[RTW] New proposed charter text. Please review before the BoF.

[Cullen Jennings]

So at one point we had some text along the lines of the proposed
 
    "Solutions that require mediation functions to achieve interoperation might be acceptable, provided such functions are not unduly complex, expensive or detrimental to performance."

But I think that sort of text does not turns out to be useful in a charter. It is so vague and ambiguous that it does not guide the work. So lets dig into the things that might guide the work and see what we could do in a charter. 

Clearly we need extensibility and ways of negotiating new things. There many points were we need extensibility but video codecs are one of them. Twelve years down the road, its highly likely that there will be new codecs in browsers beyond what we have today - who knows, perhaps they will do 3D. Clearly we need a way for an browser application to find out which codecs that browser supports then use that for negotiation. But, in as Henning nicely put it, to avoid failure of negotiation, we need some minimal set of things all browsers have. 

With something like crypto algorithms for SRTP, the odds are good we can easily find something that was both acceptable to browsers and was widely supported by legacy voip equipment that does SRTP. With something like narrowband audio codecs it gets harder, but it is probably possible to come to something workable for a many cases. Phone BCP did. As you point out, video in some way has the highest cost of interoperability failure, but the bad news is that unfortunately legacy voip equipment already has widespread interoperability failure on video codec choices. Probably the most prevalent codec in interactive sessions is H.263 but much of the new equipment does not support that. Even the equipment that supports H.264 AVC (ignoring SVC), there are so many variants that much of the H.264 systems have no common way of interoperating with other H.264 systems. To have widespread support for legacy video equipment, browsers would have to implement a lot of codecs and profiles. H.265 is slated to be completed before I would expect this work to be deployed in the bulk of browser. Then there is the MPEG work on codecs that proponents hope will be royalty free. Not to mention some other obvious candidates that are widely used. GIven the lack of a common video codec in existing legacy equipment, I don't see how this working group could fix that. However, I think this working group needs to pick something that ensure that equipment compliant with specifications from the proposed WG does not suffer from the same video interoperability failure as legacy equipment. 

On the topic of what will be the hardest point of interoperability with legacy systems, I think it will be ICE like connectivity check issues. Every proposal I have seen so far has solved the problem of authorizing where the browser can send packets RTP by some connectivity check scheme such as ICE. However, ICE is not implemented the major of legacy voip equipment and I think this will end up being the biggest factor reducing what legacy equipment can work. Someone in the working group might think of a clever way to avoid this problem but I have not herd one yet. 

I want the charter to clearly allows the WG to figure out the right way to map to existing SIP systems. To do this, I think the browser has to be able to report what codecs and other extensibility options it supports then allow the application to influence what gets used to allow negotiation with the far end. In my mind, the current charter, particularly point 5 and 6, provide that. 

With regards to backwards compatibility with legacy voip systems, it's going to be a balance with tradeoffs.  We can't have a charter that says everything must support all legacy devices because the legacy device can't even talk to all other legacy devices. I want the WG to be able to make these decision and have consensus the balance right before publishing the specifications that would come out of the WG. But making theses choices is the work in this working group is largely about. 

The WG needs to define an architecture that allows the sort of long term extensibility that David mentioned while at the same time make sure that systems compliant with the work can done suffer from interoperability failures. From my point of view the current charter would give the WG the room to make theses choices. 

Cullen 



On Mar 21, 2011, at 8:36 AM, Elwell, John wrote:

> OK, so what I think Christer and Andy are saying is that it is not just about having the right the mandatory-to-implement protocols, codecs, etc., but also about sufficient control at the API so that appropriate ones can be selected when interoperating with existing equipment. I would agree with that. Do Christer and Andy consider my text proposal for the charter to be sufficient, or does it need changing or extending?
> 
> John
> 
>> -----Original Message-----
>> From: Hutton, Andrew
>> Sent: 21 March 2011 14:16
>> To: Elwell, John; Christer Holmberg; Ted Hardie
>> Cc: rtc-web at alvestrand.no
>> Subject: RE: [RTW] New proposed charter text. Please review
>> before the BoF.
>> 
>> Hi,
>> 
>> I think Christer's comment is really important.
>> 
>> If the web application is going to be responsible for
>> ensuring interoperability with SIP or anything else then the
>> API between the browser and the web application needs to be
>> flexible and powerful to enable this interoperability to take
>> place. So I believe the charter needs to make it clear that
>> the input for this API that the IETF provides to W3C will
>> take in to account the fact that the web application has to
>> achieve interoperability with non RTC-Web based systems of
>> which SIP/SDP based systems are the obvious example.
>> 
>> Regards
>> Andy
>> 
>> 
>> 
>>> -----Original Message-----
>>> From: rtc-web-bounces at alvestrand.no
>>> [mailto:rtc-web-bounces at alvestrand.no] On Behalf Of Elwell, John
>>> Sent: 21 March 2011 07:41
>>> To: Christer Holmberg; Ted Hardie
>>> Cc: rtc-web at alvestrand.no
>>> Subject: Re: [RTW] New proposed charter text. Please review
>>> before the BoF.
>>> 
>>> 
>>> 
>>>> -----Original Message-----
>>>> From: Christer Holmberg [mailto:christer.holmberg at ericsson.com]
>>>> Sent: 19 March 2011 11:22
>>>> To: Ted Hardie; Elwell, John
>>>> Cc: rtc-web at alvestrand.no
>>>> Subject: RE: [RTW] New proposed charter text. Please review
>>>> before the BoF.
>>>> 
>>>> Hi,
>>>> 
>>>> Regarding interoperability (or, creating new solutions in
>>>> general), there are a couple of things we need to consider.
>>>> 
>>>> One is of course the protocols supported by the browser (e.g
>>>> RTP, ICE etc).
>>>> 
>>>> The second, which I THINK John is referring to, is to ensure
>>>> that web applications (whether they are SIP based or based on
>>>> something else) are able to ENABLE the browser protocols and
>>>> features in a good and effective way. For that we need to
>>>> ensure that the API between the browser and the web app is
>>>> powerful enough.
>>> [JRE] That might be one of the consequences - if the browser
>>> is able to use a protocol that is not interoperable with
>>> existing equipment and a protocol that is interoperable, then
>>> clearly an application wanting to interoperate would want to
>>> be able to choose the latter.
>>> 
>>> That was not the main point of my proposed addition to the
>>> charter - the main point was that the set of
>>> mandatory-to-implement protocols in the browser should
>>> include those that can interoperate in a reasonable way with
>>> existing equipment. Based on discussions to date, the most
>>> problematic is likely to be codecs, particularly video
>>> codecs, because of the cost and performance degradation when
>>> transcoding has to be done.
>>> 
>>> John
>>> 
>>> 
>>>> 
>>>> In the ucreqs document, we have tried to differentiate
>>>> between those two things by having separate browser
>>>> requirements and API requirements.
>>>> 
>>>> Regards,
>>>> 
>>>> Christer
>>>> 
>>>> 
>>>> 
>>>> ________________________________________
>>>> From: rtc-web-bounces at alvestrand.no
>>>> [rtc-web-bounces at alvestrand.no] On Behalf Of Ted Hardie
>>>> [ted.ietf at gmail.com]
>>>> Sent: Friday, March 18, 2011 7:38 PM
>>>> To: Elwell, John
>>>> Cc: rtc-web at alvestrand.no
>>>> Subject: Re: [RTW] New proposed charter text. Please review
>>>> before the BoF.
>>>> 
>>>> On Fri, Mar 18, 2011 at 10:17 AM, Elwell, John
>>>> <john.elwell at siemens-enterprise.com> wrote:
>>>>> Ted,
>>>>> 
>>>>> In the same way that charters tend to call out "horizontal"
>>>> topics for special mention, such as security, privacy and
>>>> NAT/firewall, I believe the ability to achieve interop with
>>>> existing equipment should also be given a mention.
>>>> 
>>>> Hi John,
>>>> 
>>>> Unsurprisingly, I am also a fan of interoperability.  But I think
>>>> there are multiple ways to scope interoperability.  The charter
>>>> currently does it by focusing on the interoperability we
>> get at the
>>>> protocol level, largely by choosing tools for this toolkit
>>> that match
>>>> the tools already in place (e.g. RTP, ICE, etc.).  I agree that we
>>>> shouldn't select tools that are in place but unusable.  I
>>> think we end
>>>> up having to trust the working group on this anyway, as it
>>> will decide
>>>> which tools are unusuable even if we insert a "don't pick unusable
>>>> tools" section.
>>>> 
>>>> So the basic question boils down to:  do we also want to
>>> assert a need
>>>> for interoperability with a specific application or set of
>>>> applications in the charter?
>>>> 
>>>> My personal take is no.  Those are use cases, and important
>>> ones.  But
>>>> the charter is focused on building a toolkit "to enable
>>> innovation on
>>>> top of a set of basic components."  I have no doubt that
>>> someone will
>>>> be able to build a fully backwards compatible service using these
>>>> components and a gateway,.  But the more important charter
>>> goal is to
>>>> get the component set right to enable them to build more
>> than that.
>>>> And I think calling out that specific use case is more
>>> likely to limit
>>>> our vision than to result in more clarity in picking protocols and
>>>> tools.
>>>> 
>>>> Again, just my personal opinion,
>>>> 
>>>> regards,
>>>> 
>>>> Ted Hardie
>>>> 
>>>>> Note that I stop short of demanding backwards compatibility,
>>>> but instead make a somewhat more flexible proposal.
>>>>> 
>>>>> Using defined protocols isn't necessarily sufficient if we
>>>> pick things that are not implemented. As an extreme case,
>>>> take S/MIME, for example.
>>>>> 
>>>>> John
>>>>> 
>>>>> 
>>>>>> -----Original Message-----
>>>>>> From: Ted Hardie [mailto:ted.ietf at gmail.com]
>>>>>> Sent: 18 March 2011 17:01
>>>>>> To: Elwell, John
>>>>>> Cc: rtc-web at alvestrand.no
>>>>>> Subject: Re: [RTW] New proposed charter text. Please review
>>>>>> before the BoF.
>>>>>> 
>>>>>> On Fri, Mar 18, 2011 at 5:04 AM, Elwell, John
>>>>>> <john.elwell at siemens-enterprise.com> wrote:
>>>>>>> Ted,
>>>>>>> 
>>>>>>> I think there needs to be some mention of interop with
>>>>>> existing real-time applications using IETF protocols (such as
>>>>>> SIP, RTP). Something along the lines of:
>>>>>>> "Work should take account of browser-based applications
>>>>>> that require to interoperate with existing applications using
>>>>>> IETF protocols such as SIP and RTP and commonly deployed
>>>>>> audio and video codecs. Solutions that require mediation
>>>>>> functions to achieve interoperation might be acceptable,
>>>>>> provided such functions are not unduly complex, expensive or
>>>>>> detrimental to performance."
>>>>>>> 
>>>>>> 
>>>>>> So, my personal read is that these work items:
>>>>>> 
>>>>>> "4.     Define which RTP functions and extensions shall be
>>>> supported
>>>>>> in the client and their usage for real-time media,
>>> including media
>>>>>> adaptation to ensure congestion safe usage.
>>>>>> 5.     Define what functionalities in the solution,
>> such as media
>>>>>> codecs, security algorithms, etc., can be extended and how the
>>>>>> extensibility mechanisms works.
>>>>>> 6.     Define a set of media formats that must or should
>>>> be supported
>>>>>> by a client to improve interoperability."
>>>>>> 
>>>>>> and this general statement which follows the work items:
>>>>>> 
>>>>>> "This work will be done primarily by using already defined
>>>> protocols
>>>>>> or functionalities. "
>>>>>> 
>>>>>> Those seem to cover the same ground as your "take
>>> account of".  If
>>>>>> your aim is to get to something more concrete, like "you
>>>> must be able
>>>>>> to build a browser-based SIP softphone using the tools
>>>> identified by
>>>>>> this group", I think you need to unpack that a bit more.
>>>> Some of the
>>>>>> presence aspects of that toolkit, for example, may not be
>>>> identified
>>>>>> here.
>>>>>> 
>>>>>> To put this another way, my personal read is that
>> we're aiming to
>>>>>> create the toolkit needed to build real-time media
>>>> applications in web
>>>>>> contexts.  Any specific application, including a SIP
>>>> softphone, would
>>>>>> be a use of the toolkit; those will be covered by
>> something like
>>>>>> draft-holmberg-rtcweb-ucreqs-00.txt as a use case, rather
>>>> than in the
>>>>>> charter.
>>>>>> 
>>>>>> Again, just my personal take on this.
>>>>>> 
>>>>>> regards,
>>>>>> 
>>>>>> Ted Hardie
>>>>>> 
>>>>>> 
>>>>>>> John
>>>>>>> 
>>>>>>>> -----Original Message-----
>>>>>>>> From: rtc-web-bounces at alvestrand.no
>>>>>>>> [mailto:rtc-web-bounces at alvestrand.no] On Behalf Of
>> Ted Hardie
>>>>>>>> Sent: 17 March 2011 16:18
>>>>>>>> To: rtc-web at alvestrand.no
>>>>>>>> Subject: [RTW] New proposed charter text. Please review
>>>>>>>> before the BoF.
>>>>>>>> 
>>>>>>>> Name: Real-Time Communication in WEB-browsers (RTCWeb)
>>>>>>>> 
>>>>>>>> There are a number of proprietary implementations that
>>>>>> provide direct
>>>>>>>> interactive rich communication using audio, video,
>>>> collaboration,
>>>>>>>> games, etc. between two peers' web-browsers. These are not
>>>>>>>> interoperable, as they require non-standard extensions or
>>>>>> plugins to
>>>>>>>> work.  There is a desire to standardize the basis for such
>>>>>>>> communication so that interoperable communication can be
>>>>>> established
>>>>>>>> between any compatible browsers. The goal is to enable
>>>>>> innovation on
>>>>>>>> top of a set of basic components.   One core component
>>>> is to enable
>>>>>>>> real-time media like audio and video, a second is to
>>>>>> enable datagram
>>>>>>>> and byte stream data transfer directly between clients.
>>>>>>>> 
>>>>>>>> This work will be done in collaboration with the W3C.
>>>> The IETF WG
>>>>>>>> will produce architecture and requirements for selection
>>>>>> and profiling
>>>>>>>> of the on the wire protocols. The architecture needs to be
>>>>>> coordinated
>>>>>>>> with W3C.  The IETF WG work will identity state
>>>>>> information and events
>>>>>>>> that need to be exposed in the APIs as input to W3C. The
>>>>>> W3C will be
>>>>>>>> responsible for defining APIs to ensure that
>>>> application developers
>>>>>>>> can control the components.
>>>>>>>> 
>>>>>>>> The security goals and requirements will be developed by
>>>>>> the WG. The
>>>>>>>> security model needs to be coordinated with the W3C.
>>>> The work will
>>>>>>>> also consider where support for extensibility is needed. RTP
>>>>>>>> functionalities, media formats, security algorithms are
>>>> example of
>>>>>>>> things that commonly needs extensions, additions or
>>>>>> replacement, and
>>>>>>>> thus some support for negotiation between clients
>> is required.
>>>>>>>> 
>>>>>>>> The WG will perform the following work:
>>>>>>>> 1.     Define the communication model in detail, including
>>>>>> how session
>>>>>>>> management is to occur within the model.
>>>>>>>> 2.     Define a security model that describes the security
>>>>>> goals and
>>>>>>>> how the communication model can achieve these goals.
>>>>>>>> 3.     Define how NAT and Firewall traversal is to occur.
>>>>>>>> 4.     Define which RTP functions and extensions
>> that shall be
>>>>>>>> supported in the client and their usage for real-time
>>>>>> media, including
>>>>>>>> media adaptation to ensure congestion safe usage.
>>>>>>>> 5.     Define what functionalities in the solution,
>>>> such as media
>>>>>>>> codecs, security algorithms, etc., that can be extended
>>>> and how the
>>>>>>>> extensibility mechanisms works.
>>>>>>>> 6.     Define a set of media formats that must or should
>>>>>> be supported
>>>>>>>> by a client to improve interoperability.
>>>>>>>> 7.     Define how non RTP datagram and byte stream data
>>>>>> communication
>>>>>>>> between the clients can be done securely and in a
>>>>>> congestion safe way.
>>>>>>>> 8.     Provide W3C input for the APIs that comes from the
>>>>>>>> communication model and the selected components and
>>>>>> protocols that are
>>>>>>>> part of the solution.
>>>>>>>> 
>>>>>>>> 
>>>>>>>> This work will be done primarily by using already defined
>>>>>> protocols or
>>>>>>>> functionalities. If there is identification of missing
>>>> protocols or
>>>>>>>> functionalities, such work can be requested to be done
>>>> in another
>>>>>>>> working group with a suitable charter or by requests for
>>>>>> chartering it
>>>>>>>> in this WG or another WG. The following topics will be
>>>> out of scope
>>>>>>>> for the initial phase of the WG but could be added after a
>>>>>> recharter:
>>>>>>>> RTSP, RSVP, NSIS, Location services, IM & Presence,
>>>>>> Resource Priority.
>>>>>>>> 
>>>>>>>> The products of the working group will support security
>>>>>> and keying as
>>>>>>>> required by BCP 61 and be defined for IPv4, IPv6, and
>>> dual stack
>>>>>>>> deployments. The Working Group will consider the
>>> possibility of
>>>>>>>> defining a browser component that implements an
>>> existing session
>>>>>>>> negotiation and management protocol. The working group
>>>>>> will follow BCP
>>>>>>>> 79, and adhere to the spirit of BCP 79. The working
>>> group cannot
>>>>>>>> explicitly rule out the possibility of adopting encumbered
>>>>>>>> technologies; however, the working group will try to avoid
>>>>>> encumbered
>>>>>>>> technologies that require royalties or other encumbrances
>>>>>> that would
>>>>>>>> prevent such technologies from being easy to use in web
>>>> browsers.
>>>>>>>> 
>>>>>>>> Milestones:
>>>>>>>> 
>>>>>>>> Aug 2011 Architecture and Security and Threat Model
>>> sent to W3C
>>>>>>>> 
>>>>>>>> Aug 2011 Use cases and Scenarios document sent to W3C
>>>>>>>> 
>>>>>>>> Sept 2011 Architecture and Security and Threat Model to IESG
>>>>>>>> as Informational
>>>>>>>> 
>>>>>>>> Sept 2011 Use cases and Scenarios for RTCWeb document sent
>>>>>> to IESG as
>>>>>>>> Informational
>>>>>>>> 
>>>>>>>> Dec 2011 RTCWeb and Media format specification(s) to
>>> IESG as PS
>>>>>>>> 
>>>>>>>> Dec 2011 Information elements and events APIs Input to W3C
>>>>>>>> 
>>>>>>>> Apr 2012 API to Protocol mapping document submitted to
>>>> the IESG as
>>>>>>>> Informational (if needed)
>>>>>>>> _______________________________________________
>>>>>>>> RTC-Web mailing list
>>>>>>>> RTC-Web at alvestrand.no
>>>>>>>> http://www.alvestrand.no/mailman/listinfo/rtc-web
>>>>>>>> 
>>>>>> 
>>>> _______________________________________________
>>>> RTC-Web mailing list
>>>> RTC-Web at alvestrand.no
>>>> http://www.alvestrand.no/mailman/listinfo/rtc-web
>>> _______________________________________________
>>> RTC-Web mailing list
>>> RTC-Web at alvestrand.no
>>> http://www.alvestrand.no/mailman/listinfo/rtc-web
>>> 
>> 
> _______________________________________________
> RTC-Web mailing list
> RTC-Web at alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/rtc-web