Next Stream Ltd. Company Blog

We mentioned on several occasions in this blog, that we are working on IPTV solutions. Just like any other software company, we are trying to avoid reinventing the wheels, by using solid, proven solutions. So, we use Linux. And for us, most of the time Linux works great. We love Ubuntu. We have several boxes with different versions of Ubuntu Server and Ubuntu Desktop. Even this post is written on Ubuntu Desktop. We love other OSs too (Go FreeBSD !!!), but the focus of this post is our latest experience with Linux.

In our test network, we have several multicast streams. We use them for Set-Top-Box testing, middle-ware testing, fun, etc. And they all work fine. Until yesterday. We were testing a border device we’re developing and strange problem occurred. Our application is supposed to read from different streams (different multicast ip + port combinations) and process the data in a different manner depending on the source. But something wrong happened. After some testing and debugging we found that all multicast traffic reaches all readers. Kernel should register every reader to unique pair of multicast source address and port and send traffic only from this address and port to that reader. But apparently there is a bug with that. We thought it is a problem with our code and tried to debug it for several hours without any luck.

Later that day a person with more experience in this matter told us that it’s possibly a Linux kernel bug. Seems reasonable. Now, let’s confirm it.

Moral of the story: It is not always your code that is the problem, just most of the time.

Management General

Our IPTV platform is shaping up. We successfully modified our middle-ware to be compatible with our set-top-box of choice. After dealing with several industry un-standardised protocol extensions and other small glitches we are having stable multicast and unicast streaming. HD content is nicely presented on the TV. Everything works like a charm.

We examined several other set-top-boxes and home entertainment appliances to get a general idea what degree of usability the competitors provide. We are actually not that impressed. Most of the boxes are too slow. Some of those are ugly too. However, we have the basic design. Several screen layouts are being tested and modified. A bit more work and our interface will be fast and usable enough. What we need right now is a graphical designer with experience in user interface development. But, we’re working to hire such person too.

We will keep several features hidden for now. Let’s not spoil the fun of experimenting with our solutions later, but a short report I can say channel management, volume management, on-screen display and everything else you’d expect from a consumer device is now operational. Eye-candy is coming up next.

This, of course, allows us to provide OEM solutions. If you prefer custom design, custom layout, branded interface, just your logo up there or a full-blown advertisement solutions, we can accommodate your needs, fulfill your requests.

For the end of this post we would like to tease you a bit. During our High-Definition tests we were able to play 40Mbps (forty megabits per second) 1080p content using our distribution and client equipment. This is currently known as Blu-Ray. Nice and sharp picture, nice and sharp …

Management General

We are coming to a success. With some help and lots of work we are hammering problem after problem. We have stable playback on our set-top-boxes. Some modules are still rough around the edges, but the whole system is shaping up. Anyway, we can present Telco/ISP oriented, prototype IPTV system, capable of Multicast and Unicast video distribution. What is more important, flexible enough to adapt to the needs of any Internet Service Providers or Telecoms.

Our IPTV design currently consists of several layers and a number of modules:

• Video Data Acquisition
• Media Library
• User Management
• Load-balanced Content Distribution
• Theme-able User Interface
• HD-enabled Customer Premises Equipment

Basic prototypes of all of the above are present and working. Our work is focused on what I can describe as “commercialisation” of the particular products. We are fixing bugs, adding relevant features. Making devices and appliances user-friendly and intuitive is a main goal. We understand that the packaging sells. We want our customers happy.

But it is not all bells and whistles. We are fighting every day with the lack of standardisation in IPTV. Many protocols exist. Many hardware and software vendors. And all of them usually interpret every protocol in different way. Which is worse, many IPTV related protocols make room for disambiguations. It is usually, up to the implementor to decide how te deal with this. And most implementors do so they favour particular device or manufacturer. Which breaks compatibility, which requires dirty workarounds in the code.

Well, the RFCs give you freedom. They actually let you create unique product. When we speak of IPTV, they even let you break compatibility with proprietary extensions. But please, do know what you are dealing with, and please, do it right. One of my recent findings was that a big IPTV company “extends” particular IPTV related protocol with encapsulation of XML data in the payload. And I can find some logic in this. XML is, de facto, the standard for data representation these days. XML is well understood. XML is wide-spread. Many libraries exist to work with XML content.

What I don’t understand is how you can ship a product that works with malformed XML. And further more, why  would this product refuse to work with correct XML ?!

Please, make all our lives easier. Obey the standards !

Management Advertising, General, Tech

In a previous post we discussed some problems we had with High-definition Set-Top-Box we plan to use for our IPTV work. After several days of work, digging network traffic dumps and patching middle-ware components we can safely say, we fixed this problem. But a new problem emerges. We are currently working on a permanent solution to rework the whole device firmware with the device manufacturer. Great thanks to Michael and John for the cooperation and understanding.

Until we finish that, we plan to release a product targeted at Telcos/ISPs that will take care of acquisition of data for IPTV networks. We would like to provide full, flexible IPTV solution for different setups. Starting from Telcos/ISPs and going “down” to public buildings, advertisement and entertainment installations. Organising appliance requires more than just putting it all together. We want it to have nice look and feel, be intuitive to work with, save you more troubles than it causes. A preliminary design is ready. Some of the prototyping too.

What is more interesting about this appliance is that we will use it to introduce internal project we’ve been working on for some time now. We are talking about secure, centralised configuration solution, scalable enough to handle datacenters. Our goal is to lower the bar for system administration of large number of devices, make it accessible and comfortable. We have several more devices planned down the road that will probably follow the same architectural schematics and depending on the success of this we may even provide stand-alone builds of the configuration system as separate product.

Stay tuned !

Management General, Tech