 Google has decided that YouTube demands such a huge transcoding workload that it needs to build its own server chips. The company detailed its new "Argos" chips in a YouTube blog post, a CNET interview, and in a paper for ASPLOS, the Architectural Support for Programming Languages and Operating Systems Conference. Just as there are GPUs for graphics workloads and Google's TPU (tensor processing unit) for AI workloads, the YouTube infrastructure team says it has created the "VCU" or "Video (trans)Coding Unit," which helps YouTube transcode a single video into over a dozen versions that it needs to provide a smooth, bandwidth-efficient, profitable video site. Google's Jeff Calow said the Argos chip has brought "up to 20-33x improvements in computing efficiency compared to our previous optimized system, which was running software on traditional servers." The VCU package is a full-length PCI-E card and looks a lot like a graphics card. A board has two Argos ASIC chips buried under a gigantic, passively cooled aluminum heat sink. There's even what looks like an 8-pin power connector on the end because PCI-E just isn't enough power. Google provided a lovely chip diagram that lists 10 "encoder cores" on each chip, with Google's white paper adding that "all other elements are off-the-shelf IP blocks." Google says that "each encoder core can encode 2160p in real-time, up to 60 FPS (frames per second) using three reference frames." The cards are specifically designed to slot into Google's warehouse-scale computing system. Each computes cluster in YouTube's system will house a section of dedicated "VCU machines" loaded with the new cards, saving Google from having to crack open every server and load it with a new card. Google says the cards resemble GPUs because they are what fit in its existing accelerator trays. CNET reports that "thousands of the chips are running in Google data centers right now," and thanks to the cards, individual video workloads like 4K video "can be available to watch in hours instead of the days it previously took." Factoring in the research and development on the chips, Google says this VCU plan will save the company a ton of money, even given the below benchmark showing the TCO (total cost of ownership) of the setup compared to running its algorithm on Intel Skylake chips and Nvidia T4 Tensor core GPUs. Because YouTube is the world's biggest video site, keeping it running was originally seen as an impossible task until Google bought the company in 2006. Since then, Google has aggressively fought to keep the site's cost down, often reinventing Internet infrastructure and copyright in order to make it happen. Today, the primary infrastructure problem YouTube needs to solve for end-users is providing video that works just right for your device and bandwidth while maintaining quality. That means using a codec that is supported by your device and picking a resolution that matches your display (and not blowing up your Internet connection with a massive file). For Google, that means transcoding a single video into a lot of other videos. You can see part of this work yourself just by clicking on the gear for an 8K video, where you'll see nine total resolutions created from a single upload: 144p, 240p, 360p, 480p, 720p, 1080p, 1440p, 2160p, and 4320p. These are all different video files, and everyone needs to be created from the original 8K uploaded file—and keep in mind, this is just for your specific device. Google also needs to offer some of those nine resolutions in multiple codecs, which dictate how the video is compressed on its way over the Internet. The company wants to offer videos in the most advanced, efficient codec available to save on bandwidth, which is a massive part of YouTube's costs. Decoding a video codec gobbles up processing power, though, and on cheaper mobile devices, decoding won't happen smoothly and efficiently without dedicated hardware acceleration support for each new codec. That means Google only gets to use the best codecs on new devices, and it needs to keep copies of the video around in older codecs for older devices. Today, modern devices usually get the efficient VP9 codec, while the more compatible H.264 is kept around for devices that aren't on the cutting edge. No one truly knows the depths of YouTube's video codec selection, but the site also generally supports devices going back almost 10 years, including "low-resolution flip phones," according to the ASPLOS paper. So there are some pre-H.264 codecs, like 3GP, for ancient devices. Learn more by visiting OUR FORUM.
 Executives at Instagram are planning to build a version of the popular photo-sharing app that can be used by children under the age of 13, according to an internal company post obtained by BuzzFeed News. “I’m excited to announce that going forward, we have identified youth work as a priority for Instagram and have added it to our H1 priority list,” Vishal Shah, Instagram’s vice president of product, wrote on an employee message board on Thursday. “We will be building a new youth pillar within the Community Product Group to focus on two things: (a) accelerating our integrity and privacy work to ensure the safest possible experience for teens and (b) building a version of Instagram that allows people under the age of 13 to safely use Instagram for the first time.” The current Instagram policy forbids children under the age of 13 from using the service. According to the post, the work would be overseen by Adam Mosseri, the head of Instagram, and led by Pavni Diwanji, a vice president who joined parent company Facebook in December. Previously, Diwanji worked at Google, where she oversaw the search giant’s children-focused products, including YouTube Kids. The internal announcement comes two days after Instagram said it needs to do more to protect its youngest users. Following coverage and public criticism of the abuse, bullying, or predation faced by teens on the app, the company published a blog post on Tuesday titled “Continuing to Make Instagram Safer for the Youngest Members of Our Community.” That post makes no mention of Instagram’s intent to build a product for children under the age of 13, but states, “We require everyone to be at least 13 to use Instagram and have asked new users to provide their age when they sign up for an account for some time.” The announcement lays the groundwork for how Facebook — whose family of products is used by 3.3 billion people every month — plans to expand its user base. While various laws limit how companies can build products for and target children, Instagram clearly sees kids under 13 as a viable growth segment, particularly because of the app’s popularity among teens. In a short interview, Mosseri told BuzzFeed News that the company knows that “more and more kids” want to use apps like Instagram and that it was a challenge verifying their age, given most people don’t get identification documents until they are in their mid-to-late teens. “We have to do a lot here,” he said, “but part of the solution is to create a version of Instagram for young people or kids where parents have transparency or control. It’s one of the things we’re exploring.” Mosseri added that it was early in Instagram’s development of the product and that the company doesn’t yet have a “detailed plan.” Priya Kumar, a Ph.D. candidate at the University of Maryland who researches how social media affects families, said a version of Instagram for children is a way for Facebook to hook in young people and normalize the idea “that social connections exist to be monetized.” “From a privacy perspective, you're just legitimizing children’s interactions being monetized in the same way that all of the adults using these platforms are,” she said. Kumar said children who use YouTube Kids often migrate to the main YouTube platform, which is a boon for the company and concern for parents. “A lot of children, either by choice or by accident, migrate onto the broader YouTube platform,” she said. “Just because you have a platform for kids, it doesn’t mean the kids are going to stay there.” The development of an Instagram product for kids follows the 2017 launch of Messenger Kids, a Facebook product aimed at children between the ages of 6 and 12. After the product’s launch, a group of more than 95 advocates for children’s health sent a letter to Facebook CEO Mark Zuckerberg, calling for him to discontinue the product and citing research that “excessive use of digital devices and social media is harmful to children and teens, making it very likely this new app will undermine children’s healthy development.” Facebook said it had consulted an array of experts in developing Messenger Kids. Wired later revealed that the company had a financial relationship with most of the people and organizations that had advised on the product.Details can be found on OUR FORUM.
 Google is going it alone with its proposed advertising technology to replace third-party cookies. Every major browser that uses the open-source Chromium project has declined to use it, and it’s unclear what that will mean for the future of advertising on the web. A couple of weeks ago, Google announced it was beginning to test a new ad technology inside Google Chrome called the Federated Learning of Cohorts or FLoC. It uses an algorithm to look at your browser history and place you in a group of people with similar browsing histories so that advertisers can target you. It’s more private than cookies, but it’s also complicated and has some potential privacy implications of its own if it’s not implemented right. Google Chrome is built on an open-source project, and so FLoC was implemented as part of that project that other browsers could include. I am not aware of any Chromium-based browser outside of Google’s own that will implement it and very aware of many that will refuse. One note I’ll drop here is that I am relieved that nobody else is implementing FLoC right away, because the way FLoC is constructed puts a very big responsibility on a browser maker. If implemented badly, FLoC could leak out sensitive information. It’s a complicated technology that does appear to keep you semi-anonymous, but there are enough details to hide dozens of devils. Anyway, here’s Brave: “The worst aspect of FLoC is that it materially harms user privacy, under the guise of being privacy-friendly.” And here’s Vivaldi: “We will not support the FLoC API and plan to disable it, no matter how it is implemented. It does not protect the privacy and it certainly is not beneficial to users, to unwittingly give away their privacy for the financial gain of Google.” As you probably know, Opera has a long history of introducing privacy features that benefit our users: it was the first major browser to introduce built-in ad blocking, browser VPN and other privacy-centric features. The significance now is the end of third-party cookies, which will reduce the amount of cross-website tracking on the web. While we and other browsers are discussing new and better privacy-preserving advertising alternatives to cookies including FloC, we have no current plans to enable features like this in the Opera browsers in their current form. Generally speaking, we do, however, think it’s too early to say in which direction the market will move or what the major browsers will do. DuckDuckGo isn’t thought of as a browser, but it does make browsers for iOS and Android. On desktop, it’s already made a browser extension for other browsers to block it. And the Electronic Frontier Foundation, which is very much against FLoC, has even made a website to let you know if you’re one of the few Chrome users who have been included in Google’s early tests. But maybe the most important Chromium-based browser not made by Google is Microsoft Edge. It is a big test for Google’s proposed FLoC technology: if Microsoft isn’t going to support it, that would pretty much mean Chrome really will be going it alone with this technology. As for Apple’s Safari, I will admit I didn’t reach out for comment because at this point it’s not difficult to guess what the answer will be. Apple, after all, deserves some credit for changing everybody’s default views on privacy. However, the story here is actually much more interesting than you might guess at first. John Wilander is a WebKit engineer at Apple who works on Safari’s privacy-enhancing Intelligent Tracking Prevention features. Wilander’s reply jibes with Microsoft’s statement that “the industry is on a journey” when it comes to balancing new advertising technologies and privacy. But it speaks to something really important: web standards people take their jobs seriously and are seriously committed to the web standards process that creates the open web. Read this posting in its entirety on OUR FORUM.
|
Latest Articles
|