Steinwurf’s OTAcast

We’ve created a new test bed to show the benefits of using OTAcast for file distribution over multicast/broadcast enabled networks. We'll initially use 16 Raspberry Pi's as receivers. We hope to be able to transmit files to all 16 receivers almost as fast a traditional (e.g. FTP) system would send to a single device.

 

Since OTAcast is powered by next generation erasure correcting codes, we can efficiently deal with

(1) packet loss;

(2) devices joining at different times ;

(3) differences in receiver capabilities i.e. fast or slow receivers and asymmetric packet losses across different devices.

Components assemble! - OTAcast test bed components are ready to go!

Components assemble! - OTAcast test bed components are ready to go!

What is OTAcast?

OTAcast is a very powerful content distribution tool which can be used to successfully serve a very large number of clients in any condition. It offers customers the ability to optimally serve a large number of devices and cope with network losses in exchange for paying a small payload increase upfront.

An alternative to OTAcast may be to use UFTP, which is equally fast for a small number of devices in a perfect environment with no losses, but struggles to efficiently serve a large number of devices, especially when having to deal with losses, even minimal losses, when the number of devices involved starts to scale up.

What speeds can OTAcast offer?

We’ve been working a lot on optimising OTAcast recently and are continuously improving the throughput speeds achievable simultaneously to all devices. So far OTAcast has reached speeds (ultimately depending on the device CPU) in our testing well above 740mbit/s (and in fact closer to 2Gbit/s).

 

We expect the forthcoming benchmarks to show that:

  • By increasing the number of devices, the performance of OTAcast will not be compromised, with UFTP unable to match that performance, even in a zero loss environment. Our own benchmark analysis shows that as the number of receivers increase UFTP slows down - whereas OTAcast remains stable (due to it’s stateless nature).

  •  When experiencing any loss, OTAcast materially outperforms UFTP.  Zero loss networks – especially when dealing with many devices - are not realistic and our direct experience shows that even wired networks in airplanes experience losses. For this reason our analysis include a range of losses to mimic real life networks.

 What other unique features make OTAcast stand out?

  • OTAcast allows on-the-fly joining and leaving. – this means devices can be dropping in and out of the network due to their mobility, or extreme network conditions, and still receive files in the most efficient manner. Even if receivers are not fully synchronized OTAcast will work well.

  • OTAcast also deals well with networks with fluctuation capacity. In cases where networks are used for multiple things at the same time, sometimes the available capacity drops, which may result in packet loss – OTAcast can easily overcome this issue.

 

If you have a system of devices which regularly need file system of security updates, whether in IoT, industry 4.0, IFE networks in aircraft, satellite networks or autonomous or driverless vehicular and logistics networks, you too can join us in testing and evaluating OTAcast.

Get in touch to learn more and apply for a 3 Month Evaluation License to see the results for yourself.

In the meantime, stay tuned to see how our new testbed performs with OTAcast, and contact us with suggestions on other scenarios we might test.

Other Related Links:

Use Cases - Multicasting

OTAcast - Product Page

Blog - Multicast ECC/FEC

Previous
Previous

Using FEC to Reduce Jitter in TCP

Next
Next

Steinwurf and the Holy Grail for Industry 4.0: High Reliability with Ultra Low Latency - Q&A with Muriel Medard