College kids everywhere collectively scream:

Netflix’s options include limiting the number of people or computers that could share one account or charging extra, for example, to let a college student in another town use his parents’ account. The company today doesn’t actively restrict the number of PCs, tablets or phones that can use one account or the number of people who can sign in, though it limits each account to playing two video streams simultaneously.

More at Bloomberg

With speculation that Lenovo is looking to purchase IBM’s commodity server business comes questions about the potential reactions by the US government and regulators. The Lenovo acquisition of IBM’s PC business was approved less than a decade ago. But at roughly the same time, China National Offshore Oil Corporation’s attempted buyout of Unocal came under Congressional scrutiny partly because of security concerns. Lenovo is no Huawei, but Huawei’s label as a “national security threat” by the US House Intelligence Committee may further stoke the inevitable opposition from a Chinese company’s purchase of one of the largest server manufactures by market share.

Or maybe not. The NYT reported in 2006 (after Lenovo’s PC acquisition) that:

The State Department will use the 16,000 desktop computers it purchased from Lenovo, just not on the computer networks that carry sensitive government intelligence.

If you look at the PCs from photos taken in the White House Situation Room, they all appear to be HP. However, my understanding is the Lenovo ThinkPad is still a workhorse for the federal government.  Wouldn’t be surprising if the same rule cited in the article apply to servers as well.

Job postings for the top five programming languages have shrunk in both absolute and relative terms according to the job trends tool on indeed.com.  Stephen O’Grady from Redmonk, an analyst firm, finds the top programming languages every quarter using a methodology developed originally by Dataists. O’Grady counts the number of Github and StackOverflow  projects for each programming language and then ranks the totals.

The top five programming languages for the first quarter of 2013 are, in order: Java, Javascript, PHP, Python, and Ruby. Matching jobs listed on Indeed, which aggregates open positions, for those languages declined as a percentage of overall jobs:

Job listings also shrunk in terms of relative growth in the past few months (Ruby is missing since it underwent large growth that drowned out the others, but the pattern was the same):

The change doesn’t appear to be seasonal or part of an algorithm change at Indeed. The trend has cratered in the past few months and I’m not exactly sure why. Any clue?

Adrian Cockcroft, cloud architect for Netflix, has a new deck up at slideshare titled “Netflix and Open Source,” which has a bunch of great facts about Amazon’s infrastructure as a service (IaaS) cloud. It also spells out how Netflix uses the technology, in its view, to obtain a competitive advantage in the market.

Much of the potential savings in cloud comes from the ability to turn it off. Netflix, for those in countries where it is not available, offers a service that allows customers to watch videos instantly over the internet. You can imagine how the usage of the service spikes after work hours and then fall back down as people go to sleep. Looking at slide 20 in the slide deck, that is exactly what happens. Netflix is renting a little over 200 compute instances (like servers) from Amazon at the low point (after midnight) and over 600 compute instances during the peak evening hours.

Other interesting facts:

• Zero to 500 instances in a little less than 8 minutes. (slide 19)
• The average lifetime of an instance is 36 hours. (slide 20)
• Amazon, based on public IP addresses, has a max capacity of 3.7 million instances, and has been more than doubling in size every year – faster than Moore’s law. (slide 22)

Perhaps the most interesting slide is how Netflix views itself compared to traditional IT:

Cloud isn’t simply a placeholder for future road map or a test bed, it is fundamental to the business strategy.

The whole slide deck is interesting and you should have a look. At the end, Cockcroft talks about a new competition that Netflix announced, with cash-money-prizes, to develop programs centered around the cloud. 

Bitcoin is currently valued at over $100 USD (via http://bitcoin.clarkmoody.com/) 

The rapidly growing value in terms of United States dollars (USD), and possible future volatility in the exchange rate, make bitcoin an unappealing option as a medium of exchange or a unit of account – two primary functions of money.

Even if bitcoin is not a unit of account as Tim Lee contends, the markets in which items can be priced in bitcoin are extremely limited. Essentially, the markets need to adjust the price of listed goods as rapidly as the exchange rate so that the goods do not devalue. For example, if you want to receive $1 (USD) for a beer and the current exchange rate is $1 USD for every 1 bitcoin then you would put the beer on the beer bitcoin market for 1 bitcoin. If the exchange rate shifts to $1 USD for every 2 bitcoin but the price on the market for the beer is still 1 bitcoin then you are potentially only getting $0.50 for the beer. Not many markets can change pricing this quickly. A lot of billing – say for electricity – is done monthly and isn’t settled up until a month after the statement is given to the customer. Bitcoin is useless in those cases.  

In practice, there is going to be a lot of currency exchange risk as well because transactions do not happen as rapidly as exchange rate fluctuations. It takes time to exchange US dollars to bitcoin and make a purchase. The exchange rate can jump up and down in that time.  

Also, bitcoins are often pitched as a currency system that is independent from central bank controlled money supply. If all exchanges and purchases are happening instantaneously though, then I fail to see how it lives up to that hype. 

Amazon Web Services (AWS) released yet another feature to win more enterprise clients. This time it was a hardware security module (HSM), an appliance for storing keys, but in the past it has been virtual private clouds, single tenant compute options, direct connections with co-location providers, etc. Here are two thoughts on Amazon’s rapid product growth:

• Products aren’t perfect and pointing out the shortcomings of offerings is part of an analyst’s job. AWS, though, drops new tracks faster than Jay-Z in his retirement. In the past, I have critiqued cloud computing providers for not publicly supplying information on industry standard data center audits. Amazon now has that. I have said, repeatedly  that performance benchmarking is needed. Amazon now has guaranteed IOPS,  a dude dedicated to benchmarking, and they joined SPEC.org. That isn’t disappointment. It is a tip, to myself included, that when reading a critique, it may be dated even if it only came out months ago.

• Building a business around a, yet, un-offered AWS feature may prove more difficult. Amazon has taken the DIY approach and hasn’t made any acquisitions. That eliminates what would be a key exit strategy for venture capital. And, unlike some build-don’t-buy failures by others, AWS’ appurtenances to its core product get uptake from customers. There is a filtering process as well. If your business goes gangbusters because it is popular that means there was demand, and Amazon has been responding rapidly to demand for new products. That doesn’t mean tackling feature gaps in AWS can’t be a successful business, there are already notable exceptions. It does mean there is an extra complication.