[opendtv] Can FireWire again save the day?
- From: "Manfredi, Albert E" <albert.e.manfredi@xxxxxxxxxx>
- To: <opendtv@xxxxxxxxxxxxx>
- Date: Thu, 11 May 2006 11:42:04 -0400
We've heard this before, almost verbatim. The 1394 Trade Association is
sounding today remarkably like the ATM Forum sounded in the mid to late
1990s. When the link layer scheme fails to establish itself as had been
foreseen, all bets were placed on reviving interest by layering IP over
the technology.
One big problem with ATM was that in a greater IP internetwork, ATM's
special features become virtually impossible to depend on. It's hard to
go through IP routers and maintain those virtual circuit features, and
IP over ATM was never more than small islands in an ocean of IP over
best effort packet switched nets anyway.
And on the other hand, things IP can otherwise do easily became so
cumbersome when IP was layered over ATM (one good example being IP
multicast).
Home nets are smaller, confined, so one might argue the situation is
different. But that cuts both ways. The smaller scale and non-transit
properties of a home network are exactly what makes simple Ethernet link
layers adequate in such nets, even for time-sensitive comms. Just throw
bandwidth at the problem, if all else fails. If IEEE 1394 needs 400
Mb/s, throw 1 Gb/sec Ethernet at it.
Bert
--------------------------------------
May 01, 2006
Can FireWire again save the day?
http://www.digitaltvdesignline.com/howto/showArticle.jhtml?articleID=3D18=
7
201490
By Junko Yoshida
Dresden, Germany -- The latest "Mission Impossible" has nothing to do
with Tom Cruise; it's the revival of IEEE 1394 as the fundamental
networking technology for the home.
The early buzz over IEEE 1394, also called FireWire, fizzled long ago.
Many in the industry lost hope for 1394 as a linchpin of home networking
after the technology fell victim to business politics and industry
infighting. Nonetheless, promoters who gathered in Dresden last week for
a 1394 Trade Association quarterly meeting insisted that IEEE 1394 is
not just back, but never really went away. Their new strategy is to
exploit consumers' appetite for HDTV, building on the momentum of the
High-Definition Audio-Video Network Alliance's emerging specification
for sending 1394 signaling over coaxial cables, a technique called "1394
over coax."
"Look around," said Jalil Oraee, the founder and chief technology
officer at Oxford Semiconductor Ltd. (Abingdon, England). After all the
talk about wireless or wired Ethernet-based home networks, or the good
old coaxial-based home network promoted by the Multimedia Over Coax
Alliance, there are still no other technologies with enough bandwidth
and distance to distribute multiple high-definition TV streams while
offering clock synchronization, guaranteed delivery of data and
quality-of-service, according to Oraee.
Where best-effort-based Ethernet devices could easily crash and
introduce random delays, "with 1394, we can send not only IP [Internet
Protocol] packets, but also [ensure] quality-of-service," Oraee said.
Technologically speaking, many in the industry agree that IEEE 1394 may,
in fact, still have a lot going for it. Its failure to ignite the
home-networking market "wasn't the 1394 industry's fault," said Richard
Doherty, research director of The Envisioneering Group (Seaford, N.Y.).
Rather, the consumer electronics industry's initial enthusiasm for 1394
was "derailed by cable industry inaction" and "possible subterfuge,"
Doherty said.
In fact, at one point IEEE 1394 appeared to have the inside track as the
preferred home-networking medium. A long, hard negotiation between the
cable and CE industries produced an agreement mandating that U.S. cable
operators install 1394 in every set-top box. The agreement was ratified
by the Federal Communications Commission.
Despite the FCC's longstanding approval, however, James Snider,
executive director of the 1394 Trade Association, last week complained
that "most cable operators in the United States are not delivering
functioning 1394 ports."
Doherty confirmed this. "The 1394 connector is there, [but] the software
protocols are not." He added, "Comcast flatly refuses to give cable
subscribers a 1394 working box, despite the 2004 FCC directive that they
must."
Other industry analysts, who spoke on condition of anonymity, observed
that Scientific-Atlanta, one of two set-top vendors dominating the U.S.
market, won't provide the 1394 software stack to make its boxes work
with 1394. Why? Because U.S. cable operators and cable-box vendors
prefer not to see their traditionally closed business invaded by the
consumer electronics industry.
Brian O'Rourke, senior analyst at In-Stat, put it bluntly: "As an
interface technology, 1394 has done fairly well. As a networking
technology, it has been a failure." But even O'Rourke believes the
emerging High-Definition Audio-Video Network Alliance may be "a boon" to
1394 in the CE world. "HANA offers 1394 an opportunity to become a
viable networking technology," he said.
HANA promises to enable households to stream five high-definition
signals within the home simultaneously. HANA maintains that a
400-Mbit/second data rate, enabled by the 1394 specs, can do the job
while keeping programming content secure by means of the so-called "5C"
content-protection scheme. HANA has backing from a wide range of
companies, including movie studios (Warner Bros. and NBC/Universal),
cable companies (Charter), CE companies (Samsung) and chip makers (Texas
Instruments).
"We see 400 Mbits/s as a required minimum," said Hans van der Ven, a
chairman of the 1394-over-coax working group within the 1394 Trade
Association. Even if they don't own five HDTV sets, consumers doing a
trick play such as fast-forward on a 20-Mbit/s HD stream on a home
network could easily use up five times that bit rate, he explained. HANA
has a road map to move to a data rate of 800 Mbits/s by 2009, van der
Ven added.
Asked about 1394's role in HANA, van der Ven, an independent consultant
at PC & TV-AV Connections (Stamford, Conn.) who formerly worked for
Panasonic, said, "While HANA promotes its vision, we write standards."
The 1394-over-coax specification aims to leverage the coaxial cable
already installed in many U.S. homes (see story, below). "You need 1394
for two reasons," van der Ven summed up: "when you want to record HD
programming, and if you want to move it around anywhere else in the
house."
Although IEEE 1394 was eclipsed by the high-definition multimedia
interface as a point-to-point interface between a cable set-top (or HDTV
receiver) and a display, HDMI doesn't work well as the interface between
a set-top and a home recording device. Since it's a one-way interface
for sending uncompressed digital signals from one device to another,
HDMI will end up transferring too large a file to a recording device,
thus compromising the recording device's storage capacity. In contrast,
1394, which is designed to send compressed signals, is optimized for
recording content from a digital TV, as well as playing it back.
Bandwidth junkies
Even Microsoft Corp. seems to agree. While the computer software giant
won't be ready to support 1394b in the upcoming OS Vista, it plans to
ship 1394b support in the service pack. "We have PC OEMs' interest; 1394
will be important for high-bandwidth junkies as an interface for
external storage systems," said Mark Slezak, program manager of
Microsoft's Windows Device Experience Group.
HANA, however, lacks a consumer electronics heavyweight--such as Sony or
Matsushita--in its camp. Both Sony and Matsushita are pursuing Ethernet
as an A/V home-networking medium via the Digital Living Network Alliance
(DLNA).
Acknowledging that Panasonic has not joined HANA, Paul Liao, vice
president and CTO of Panasonic Corp. of North America, said, "Today, all
sorts of networks--WAN, LAN, data, voice, video--are moving toward being
based on IP protocols. For this reason, Panasonic sees there are great
benefits for home video and audio entertainment networks to use IP-based
protocols rather than IEEE 1394 protocols, which HANA is based on."
The 1394 Trade Association doesn't agree, nor do HANA promoters. HANA
chairman Jack Chaney, the director of the DMS Labs at Samsung, flatly
said, "HANA is IP-based and guarantees delivery." Peter Johansson, vice
president of Congruent Software Inc. (Bellevue, Wash.), concurred.
"Ethernet and Internet Protocol are not one and the same thing. IP is so
easy to carry . . . Ethernet carries IP, 1394 carries IP and coax
carries IP."
Chaney remains confident that other companies will join HANA.
"Large corporations' strategies are complex and very secret," he said.
"If the objectives of HANA and the 1394 Trade Association are achieved
on schedule, then these large corporations will be part of the flow to
HANA."
The technical work within the 1394 group is converging on IP as the
control plane for all network devices (that is, TCP/IP encapsulation for
A/V devices). Johansson is proposing to transport A/V streams as
out-of-band isochronous data, while connecting link segments with L3
bridges.
In the proposed HANA-based home network architecture using 1394 and
1394-over-coax, the group plans to put in place a "bridge" that
translates a 1394 bus to 1394-over-coax or, in the future, another
controller that bridges 1394 and Ethernet networks. It will also install
a proxy device that helps the network discover what legacy devices are
out there, and send control commands. The proxy device would translate
HANA commands--transported on IP--into A/V device commands and send
them, Johansson explained. "Proxy devices are like having a guide and an
interpreter in one," he said.
Indeed, beyond the lack of clock synchronization, the biggest issue with
DLNA for 1394 promoters is its treatment of legacy A/V devices,
according to van der Ven. "The new Ethernet router that comes with the
latest Universal Plug & Play features talks to DLNA devices, but it
stops legacy devices from interfering"--effectively banishing the legacy
products from new home network technologies.
All material on this site Copyright 2006 CMP Media LLC. All rights
reserved
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts: