MoovAtom Version 2 Player and API

We are proud to announce that we've released our latest API and Player. The API now lets you POST jobs to it just like you would a form, and edit the player parameters remotely. We've also added support for JSON on top of the existing XML support.

The new player has a much better interface, better sharing options, and has better branding capabilities. We haven't even scratched the surface on all the player editing options, but our clients now have a much better foundation to deliver from.

Here is an example of the player:

Hackday.TV Video and Presentations

Here is some video from the Hackday.tv event. The video focused hackday had some of the best hacks we've seen come out of a 24 hour sprint. Words2Clouds used MoovAtom's API to quickly incorporate streaming video, we brought some cakey goodness for all the hackers, and made tons of friends along the way. All in all a testament to the innovation going on in New York City right now. Big thanks to General Assembly for hosting the event, and Shelby.tv for organizing the madness.



And now for some awesome pictures!

Hackday.tv / General Assembly / API

MoovAtom released our API beta last month and will be launching it at Hackday.tv at General Assembly in NYC September 10th.

You can check out the API by clicking here

Edit: Live at the event! Check the image posted in Nasdaq's instagram feed.

60,000 Streams / Updates

MoovAtom has passed a milestone of 60,000 streams delivered. Not bad for only being out of beta for a month. We've done single events with 1,200 simultaneous streams and 600 GB of data streamed over 2 hours, and have tons of videos and channels being added to our system every day. We're now delivering video for Ad networks, e-commerce platforms, the HR industry, Insurance companies, and film makers, and we're growing!

You can also now sign up online through our automated payment gateway, and you can cancel any time without the need for human intervention. You decide how much to use and only pay for the data you stream. No contracts! Simple right?

Check us out here for more info:

MoovAtom

Nginx vs. Django/Apache for serving static files.

This goes without saying really and is a relatively well known nugget, but this post shows the quick and dirty of nginx blowing the pants of django/apache/mod_wsgi for serving static images from a small EC2 instance. Serving static files with django is strongly advised against, so if you're doing it, change the front end stuff over to nginx or lighttpd and proxy back to django/apache.

command: ab -n 10000 -c 200 http://testdomain.com/static/test.xml

Apache/Django - 303.51 req/sec

Document Length: 3247 bytes
Concurrency Level: 200
Time taken for tests: 32.948 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 34760000 bytes
HTML transferred: 32470000 bytes
Requests per second: 303.51 [#/sec] (mean)
Time per request: 658.953 [ms] (mean)
Time per request: 3.295 [ms] (mean, across all concurrent requests)
Transfer rate: 1030.28 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 5 67.5 0 3001
Processing: 83 644 1668.2 323 16566
Waiting: 83 637 1668.2 320 16566
Total: 84 649 1669.9 326 16571

Percentage of the requests served within a certain time (ms)
50% 326
66% 407
75% 502
80% 580
90% 819
95% 1096
98% 8411
99% 9415
100% 16571 (longest request)


Nginx - 4705.07 req/sec

Document Length: 3247 bytes
Concurrency Level: 200
Time taken for tests: 2.125 seconds
Complete requests: 10000
Failed requests: 0
Write errors: 0
Total transferred: 35271448 bytes
HTML transferred: 32471168 bytes
Requests per second: 4705.07 [#/sec] (mean)
Time per request: 42.507 [ms] (mean)
Time per request: 0.213 [ms] (mean, across all concurrent requests)
Transfer rate: 16206.52 [Kbytes/sec] received

Connection Times (ms)
min mean[+/-sd] median max
Connect: 4 20 10.4 20 47
Processing: 6 22 11.0 22 60
Waiting: 6 22 11.0 22 59
Total: 12 42 14.1 41 89

Percentage of the requests served within a certain time (ms)
50% 41
66% 48
75% 52
80% 54
90% 60
95% 66
98% 75
99% 75
100% 89 (longest request)

Hardly scientific but it's definitely worth showing how significant the difference is.


MoovAtom

Live Streaming "The Raid" documentary premiere in LA

So, MoovAtom went out to LA and streamed the world premiere of the new World of Warcraft documentary from Blue Loon Films' "The Raid". At peak, we had 1,200 simultaneous viewers watching a ~500kbps stream on six continents and over 30 countries. This level of traffic continued for solid 2 hours and the streaming servers handled it extremely well. The setup included a single xlarge Wowza EC2 instance in the US West region as the ingest point, running WireCast on a Macbook Pro i7 pushing the stream. The live stream actually stayed up for around 7 hours solid without a hiccup on any point of the chain. This says quite a bit about what type of bandwidth can be pushed on a single EC2 instance: Roughly 585 Mbps. We had our servers ready in the US East region, but we didn't even need them. Those 1200 people put the single c1.xlarge instance at an average 5% load with spikes up to 6%.

Here is a teaser:

MoovAtom