Author Archives: Nathan Hinkle

SSDs: Speeding Up Storage

06 Tuesday Dec 2011

Posted by in Reading Summary

≈ 1 Comment

This is a summary of “Engadget Primed: SSDs and you,” the “going further” reading from November 7th, when we talked about hardware enhancements. Although the main focus of the article was on modern advances in solid state storage, it started by chronicling previous important developments in the history of data storage. The article is very lengthy, so I’ve tried to summarize it as briefly as I can while still getting the important points, but this post will be a long one.

One of the first systems it mentioned was IBM’s RAMAC, which we also discussed in class. The system included IBM’s first disk storage unit, with a whopping storage capacity of 4.4 MB. We’ve come a long way, in an era where a 60GB disk is considered tiny, and 1TB is par for the course.

IBM 350 Disk Storage Unit - Engadget

After recounting the early players in mechanical storage, the article goes on to explain the technology behind modern mechanical drives like the ones still in predominant usage today. Surprisingly, at its roots the technology is barely different from that in the first IBM hard drives. A hard drive has one or more spinning magnetic platters. Bits (0’s and 1’s) are recorded based on how each magnetic grain is polarized. The primary difference between early rotary hard drives and newer models is the speed at which they rotate (the IBM 350 spun at 1200RPM, modern hard drives typically spin at 7400 RPM), and most importantly, the data density. There is a limit to how much data can physically be stored in a given area of a magnetic disk though:

Eventually, though, magnetic storage runs into fundamental laws of physics. In this case, those immutable rules are represented by the superparamagnetic effect (SPE). Once we shink magnetic grains below a certain threshold, they become susceptible to random thermal variations that can flip their direction.

Essentially, traditional hard drives are physically running out of space to store any more data. Manufacturers have pushed the limits to about 3TB, but at some point, it’s not possible to store more data without drastically affecting performance.

Finally, after working through the background information, the article delves into SSDs and how they work. I’ve been using an SSD for about 6 months now, and had read some articles about them previously, but hadn’t learned about the inner workings in nearly this much detail. I suggest you read the article, because I can’t possibly fit all of the details into this space, but here’s an overview.

So how does flash work, and what makes it different from traditional magnetic drives? The short answer is that instead of storing data magnetically, flash uses electrons to indicate ones and zeroes. You might already recognize why this is a plus: no moving parts. That means no noise, no head crashes, and greater energy efficiency since you don’t have to move a mechanical arm. And unlike DRAM, it’s non-volatile — it doesn’t need constant power to retain information.

Non-magnetic storage has actually existed for many years, and has previously been used in specialized applications such as space probes and data acquisition systems for oil exploration. The first consumer-targeted flash storage for use as a storage disk on a regular computer showed up around 2005 in a Samsung laptop. With 32GB of flash storage, the laptop cost almost $4000. If you thought SSDs are expensive now, think again.

The article goes on to explain in-depth how the underlying physics of solid state storage work. To summarize a few of the most interesting points:

  • There are two types of flash memory: SLC and MLC (single level cell and multi level cell). MLC memory can store twice as much data in a given amount of space, but takes longer to read and write. MLC also degrades faster.
  • SSDs slow down over time as they fill up with data. This has to do with how the memory cells are wear leveled and how data is actually written to the SSD. Some SSDs actually ship with extra space built in (which is used by the device but not reported to the operating system) to account for this.
  • SSDs wear out relatively quickly. Most MLC-based flash has a limit of around 100,000 cycles. This limit can be reached in as little as a year. An article on Coding Horror about the hot/crazy scale of SSDs recounts numerous SSD failures, with none lasting more than 2 years.
  • SSDs are more power efficient than HDDs because they have no moving parts
  • The controller chip being used has a huge impact on performance. Early SSD controller were often low-quality, resulting in poor performance and longevity.

SSDs are still a rapidly evolving technologies. Just recently have prices started to come down into the $1/GB range. As the technology advances, it will become increasingly affordable – right now, SSDs are primarily used by developers, gamers, and other power users, but they are starting to make their way into the mainstream, especially when integrated with consumer products like tablets and ultralight laptops. The big question that remains is, are they worth it?

In my view, not quite yet. As the price continues to drop, SSDs will soon reach the point where they are economically feasible for everyone, but presently, it’s impractical to store large amounts of data on them. A popular choice right now is to have a relatively small (60-160GB) SSD with boot files, applications, and some working data on it, and a secondary mechanical hard drive to store large files for which transfer speed is less important. This is a challenge in laptops though, where space is at a premium and it’s often difficult to fit both a hard drive and solid state drive into one machine. The solution I use is an optical bay caddy: I’ve removed the DVD burner from my laptop, and put a hard drive caddy in its place. I have a 128GB SSD as my primary drive, and store large files on the hard drive. Some laptops are now shipping with a hard drive and a small SSD both built in, as well.

Solid state storage is a fascinating technology, both for its physical underpinnings and the effects it’s having on the computing industry. I foresee a time in the near future when SSDs will be standard fare, and we will all wonder how we lived without them. For those of us who’ve already had a taste, that question has already presented itself.

Computers: The Ultimate Machines of War

31 Monday Oct 2011

Posted by in War is in the air

≈ 1 Comment

War is the driver of history – many of our most prolific inventions have been the result of advances in military technology. Wars have influenced the development of computers in particular, both necessitating their advancement and at times delaying researchers’ progress. The earliest computing machines were used primarily for generating mathematical tables which could be applied to nearly any industry, and for tabulating massive amounts of data. Military leaders quickly recognized how computers could be applied to their needs though. One area in which this was particularly evident was with cryptography – the business of encrypting and cracking secret messages. The enigma machine was one of the first electromechanical devices used for the encryption and decryption of secret messages. Built by German engineers Arthur Scherbius and Richard Ritter in the 1920s, it used a series of rotors with integrated circuits to route keypresses through the machine, encoding each letter. After each letter had been encoded, it would advance the rotors, so each letter in the message would be encrypted differently. With various other mechanisms to complicate reversing the code,  there were about 10,000,000,000,000,000 possible combinations (Singh, pg. 136). The enigma machine was used by Nazi Germany in World War II to encrypt nearly all of their radio transmissions. Deciphering this information was of crucial importance to the British military, and an entire campus at Bletchley Park was established as a base of operations for the cryptoanalyists. One of the foremost researchers, Allan Turing, developed his own modifications of the enigma machines, designed to brute-force the ciphers based on known pieces of information. These machines, which he dubbed “bombes”, were another significant step in the progress of computin. By the end of the war, 49 of them were in use at Bletchley Park (Singh, pg. 181). War did not solely advance the progress of computers though. With significant resources being put into fighting World War II, available computers were almost exclusively purposed for wartime calculations. An example of this was the Harvard Mark I, often regaled as the first ever fully functional and stable electronic computer. Built by IBM under the direction of Howard Aiken, it was donated to Harvard University for their use in research.

When the Mark I was completed in 1944, IBM gave it to Harvard as a gift. That spring it was installed at the university but was immediately leased for the duration of the war by the US Navy, desperate for gunnery and ballistics calculations. Aiken, a naval reserve officer, was put in charge of the Mark I for the Bureau of Ships – Williams, pg. 112

Although the actual construction of the Mark I was completed rapidly, it was quickly commissioned by the Navy, and researchers at the University were shortchanged of the opportunity to truly take advantage of these new computing resources. Had it not been for the war, computer research may not have been quite so rapid, but the technology would have gotten into the hands of civilians much faster.

The role of women in the history of computing goes back as far as Babbage’s era. Frequently, women were employed in the task of completing the manual calculations (human “calculators”) for men’s work in engineering, astronomy, and other fields. The division was largely a matter of sexist beliefs that only men should do the actual innovation, but that manual computation was a waste of their time (Ceruzzi, pg. 240). Nevertheless, jobs as “computers” were very popular with women, as they were still a step up from the common secretarial work which was oftentimes the only other option. The women who took these jobs were often very proud of their work, for though the labor was menial, it did require significant mathematical abilities (Ceruzzi, pg. 239).

It is therefore unsurprising that many of the first computer “programmers” were women as well – it was merely an extension of the existing tradition of men determining what needed to be calculated, and women executing the calculation. Many of the women programmers were hired because they were mathematicians – one of the few fields women could study at the advanced level – even though they had absolutely no formal training with computers (Williams, pg. 113).

Computers are crucial in modern warfare, to a far greater extent than Aiken or Hopper could have ever predicted. Fighter jets are flown by advanced electronics, computer-powered satellites beam high-resolution imagery to military intelligence agencies, and cyber-espionage is of increasing concern for governments worldwide.

One of the areas in which military technology has come to depend extensively on computers is with unmanned aerial vehicles (UAVs), often referred to as “drones”. The United States Air Force and other government agencies have been using UAVs since the mid 1990s. Originally, they served for reconnaissance missions, but have since been adapted with powerful missiles and other weapons (USAF, 2010). Some have questioned the ethical implications of such dangerous weapons being controlled by soldiers who operate them across the world from where the fighting occurs. Most of the approximately 700 Predator Drones in Iraq are controlled from an Air Force Base in Nevada, where soldiers may have limited perception of what is actually happening on the ground when they fire their missiles (Harris, 2006).

Another area in which computers are of increasing importance in modern warfare is not on the battlefield, but with cyberwarfare. A highly publicized example was the Stuxnet virus, which infected industrial control systems running uranium enrichment systems in Iran. It is unknown to this day who was responsible for the virus, but investigations have implicated both the Israeli and American governments as possible perpetrators of the attack. The virus caused centrifuges to spin out of control, resulting in serious damage to their enrichment systems. This problem set back their nuclear program by years, which may have disrupted Iran’s attempts to build an atomic bomb. With various nations establishing specific cyberdefense units in their military, this type of sabotage may become the next major forefront of global wars.

John von Neumann was quoted as saying,

If you say why not bomb them tomorrow, I say, why not today. If you say at five o’clock, I say why not one o’clock. – Rheingold

In this quote, von Neumann is referring to “them” as the USSR. He was very closely involved with the development of atomic weapons before and during the cold war, and was a strong advocate of a preventative strike against the USSR. This was at a time of intense fear of global war, and von Neumann believed that if the United States did not attack the USSR first, they would surely be decimated. Nevertheless, his views were seen by many as being extreme. A majority of the American population was not advocating for a preemptive attack against the USSR. Looking back at the cold war era from our current vantage point gives us perspective on just how devastating an all-out nuclear war would have been – which von Neumann’s proposed attack surely would have provoked. Yet had I been alive in that time, it is difficult to say whether I would have agreed or disagreed. To be sure, there was significant fear of the Soviets, but as a generally pacifistic individual, I believe I would have advocated against starting a war with such a formidable opponent.

 

Sources:

CERUZZI, P. E. (1991). When Computers Were Human. IEEE Annals of Computing History, 13(3), 237-244.

Harris, F. (2006, June 2). In Las Vegas a pilot pulls the trigger. In Iraq a Predator fires its missile. The Telegraph.

Rheingold, H. (2000). Johnny Builds Bombs and Johnny Builds Brains. In Tools for Thought.

Singh, S. (2000). The Code Book: The Science of Secrecy from Ancient Egypt to Quantum Cryptography.

USAF. (2010, July 20). MQ-1B Predator. Retrieved October 29, 2011, from US Air Force Information Factsheets: http://www.af.mil/information/factsheets/factsheet.asp?fsID=122

Williams, K. (n.d.). Improbable Warriors: Mathematicians Grace Hopper and Mina Rees in World War II.

 

The Transcontinental Calculation

17 Monday Oct 2011

Posted by in Alternate History

≈ Comments Off on The Transcontinental Calculation

In the first half of the 1800s, getting from one side of the United States to the other was a significant affair. There were no planes back then (of course), no cars, and the overland route via covered wagon was even more treacherous than the video games of our youth alluded to. In 1863, workers broke ground on the US transcontinental railroad. By 1869, one could ride from Nebraska to California in a week, instead of the six treacherous months previously required. It was one of the most remarkable feats of civil engineering – not to mention sheer labor – of the 19th century. Not only did the railroad unite the country with transportation, it also enabled a new era of communication: telegraph lines were installed next to the railroad, allowing messages to be sent instantly across the country.

One can imagine the number of calculations required to build a single trestle, let alone an entire 1780 miles of railroad. Only about 25% of the workers involved in the project were actually physical laborers doing the blasting, digging, and other heavy work. It is some of the other workers who would have most benefited from access to a difference engine.

For one, there is the obvious need for engineers to perform calculations regarding where it is most efficient to lay the route, how strong bridges must be to support the trains, how far it can be between refueling points without trains running out of coal and other supplies, and countless other mathematical problems. Indeed, any engineering project in that era would have benefited greatly from access to more accurate and varied tables of numbers.

A less immediately obvious, but undeniable application for a difference engine is all of the accountants and workers responsible for ensuring sufficient supplies. Given a certain number of expected miles of construction, how many railroad ties does one need to order? How much rail? When should you send the shipments of materials to optimize the number of trains you send, without losing valuable work time for lack of parts? Indeed, the benefit to accounting and management might have been greater than the advantages for the engineers.

Had the difference engine been available in 1863, would it have had any lasting impacts in the context of the railroad, or would it merely have eased the burden on overworked engineers and accountants? It’s hard to say. Even if use of a difference engine had made it cheaper to construct the railroad, it might not have been completed any sooner – the primary delays were caused by bad weather and treacherous conditions. Where the tables derived from a difference engine might have been more useful would have been for the hundreds of people starting new businesses in the recently opened up territories of the west. The railroad lead to a massive expansion of the population in the western US, and to be certain, many of them would have found the tables that a difference engine would have made available to be quite handy.

Wednesday 10/12: Cryptography

13 Thursday Oct 2011

Posted by in Class Summary

≈ 1 Comment

Class started with a reminder that homework assignment #2 is due Monday, and due Wednesday a selection from the IEEE archives for the final project.

The anonymous self-written tests from our previous class were discussed, and praised for their depth and insights. We expanded on an answer to “why did Hollerith use punched cards instead of tape”, noting that cards could be resorted and accessed randomly, while tape constrains one to always reading data in a particular order. This is analogous to the difference today between “random access” and “serial access”.

RMS LusitaniaNext, we went back in time to World War I. In May 1915, the RMS Lusitania, a British merchant ship with many American citizens aboard, was sunk by German U-boats. This was before America had entered World War I. Germany was starting to flex its naval muscles with its U-boat fleet. In 1917, Germany sent the famous Zimmerman Telegram to its ambassador in Mexico, seeking to form an alliance with Mexico. Germany would assist Mexico in reclaiming former Mexican territory including Texas, New Mexico, and Arizona; in return, Mexico would form an alliance with Germany, attacking the United States to prevent them from engaging Germany in Europe. British intelligence intercepted the note during its initial transmission, and it was decoded by their cryptographers. The note was published in American newspapers, purporting that it had been intercepted by spies when the note was in (unencrypted) transmission via telegram across Mexico. The revealing of this information prompted America’s entrance to the war. It wasn’t until 1923 however that Churchill admitted that British cryptographers had in fact intercepted and decoded the message – throughout the war, the Germans had no idea that their private communications were being intercepted.

Enigma machine in use

An enigma machine being used in Russia. - German Federal Archives, via Wikimedia Commons

This dramatic breach of security for the Germans led to the development of the enigma machine. Over 30,000 of the machines were built during the course of World War II. The mechanisms of this machine were the subject of our reading for today, an excerpt from The Code Book.


To get a better sense for how the enigma machine operates, we used paper enigma simulators to decode a secret message on the board: MPXNCZJA. The scramblers were arranged in order of 1-2-3, and the day’s key was MCX. Recall that the first three letters are the message key, leaving NCZJA as the message. Decoding the message and taking it to the honors college office was rewarded with a prize.

We wrapped up our discussion by reflecting on what ultimately led to the failure of the system: not the encryption itself, but its human users. The use of repeated, predictable phrases, easy-to-guess message keys, and other patterns to exploit led to the messages being decrypted far faster than they would have been if better practices had been followed. As with many security systems, the weakest link in the chain of security is not the system itself, but those who must operate it.

RIP Steve Jobs

06 Thursday Oct 2011

Posted by in People

≈ 1 Comment

Most of you have probably already heard this news, but I think it’s worth mentioning in the context of this class. Steve Jobs has been very influential in making computers accessible to non-technical users, improving user interface, and bringing computing to mobile devices. He passed away Wednesday morning at age 56. Though I’m not personally an Apple user, I deeply respect Steve Jobs for the tremendous impacts he has made on the evolution of personal computing. Rest in peace.

RIP Steve Jobs: 1955-2011

Speeding through the series of tubes

03 Monday Oct 2011

Posted by in Personal History

≈ Comments Off on Speeding through the series of tubes

My Personal Computing History

The first time I ever used a computer was in first grade, sitting in the back of our classroom at the row of then-new, now-ancient Apple Macintoshes. While most of my classmates seemed enthralled by video games, I was already looking for more “productive” uses. My aunt was one of the few people my family knew with a computer and “The Internet”, so I would send her emails from the school computer. I was fascinated by the ability to send messages such long distances so quickly. I still hadn’t quite grasped the concept of The Web, but was already learning to value electronic communication.

Macintosh PowerPC

The first computers I used were PowerPCs like this one

We got our first computer when I was in about second grade. I went with my dad to browse the options at a local store. Nowadays I make all of the technology decisions in the house, but at that point I was just tagging along for fun. We ended up with a Compaq beige box, loaded with Windows 98 and Microsoft Works. With 32 MB of RAM, a 5 GB hard drive, and a 360 MHz processor, it wasn’t anything to write home about, but was enough to write emails. I didn’t really use the computer much at all for the first few years. My dad cajoled me into learning to type with JumpStart Typing for Kids, the first video game (of sorts) that I ever played. I thought it was incredibly lame, and quickly figured out that you didn’t have to type accurately to win the “free-form typing mode” challenges which let you level-up, you just had to type fast. By holding down the a key long enough, I could win every level. While the game didn’t really serve its intended purpose, I suppose it still provoked me to use some creative thinking skills.

The home screen for the JumpStart Typing game

When I got to middle school, I started to develop a real interest in computers. I quickly became the go-to person for fixing problems, and got my first taste of programming with a lego robotics class. We got our first good computer when my grandparents bought us a Dell Dimension for Christmas. Over the years I’ve upgraded it several times, but it’s still humming along as our primary home computer, and runs Windows 7 smoothly on its Pentium 4 HT. In high school, I finally got my own laptop. With my new-found freedom and the seemingly undefeatable speed of a dual-core processor and a whole gigabyte of RAM!!!, I took off exploring how to put my shiny new machine to use. Today, my desktop computer has a whopping 24 GB of RAM, four times as much as our first computer’s hard drive space, and can chew through anything I throw at it. One has to wonder where it will end. If I have kids some day, what will their first computer look like? Will it be a small device with nearly infinite resources? Or will everyone be using tablets with less physical capacity, and everything running remotely from the cloud?

The impacts of computers on Me aren’t just about Me

Wikipedia LogoOne of the biggest ways that computers and the internet in particular have influenced me is with the ability to collaborate with, meet, and learn from other people. I had my first experience with online forums while trying to figure out how to modify the login screen on my laptop. After getting my question figured out, I ended up sticking around and learning more about computers than I had anywhere else previously. I spent some time editing on Wikipedia, before getting tired of the intense bureaucracy there. Super UserThe site I’ve been the most involved with is the Stack Exchange network. Founded as the programming site Stack Overflow, it’s now grown to encompass everything from English Language to Gardening. Stack Exchange is a Q&A platform, and addresses many of the shortcomings of traditional computer forums. I’m most involved on the computers site in the network, Super User. The system runs on a slightly addictive incentives system with “reputation points” (I’m wont to admit that I’ve got over 12000 of them now). Unlike other sites though, the points mean something – increased reputation unlocks various privileges, which allows for community moderation, from editing to closing posts to even deleting questions. Profile for nhinkle on Super UserMost moderation is done by regular users, although there are a handful of community moderators. In another unique twist, moderators are elected by a democratic voting system, not appointed by site admins. I was elected to be a moderator this spring, and am still actively involved on the site. In an increasingly interconnected world, it’s important to recognize the worldviews, lifestyles, and perspectives of people from other cultures. I talk with a programmer in the Netherlands. I read essays from students researching biodiesel in Brazil. I’ve made friends with a deaf diabetic from the midwest. The internet was founded on the basis of spreading knowledge and connecting the world, and I feel privileged to have taken a part in that experience.