Kill HANA when it is still young!
I do not know how this one escaped me in CES, but it looks like a new alliance for the distribution of video (specifically high definition video) has been formed at the end of 2005 and this group was represented in CES of 2006 (with actual products based on their guidelines planned for CES of 2007). The group is called High-Definition Audio-Video Network Alliance (HANA) and it is essentially an alternative to DLNA, with a very different philosophy for the ditribution of media in the digital home.
According to the HANA philosophy audio and video need their own network and should not use the home data network (which they refer to as the IT network), but rather instead they suggest using a separate network that can supposedly better handle high definition content. This separate network will employ content protection schemes (DTCP) to complement the hardware content protection built into next generation CE devices thus forming a secure connection from point to point. They further suggest that commercial content will flow only through the A/V network (which is the only one "properly secured" for it) and personal media (such as the one created by consumers via their video camera) will be allowed to flow into the A/V network from the IT network via a proper gateway, however no data flow in the other direction will be allowed.
As the link level solution for the A/V network, HANA is promoting IEEE 1394b (FireWire that can work for up to 100 meters) instead of DLNA's Ethernet, WiFi and Bluetooth. This is presumably because Firewire is isochronous and hence better suited for video, however in practice it is my opinion that they needed some link level separation between the IT network and the A/V network and so looked for what's available out there. IEEE 1394 has been around long enough, is very fast (400 Mbps to 1.6 Gbps), supports DTCP, and most importantly is not used today for networking (due to wire length limitations) but it can be now with IEEE 1394b, so they adopted it.
In terms of the software, HANA like DLNA is not inventing new protocols or data representation formats, but rather they make use of existing standards that they find most appropriate and provide guidelines for using them in the context of the usage scenario they aim to satisfy. According to them, their focus in HD video inside the home has lead them to make very different choices compared to DLNA. It seems like they do not plan to adopt UPnP, instead they focus in remoting user interfaces from one device to another via XHTML-Basic and CSS1 (kind of like WAP 2.0). This decision to adopt a WWW like mechanism is very peculiar when taking into account that existing efforts to do something like that have proven inferior to UPnP due to the sluggishness of HTML which leads to bad response times and a poor user experience (imagine scrolling through hundreds or thousands of songs where each page with 5-6 songs requires a new HTML page to be fetched and displayed, this is insane!). In contrast UPnP takes a Web services like approach and delivers only the business data to the player (XML as opposed to (X)HTML), leaving the UI to the player so that the player only needs to ppulate the data into the display and not render a new screen. While the UPnP approach is obviously less flexible it does seem to work better for those well defined scenarios that HANA is pursuing, so why not go with UPnP? Again it is my opinion that they strongly felt the need to create an alternative to DLNA and hence carefully chose a different mechanism for almsot any subject matter.
In summary, it seems like the founding members of HANA go into a lot of effort to associate the need for something like HANA with the difficulties in streaming high definition video (they set their goal for streaming up to 5 simultaneous HD streams), while downplaying the content protection issues. Considering the fact that gigabit Ethernet is already very affordable and is slowly making its way into homes as a replacement for 10/100 Ethernet, their arguments regarding the challenges with HD distribution do not hold. This is because one can easily stream 5 MPEG2 HD videos over a gigabit network and probably 10 MPEG4/AVC HD videos. In addition to that the current 400 Mbps throughput of Firewire will not allow 5 MPEG2 videos to be streamed simulatenously so where does that leave HANA? It is my opinion that the reality of the matter is that content companies (NBC Universal is a founding member...) are panicking when they realize that video distribution around the home is here to stay and fear that existing solutions are not protected enough. More precisely they do not like the open nature of existing home networks, which even with DRM, offer only link level protection when what they are looking for is system wide, end-to-end, tightly integrated protection in the hardware, software and the link.
So what do you do if you are NBC Universal and you are faced with such a gloomy reality? You start a new standard body, get some CE companies on board and hope for the best. While they extended an invitation for IT companies to join in, they so far have on board only Sun Microsystems. If you ask me, the others are probably busy with DLNA/UPnP (An Intel favourite), while Sun a late comer to this market decided to bet on HANA where it can achieve a better status as an early member than it can ever have in DLNA (and where they can also try pushing Java into the mix).
So if someone asks you about HANA, you should tell him that this is one more desperate move by content companies (and to a lesser degree perhaps also by CE companies) to resist change and push aside the PC/Mac world, an effort that is very likely to fail, although only time can tell.
Additional Reading:
HANA Overview Presentation
Two Approaches to Networked Digital Video: DLNA and HANA
According to the HANA philosophy audio and video need their own network and should not use the home data network (which they refer to as the IT network), but rather instead they suggest using a separate network that can supposedly better handle high definition content. This separate network will employ content protection schemes (DTCP) to complement the hardware content protection built into next generation CE devices thus forming a secure connection from point to point. They further suggest that commercial content will flow only through the A/V network (which is the only one "properly secured" for it) and personal media (such as the one created by consumers via their video camera) will be allowed to flow into the A/V network from the IT network via a proper gateway, however no data flow in the other direction will be allowed.
As the link level solution for the A/V network, HANA is promoting IEEE 1394b (FireWire that can work for up to 100 meters) instead of DLNA's Ethernet, WiFi and Bluetooth. This is presumably because Firewire is isochronous and hence better suited for video, however in practice it is my opinion that they needed some link level separation between the IT network and the A/V network and so looked for what's available out there. IEEE 1394 has been around long enough, is very fast (400 Mbps to 1.6 Gbps), supports DTCP, and most importantly is not used today for networking (due to wire length limitations) but it can be now with IEEE 1394b, so they adopted it.
In terms of the software, HANA like DLNA is not inventing new protocols or data representation formats, but rather they make use of existing standards that they find most appropriate and provide guidelines for using them in the context of the usage scenario they aim to satisfy. According to them, their focus in HD video inside the home has lead them to make very different choices compared to DLNA. It seems like they do not plan to adopt UPnP, instead they focus in remoting user interfaces from one device to another via XHTML-Basic and CSS1 (kind of like WAP 2.0). This decision to adopt a WWW like mechanism is very peculiar when taking into account that existing efforts to do something like that have proven inferior to UPnP due to the sluggishness of HTML which leads to bad response times and a poor user experience (imagine scrolling through hundreds or thousands of songs where each page with 5-6 songs requires a new HTML page to be fetched and displayed, this is insane!). In contrast UPnP takes a Web services like approach and delivers only the business data to the player (XML as opposed to (X)HTML), leaving the UI to the player so that the player only needs to ppulate the data into the display and not render a new screen. While the UPnP approach is obviously less flexible it does seem to work better for those well defined scenarios that HANA is pursuing, so why not go with UPnP? Again it is my opinion that they strongly felt the need to create an alternative to DLNA and hence carefully chose a different mechanism for almsot any subject matter.
In summary, it seems like the founding members of HANA go into a lot of effort to associate the need for something like HANA with the difficulties in streaming high definition video (they set their goal for streaming up to 5 simultaneous HD streams), while downplaying the content protection issues. Considering the fact that gigabit Ethernet is already very affordable and is slowly making its way into homes as a replacement for 10/100 Ethernet, their arguments regarding the challenges with HD distribution do not hold. This is because one can easily stream 5 MPEG2 HD videos over a gigabit network and probably 10 MPEG4/AVC HD videos. In addition to that the current 400 Mbps throughput of Firewire will not allow 5 MPEG2 videos to be streamed simulatenously so where does that leave HANA? It is my opinion that the reality of the matter is that content companies (NBC Universal is a founding member...) are panicking when they realize that video distribution around the home is here to stay and fear that existing solutions are not protected enough. More precisely they do not like the open nature of existing home networks, which even with DRM, offer only link level protection when what they are looking for is system wide, end-to-end, tightly integrated protection in the hardware, software and the link.
So what do you do if you are NBC Universal and you are faced with such a gloomy reality? You start a new standard body, get some CE companies on board and hope for the best. While they extended an invitation for IT companies to join in, they so far have on board only Sun Microsystems. If you ask me, the others are probably busy with DLNA/UPnP (An Intel favourite), while Sun a late comer to this market decided to bet on HANA where it can achieve a better status as an early member than it can ever have in DLNA (and where they can also try pushing Java into the mix).
So if someone asks you about HANA, you should tell him that this is one more desperate move by content companies (and to a lesser degree perhaps also by CE companies) to resist change and push aside the PC/Mac world, an effort that is very likely to fail, although only time can tell.
Additional Reading:
HANA Overview Presentation
Two Approaches to Networked Digital Video: DLNA and HANA
There's nothing to worry about. Nobody will rewire his home for Firewire. Using Ethernet/WLAN they may have a small chance (even if that would be optimistic, as they would have Microsoft and Sony as competitors), but with Firewire, without a wireless transmission? Never.
Posted by
Tim |
3:22 AM