Home > News content

Cloud service set up, AI singing, this is a Google technology conference not inferior to I / O

via:博客园     time:2018/7/25 21:32:03     readed:691

Even people who think they are very familiar with Google may not have heard the name of the Cloud Next conference.

This is outside of Google I/O, and the company has another countless developer event. On July 24, local time, one and a half months after the end of I/O, 25,000 Google Cloud developers met again in San Francisco.


"Cloud" potential in Q2 earnings

Prior to the launch, Google's parent company, Alphabet, announced its 2018 Q2 earnings on Monday, excluding revenue from acquisition costs of $26.2 billion, a 26% increase from the same period last year; excluding EU's $5 billion fine, Q2 net The profit is $3.2 billion.

A few months ago, Google Cloud CEO Diane Greene told CNBC that Google’s cloud business was underestimated. In addition to the mobile advertising business, the expansion of the cloud computing business has contributed to the revenue growth brought by Q2. Although Google did not disclose specific cloud business revenues, revenue from “other revenues” (other business revenues) covering cloud services reached US$4.425 billion, an increase of 36.5% over the same period.


Diane Greene

At the same time of growth, Google is also increasing its investment in cloud computing. As of Q2, the company's capital expenditure was $5.3 billion, equivalent to twice the $2.48 billion in the same period last year. In addition, there are 890,58 employees in Alphabet, compared with 75,606 in the same period last year. Most of the increase in the number of employees comes from engineers and sales people in the cloud computing business. According to Diane Greene, at the press conference, in order to provide better cloud services, Google The number of engineers at Cloud has tripled, more than sales people.

"Hybrid Cloud" Google Cloud Platform

With the new Google Cloud Platform platform, Google is providing a hybrid cloud service solution for businesses that can deploy on both the Google Cloud Platform and the enterprise internal environment.

Istio is a microservices management framework that allows operators to have more control over traffic management, security, and troubleshooting. Google is about to release Istio 1.0. Google Kubernetes Engine is now available for internal infrastructure, allowing users to update applications without having to migrate to the cloud. The idea behind the Cloud Service Platform is to let companies update their stacks at their own pace. In addition, Google Cloud Platform tools are tightly integrated with other services in cloud computing, such as machine learning tools.

Google announced a new CI/CD (Continuous Integration/Continuous Delivery) platform ——Cloud Build that allows customers to quickly build, test and deploy software on a large scale. Cloud Build works in a variety of environments, such as VMs, serverless, Kubernetes, and Firebase. More importantly, Cloud Build supports Docker containers and provides developers with an increasingly automated way to build, test, and deploy.

Data security and AI become the main theme of G Suite updates

The most recent link between the entire conference and the average user is the Office Suite G Suite update and new product release.


Sundar Pichai, Google CEO who attended and delivered a speech

G Suite is Google's suite of productivity tools for corporate offices, including Gmail, Docs, and Calender, which are well known to domestic users. In the past year, Google has added about 300 new features to G Suite. At today's conference, data security and AI became the main theme of G Suite's new features.

In terms of security, Google released the Security Center's investigation tool, which allows administrators to identify and manage malicious sharing and attacks in users and files in batches. The Data Region feature allows administrators to Files are set to be used only in the region (such as the US, Europe, etc.) to help managers better control data security.

In terms of artificial intelligence, the first Smart Reply and Smart Compose features released on Google I/O have more applications.

Smart Reply was first applied in a Gmail mailbox. According to Google’s live data, more than 10% of all emails on the entire platform use this feature to assist in responding. Now, Google has extended this feature to even the chat app Hangouts Chat, the AI ​​will recognize the way and habits of the user's chat, and provide users with smart and optional quick reply recommendation words.

In terms of improving efficiency, Smart Compose, which assists in email writing, may perform better than a simple chat response. Smart Compose is now open to G Suite customers and the upgrade will be pushed to customers in the next few weeks.

At the press conference, the new feature of G Suite is the document collaboration service Docs. For the convenience of writing and reading, Google Docs has introduced grammar suggestion. This feature uses machine translation-based methods to identify grammatical errors and correct them, from when to use "a" and "an" to how to use the complex concepts of dependent clauses.

In addition, Google has introduced voice commands for hardware users of Hangout Meet.

"people-oriented" AI

The last appearance of Professor Li Feifei brought the latest progress of Google Cloud in AI and machine learning, which is the second time she appeared at the Google Cloud conference. After joining Google Cloud for more than a year, she has almost become the most authoritative spokesperson for Google Cloud AI.


Dr. Li Feifei

At this Cloud Next conference, Li Feifei released her and her team's latest work, including Cloud AutoML, AI Service Center and more.

Pre-selected, trained domain-specific machine learning models are among the most popular AI features in cloud services. Cloud AutoML, launched earlier this year, is Google’s service in this area.

Today, Google launched the official version of Cloud AutoML Vision, which packages the Cloud Vision API to enterprise users. In addition, AutoML adds two new features: AutoML Natural Language and AutoML Translation, which provide users with a custom training model for custom text and language translation.

"At this point, AutoML can provide automated image generation, language classification and translation capabilities for companies and developers with AI experience in various industries. AutoML has more than 18,000 registered users, service industry across media, retail, finance. , insurance, energy, medical care, the environment, etc. The products and innovations are interlocking, and the cooperation also brings progress. I hope we can continue to improve and bring the best technology to everyone." Dr. Li Jia, Director of R&D of Google Cloud AI, said .

Google provides developers with the tools to create a unique session interface, but this is just the beginning of a conversion service center. Google believes that companies need a more comprehensive solution, integrated automation and powerful analytics. To meet this need, Google Cloud has launched Contact Center AI, a set of virtual assistants, intelligent information mining and emotions. The first solution product (industry solution product) that helps the Contact Center staff to more effectively solve problems and enhance the user experience.

"Compared to some of the basic skills that replace humans, we believe that artificial intelligence has more unlimited potential in enhancing human skills." At the end of the speech, Professor Li Feifei said.

Google Cloud Next in the Cloud Age

Most of the new products released at the Cloud Next conference are far away from the average user. It is the way that Google expresses its attitude to developers in the era of cloud computing. — I/O is not enough to expose the protagonist. Google’s cloud service requires a special session.


For those who want to know Google, Cloud Next is still a technology event worthy of attention, such as CNET's technology journalist Stephen Shankland said on Twitter: "Google Cloud Platform is particularly attractive to me because it is a Paid products: There is no privileged usage fee split between advertisers and users, and there is no bias to one party to harm the other. Paying customers means that Google's purpose is very consistent: to make products better."

A few years ago, when Google just started to force cloud services, many people were not optimistic about this. At that time, the market belonged to Amazon and Microsoft. People told CEO Diane Greene that Google would take at least ten years to prepare for corporate users.

"Microsoft has spent so long, I said, no, no need for ten years." In an interview with foreign media TechCrunch, Greene said. At the time, the Google Cloud Developer Conference was only about 2,000 people. With pressure, Google Cloud came along, and they had already prepared their own business services.

In an interview, Diane Greene said that she hates being called "the third place after AWS and Azure". Google Cloud has greater ambitions and strengths that go far beyond the current market share. potential.

When Amazon and Microsoft's cloud services are becoming the company's most profitable business, Google will not be willing to be the "third in the distance."

Source of the map: Google

Editor in charge: Song Desheng

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments