October 1, 2022
  • October 1, 2022

YouTube Argos ASIC chip for video transcoding

By on August 25, 2022 0

Hello and welcome to Protocol Enterprise! Today: why YouTube decided to invest millions in a custom video chip, Salesforce meets earnings expectations but cuts forecasts and Twitter security whistleblower report fallout continues.

See the future with Argos

Even though some chip industry insiders insist that a version of Moore’s Law is alive and well, Google vice president of engineering Partha Ranganathan helped place two big bets on the fact that the famous proclamation of 1965 is, in fact, dead. The first was the company’s AI chips, or tensor processing units (TPUs), which the company unveiled several years ago.

The second is his little discussed – and some might say underrated – YouTube chip, code name Argos.

  • Argos evolved from Project TPU, after Google engineers realized they could develop more chips to tackle other thorny computing problems beyond AI.
  • Video is an obvious choice since YouTube ingests and streams a lot of it, and it accounts for the majority of internet traffic worldwide.
  • “For something like transcoding, which is a very specific, high-intensity workload, they can get a tremendous amount of bang for their buck there,” said chip industry analyst Mike Feibus.
  • After a 10-minute meeting with YouTube CEO Susan Wojcicki, engineers under YouTube VP Scott Silver got the go-ahead.

Named after the Greek mythological monster with many eyes, the chip helps YouTube convert videos to all the different sizes and formats needed to play on thousands of devices.

  • Technically called an application-specific integrated circuit, or ASIC, the chip offers impressive performance: a 20 to 33 times increase over the general-purpose chips that YouTube previously used.
  • The design itself is what Ranganathan called “software-defined hardware”, meaning they were able to iterate on the chip design much faster than a normal hardware process.
  • The chip is largely focused on video transcoding, which means taking the raw download file and converting and compressing it into different formats so anyone can watch it.
  • Each video on YouTube has up to 15 separate copies in different formats, and around 500 hours of video are uploaded every minute.

YouTube’s decision to build its own video processors is part of a growing trend technology companies choosing off-the-shelf solutions from major chipmakers such as AMD, Intel and Nvidia are insufficient for their needs.

  • AWS built its Graviton server chips, Microsoft is reportedly working on its own server processors, Facebook has a chip unit, to name a few.
  • On the one hand, it is not a question of cost. Jay Goldberg, director at D2D Advisory, told us that building custom silicon is about creating strategic advantage.
  • With its YouTube chips, Google is trying to provide a better viewing experience – by converting more videos to formats that use less data, it can make them load faster and potentially better.
  • On the other hand, it is a question of money. Silver said the company certainly likes to save money, but what it really helps the company is to better allocate its resources.

— Max A. Cherney (email | Twitter)

DataRobot Sponsored Content

DataRobot’s AI cloud for financial services unlocks the art of the possible: DataRobot continues to attract customers in financial services who want to de-risk their AI investments and rapidly scale AI to nearly every part of their operations, resulting in improved productivity and higher customer satisfaction. clients.

Learn more about DataRobot

A little quieter in the second half

Surprise! Salesforce continues to grow, but announced on Wednesday that it will not meet its previous expectations for the second half of 2022.

The SaaS giant’s revenue rose 26% year-on-year to $7.72 billion for the quarter. But the company also lowered its guidance by a hair, down from a previous forecast of $31.7 billion to just $30.9 billion, as enterprise tech continues to search for signs that a slowdown in consumer spending is starting to trickle down to business spending.

Salesforce attributed the lower forecast to a mix of currency headwinds and subsequent shifts in customer buying habits. “The currency environment is obviously unprecedented,” co-CEO Marc Benioff said during the company’s earnings call.

But Salesforce’s performance across all of its clouds is still strong, and the company also announced its first-ever stock buyback of up to $10 billion.

The company’s Slack bet also appears to be paying off, with the messaging app propelling Salesforce’s platform business to revenue growth of more than 50% for the quarter. Salesforce’s data business, which includes Tableau and Mulesoft, had the weakest revenue growth of any cloud, but still hit $1 billion, meaning all Salesforce clouds topped the $1 billion revenue mark for the quarter.

Salesforce is still focused on growth, but careful about hiring and still has no major mergers and acquisitions plans, although Benioff teased the idea during the earnings call. “We found a great company…and that company is Salesforce,” he said. He has a point.

— Aisha counts (E-mail | Twitter)

“It’s work”

Many members of the cybersecurity community threw their support behind Peiter “Mudge” Zatko on Tuesday, following his whistleblower complaint accusing Twitter of ignoring an array of major cybersecurity issues (among others). Zatko, who was previously head of security for the social media platform, pointed to numerous shortcomings ranging from outdated software in Twitter’s data centers, non-existent monitoring for insider threats, lax controls over user access. employees in key systems.

Today, however, some have aired a contrary cybersecurity view on the subject.

“I don’t know a single working CISO who isn’t terrified of at least a dozen major issues needing immediate attention in the business — and frustrated with senior management,” wrote Edward Amoroso, former director of long-standing security at AT&T. , in a post on LinkedIn. “It’s work.”

The CEO of virtual CISO service provider SideChannel, Brian Haugli, shared similar thoughts during his conversation with me today. As a CISO, “your job is to dig up anything that could potentially be damaging, but then realistically put it within the expectations that [the executive team] can adapt,” he told me. “I don’t think there’s a single organization, even in the Fortune 500, that doesn’t have issues similar to what Twitter is going through.”

Regardless of what you think of those prospects (and Amoroso’s take on it got a lot of reactions, in the comments on his post and, of course, on Twitter), no one would argue that the apparent indifference of other leaders to Zatko’s findings is justified.

It’s just that instead of Twitter being great, “it’s the problem of security inside organizations,” as Haugli told me. “You’re constantly balancing what the business wants to be able to do, which is to go fast, and what security wants.”

—Kyle Alspach (E-mail | Twitter)

DataRobot Sponsored Content

DataRobot’s AI cloud for financial services unlocks the art of the possible: Banks need to secure a competitive advantage in an increasingly tight race to exploit the best technologies. Decision makers must not only plan a future-ready strategy, but also recognize the value of AI that could improve not only their performance internally, but also their reputation with their global customers.

Learn more about DataRobot

Thanks for reading – see you tomorrow!