August 2013

SMPTE Newswatch Masthead

Hot Button Discussion

Workflows in The Cloud

By Michael Goldman 

 

Now that media distribution companies such as Hulu, Netflix, and others have successfully pioneered ways to distribute media content using Cloud computing-based systems, the industry is now pondering what parts of the content production workflow chain might someday be moved into the Cloud.

 
As discussed last year in SMPTE Newswatch, beyond the strictly technical complications of moving huge media files through Cloud-based networks, the issue of content security has been one area that has required companies to move slowly. However, as major consumer retail and service companies are demonstrating, cloud providers have enabled encryption schemes and other security measures that their clients now trust. Industry veteran Al Kovalick, founder of Media Systems Consulting, suggests cloud providers now have the technology to keep media and data safe, and their challenge now is to earn the trust of their target industries in that regard.
 
"Security is becoming less of an issue in the sense that these companies have earned more trust from CEOs and CTOs," Kovalick says. "Trust is the important word. If you 'trust' Amazon's cloud security policies, then you are more likely to do business with them. The emphasis lately has been on how those cloud providers build that trust. Companies are using cloud services because they believe the security claims. This includes encrypting everything going in or coming out and deleting data permanently from the disc, as required. So that issue is not as big a challenge as it once was."
 
Momentum is gathering for the industry to examine how it can incorporate the Cloud into production workflows, among other applications. Those efforts picked up with the formation earlier this year of the Joint Task Force on Networked Media (JT-NM), an initiative spearheaded by the European Broadcasting Union (EBU), SMPTE, and the Video Services Forum (VSF). The basic goal of the task force is to develop documents that could become standards for IT-based networks for broadcasters, particularly in the areas of AV streaming and deterministic file transport. A byproduct of the initiative, if successful, could be the ability to move switchable, realtime video streams in Cloud-based networks.
 
"The bottom line is that these new ideas could eliminate SDI and AES/EBU links over time," Kovalick suggests. 
 
The effort is still in its early stage, but if it eventually succeeds, and if simultaneously major broadcast entities evolve their infrastructures to an IT-based way of networking, as discussed last month in SMPTE Newswatch, then potentially, the industry could reach a time when the networking, processing, and storage aspects of production workflows could start happening through the Cloud. This is not something that will happen quickly, but Kovalick thinks it is something that could happen eventually.
 
"There is, of course, already a standard for moving video over IP--SMPTE 2022-6," Kovalick points out. "That is for WAN transport, long distance work, and is not designed in any way to replace SDI. The stream is not frame-switchable as with SDI, but is fine for transporting video between long distances when it is not frame-based switched in route.
 
"However, the task force is trying to go beyond this. We hope to see short-term transmission, LAN-based, that can switch along the route without a media glitch. The goal of the task force is to create an ecosystem for streaming over Ethernet/IP. It would involve many switches, synchronization methods, path control methods, the end points--all those things cooperating together. Ideally, we want to reproduce the general functionality of SDI, but with commodity IT networks. That is the long-range goal."
 
Kovalick says the first job of the task force was to put out a call for user stories from broadcasters, describing exactly what they would like to accomplish using such a system.
 
"A typical user story would be 'I want to move an SDI payload, which typically is what the SDI link carries, over Ethernet and be able to route and switch it in a way that improves my workflows,'" he explains. "We received 137 of those user stories submitted. They will be the basis of a Request for Technology (RFT), to see what needs to be built to implement what they describe. We are already well into it and hope to have an RFT published at IBC in September."
 
"We've been doing file-based workflows moving files over Ethernet for about 10 years," Kovalick says. "It's solid and proven and it is how many businesses work today. File transport is not the same as realtime video streaming. For media companies, streaming over IT networks would be the next move; and that will take time. A typical media facility today is filled with non-IT gear, AV-specific routers, speakers, mics, control surfaces, lighting, and a lot more. Thus, using IT networks to move media will be something that coexists with much non-IT gear, likely forever."
 
Kovalick is quick to emphasize there will be no "knife-switch changeover" in broadcast facilities from an SDI infrastructure to an Ethernet infrastructure since SDI is so deeply embedded in the broadcast world. Rather, the two will co-exist even as the overall movement to an IT-based foundation marches onward over many years. The eventual goal will be the capability to take "the SDI payload" and network it near and far over an Ethernet-style system that would, by definition, eventually allow new kinds of remote processing and data storage to happen for production and other kinds of non-SDI related types of media.
 
"SDI is two things--the first part is the underlying bit structure on the wire, which is tricky [technical] stuff," Kovalick says. "It's scrambled and has a certain modulation format, certain timing, and so on. The second aspect that SMPTE defined is what rides above that? What is the actual payload? It is 1080i, 1080p, 720p, those sorts of things. That payload of what SDI carries--the audio, video, closed caption data, and metadata--SMPTE has done a marvelous job of defining all that. SMPTE should continue to develop such payload definitions. The difference is that they won't only ride over the SDI modulation format--it could [someday] ride over Ethernet. In the meantime, we will want everything to interoperate--sort of a world where you can have pure SDI and SDI-payload-over-Ethernet (SPoE), with bridges between the two. These are all long-term goals, but we think it could happen."
 
Kovalick expects the hardware companies that will have a major presence in the IT media facility universe are the ones that have a good sense of both video and IT. "These companies will work with IT gear in a way that understands the market for video/audio infrastructure," he adds. "Some of them are trying to build various kinds of fast applications for media over Ethernet."
 
The end goal is a world in which servers and other infrastructure are "standard off-the-shelf Ethernet for media companies to use," he says. "But it is really hard to move, say, 10 gigabits-a-second of realtime video to a server and out to an Ethernet port. I don't know if there will be special cards in them or not. I'm hoping not, but the jury is out as to whether a standard off-shelf server will interface with this special new world or whether we will need some special glue. Success would be to use all off-the-shelf, enterprise worthy, unmodified IT commodity gear. Whether we get there 100% is unclear, but I do think, eventually, we will get pretty close."
 
Cloud vendors large and small are pushing ahead with solutions for today's landscape and the building blocks of tomorrow's. Even Amazon itself, in the form of its AWS Cloud division, now exhibits at NAB with its partners, offering solutions of potential interest to the broadcast community. It offers solutions such as online storage, archiving, transcoding, and streaming. Smaller providers like Cloudsigma also presented interesting solutions for some of these challenges at NAB 2013, according to Kovalick, and he expects to see more of such providers and solutions in the future.
 
Kovalick, meanwhile, will be chairing an eight-paper session at the SMPTE 2013 Annual Technical Conference & Exhibition about the latest developments in the industry's goal of moving to an all-IT facility, and all the advantages that might bring, including Cloud networking for broadcasters.
 

News Briefs

State of 4K Ultra HD
Michael Grotticelli recently posted some interesting coverage about the state of the push toward widely implementing 4K Ultra High Definition (UHD) on the Broadcast Engineering website. In the article, he summarizes the Consumer Electronics Association's (CEA) newest report on the topic, dubbed "Ultra High Definition: State of the Industry."  Grotticelli suggests there is a gradually emerging market developing for 4K broadcast content designed for home viewing. He reminds readers that various technology companies are hard at work researching what he calls "next-generation compression and decompression chips" built around the HEVC codec for eventual implementation in 4K home viewing monitors. Additionally, "the CEA report cites the upscaling of high-definition resolution content and a trend toward more affordable pricing as two of the most important factors affecting the 4K market in these early days," Grotticelli writes. "Ultra HD displays can upscale HD video to Ultra HD resolution using video processing to fill in the extra resolution." Grotticelli also posted a second article that suggests the current broadcast industry infrastructure's limitations remains "the biggest hurdle to Ultra HD distribution." That article discusses the promise of HEVC, other developments, and interesting recent tests, but points out those developments are still a long way away from being implemented. UHD is the topic of the SMPTE 2013 Symposium: Next-Gen Image Formats: More, Better, or Faster Pixels? taking place in Hollywood, CA on 21 October 2013.
 
The First Laser Cinema
After a couple years of successful tests and major advances in the science of digital laser projection for cinema applications, what is being billed as the world's first permanent commercial digital laser projection installation at a major movie facility was announced this month. Christie Digital announced its technology had been selected for the installation at the Seattle Cinerama Theatre. According to Christie's release, the theater has purchased a 4K Christie Digital laser projection system with scalable light output up to 60,000 lumens. Christie's technology was the first system approved for a permanent theatrical installation, since it recently received the first U.S. FDA approval of variance to permit the sale of laser projectors deemed safe for use in such facilities. That was the first variance ever issued as of press time, but it is likely others will follow in the next several months as the laser movement picks up steam. Bryant Frazer at the Studio Daily site offers a report on the development and what we might expect next in the laser cinema rollout process. 
 
Holographic Headway
A fascinating recent article by teacher/artist Linda Law on the Digital Cinema Report site examines both the past and the future of digital holographic video presentations and dimensional imaging generally. Law reviews the entire history of the medium. From the time it was merely theoretical, to analog holograms, to the early days of digital holograms, to the spate of modern dimensional projection techniques and systems that have been used in recent years. One that received notoriety, in 2012, was the so-called "holographic" revival on a live stage of the late rap music artist, Tupac Shakur. Law also looks ahead to what might come next for the medium, including her thoughts on new recording materials, new tiny diode lasers and also pulsed laser systems that are opening up possibilities. Her primary prediction is that in the near future we will be seeing "holographic images that will be startling in their realism." As the technology improves, however, Law warns it won't mean much if it does not evolve along with a new generation of artists to help ensure the highest possible quality.