerging player in the international CPU market, and, if it delivers on its promises, stands a good chance of competing against the big dogs: Intel and AMD.

Intel buys McAfee: a bold move to stay competitive

Posted on by
Reply

Last week, to the surprise of many in the IT industry, Intel bought the computer security company McAfee. In an article on CIOUpdate.com, Larry Barrett insightfully describes the business context in which this acquisition was made.

During the recession that began in 2008, and even before it, technology companies have been focusing on innovating and diversifying their services. Google and Apple have been long-time leaders in this trend, especially in the mobile revolution. Oracle and Cisco, Barrett notes, have not been far behind in branching out of their traditional offerings.

Now, Intel is feeling the pressure of several years’ worth of low stock prices despite its considerable advancement of CPU technology and consistent quarterly profits. It seems that investors are not valuing these short-term gains and have feared that Intel is falling behind in the business arena.

Enter Intel’s last week purchase of McAfee, Inc. With this move, according to Barrett, Intel hopes to develop the technology of computer security at the microchip level, despite the current standard of protecting networks and programs at the operating-system and software realms. More importantly, Intel hopes this relatively new approach will catch on in the mobile market, where the digital security industry is having a hard time coping with the tremendous output of new devices and applications.

This is a risky move, as many commentators are skeptical of the very idea of security at the micro-hardware level. But if this gambit pays off, Intel could well be years ahead of the competition in providing secure processing power to smart phones and other mobile devices. Even Intel’s competitor, AMD, had struck outside its strict business boundaries when it purchased ATI Technologies in 2006 and claimed a stake in the GPU market.

As Barrett’s article makes clear, Intel had no choice but to make a bold move. In a business environment where doing strictly what you do well is not enough (in the eyes of investors), a company needs to change. This is a good thing, it seems to me. Imagine how much we would have missed out on if Google had just remained a search engine company or Apple had just continued to build outdated desktop PCs.

Yesterday’s (May 6) stock plunge and computational finance

Posted on by
1

For those following financial news, yesterday was a crazy day on the stock market. Financial markets around the world have already been on a perpetual roller coaster ride because of sagging economic growth in Europe, especially in Greece. But yesterday, the DOW took its single largest plunge in history – an almost 1,000 point-loss (about 10% of the market) in less than half an hour.

Analysts are straining to figure out what exactly caused this fall. Rumors and conjectures have been swirling around the media for lack of better information. Although the stock market recovered most of the value it had lost before closing on May 6, people rightly want to discern what were the causes of the brief crash.

Tim Paradis of the Associated Press reports some of the current explanations for yesterday’s price fall. One idea is that a trader mistakenly entered $16 billion instead of $16 million into a sell order. Another possible trigger is a mis-listing of the price of Procter & Gamble stock, which caused massive selling of that stock and a subsequent 37% drop in its price.

After listing these theories, Paradis concludes,

Whatever started the selloff, automated computer trading intensified the losses. The selling only led to more selling as prices plummeted and traders tried to limit their losses.

“I think the machines just took over. There’s not a lot of human interaction,” said Charlie Smith, chief investment officer at Fort Pitt Capital Group. “We’ve known that automated trading can run away from you, and I think that’s what we saw happen today.”

In a nutshell, when financial computers detected a significant drop in the stock market (be it from the erroneous $16 billion dollar sell order or from mist-quoting Procter & Gamble stock price), they saw this as a sign that stocks were losing value. Then, they automatically sold shares, which caused even more devaluation and selling.

While it is true that inadequate computational models may have been partly to blame for both the recent recession in general and yesterday’s price plunge in particular, it is important to keep this criticism in perspective.

Computational finance is still an industry in its infancy, and like with all human devices that fail at some time, it’s no use blaming the device for the failure. We should blame human error. For example, if a bridge collapses somewhere, we don’t blame the physical structure itself; we seek to find the faults in our own engineering plans that caused the accident.

Computational finance, using mathematical algorithms to make sense of the stock market, offers much promise in making sense of the seemingly-chaotic world of financial trading. As computers grow more powerful, they will be able to answer the most difficult questions in finance, and the stock market may cease to seem so arbitrary. This industry should be allowed to develop, but with sufficient checks in place and smarter algorithms to prevent shocks like the one that happened yesterday.