Who is Mechanical Turk?
I’ve been reading a lot lately about Amazon Mechanical Turk. What’s that you ask? Good question.
Mturk, for short, is Amazon’s human-powered marketplace. Just like you can provision servers on demand, with Mturk you provision small work units that you pay people to complete. These are often small tasks, such as a web search or transcribing a few seconds of audio, and for the work that you do, you get paid, from a penny per unit on up. If you’re concerned about the quality of the responses you get, you can assign each task to multiple people, so it’s done two or three times and you can verify the quality of the response.
Here’s a quick video about the service from the user perspective.
Anyone can also submit work units to Amazon to be completed. Here are three really interesting blog posts about people’s positive experiences with Mturk.
NewsCred.com used the service to help categorize RSS feeds.
iamelgringo.com used the service to verify 6,000 business URLs and addresses. His project was completed in 5 days at a total project cost of $300.
Andy Baio used Mturk to transcribe an audio interview, breaking it down into small chunks for easy transcribing. Transcriping 36 minutes of audio took only 3 hours and cost $15.40.
Are there uses for this type of service in higher ed? It sounds like it could really help on grunt work or large sets of data that need verified or cleaned up.
If you enjoyed this post, please subscribe to my RSS feed!
S3Stat tracks your cloud usage
One of the downsides of Amazon’s S3 service is that the reporting stats you get back are, to the naked eye, tough to decipher. It’s a spreadsheet full of times and values and bytes. It doesn’t give you a clear idea of performance, trends or popularity of items.
Enter S3Stat.com.
Every night, they will parse your S3 logs and transform them into Apache logs, which can be run through Webalizer and output back to an area on S3 where you can analyze to your heart’s extent. The stats are ongoing, so you can easily go back and check out last month or earlier in the year, which is nice. Here’s a quick screen grab:
S3Stat Screenshot
The one caveat is that you give S3Stat your S3 login and pass. That allows them to grab you logs and post the results back to your S3 account. This is similar to giving your login to RightScale to manage your EC2 resources. S3Stat has done nothing to question my trust in them, so we’ll continue to use the service.
Signing up for the service is easy, and you get a free 30 day trial. After that, the service is miniscule $2 a month. That’s a small price to pay for not having to wade through Amazon’s default stats.
If you enjoyed this post, please subscribe to my RSS feed!
Amazon Announces…something
This morning, Amazon Web Services announced a new content delivery service, but hasn’t given much in the way of specifics, costs or details.
It’s interesting they’d create a content delivery network since that’s what many people, myself included, are using S3 for. Perhaps this network will have less latency and better response times then S3 sometimes has.
In an email sent to customers, Amazon writes:
You’ll start by storing the original version of your objects in Amazon S3, making sure they are publicly readable. Then, you’ll make a simple API call to register your bucket with the new content delivery service. This API call will return a new domain name for you to include in your web pages or application. When clients request an object using this domain name, they will be automatically routed to the nearest edge location for high performance delivery of your content. It’s that simple.
From the email, the new content delivery service sounds separate from S3. Here’s what we do know.
1. There are no minimum fees and no commitments to use the new content delivery service. You will only pay for what you actually use.
2. A single, simple API call is all that is needed to get started delivering your content.
3. Works seamlessly with Amazon S3 – this gives you durable storage for the original, definitive versions of your files while making the content delivery service easier to use.
If the only difference between this CDN and S3 is a slight increase in performance at an additional cost, I don’t see why I should stop using regular old S3 as a delivery tool. How about a little more direction, Amazon.
If you enjoyed this post, please subscribe to my RSS feed!
