#GoogleCloudOnBoard
蠻有意思的一個seminar，這類型的seminar其實還蠻不錯的。我們很努力的想說8.30去然後拿Free Gift，怎知8.35抵達的時候Free Gift已經派完了，而且已經看到非常多的人正在交談和吃早餐。這，10點開始的seminar,8.30開始的registration,8.35分竟然就擠滿了人是怎麼回事。Google的安排其實是讓我們有機會跟不同的人交流，他們明顯的安置了幾個站點，讓有興趣的人上前聚起來問起來，看著看著，感覺這根本是IT怪獸們的聚會。有的躲在一旁自己拼命的吃，有的站在一旁自己看著書，有的竟然打開電腦工作，有的一班玩鬧著，然後有許多的在交換名片。
出去走走看看還是需要的，這世界多大呀看了才提醒自己多渺小，該努力的。
想說，他們的食物真的超好吃！
然後以下是大長篇，formal式報告。
Event Date: 17 April 2018 08:30 AM - 05:30 PM
Location: Sands Expo and Convention Centre
Attendee: Yin² & Jessie
This is an event hosted by Google - Google Cloud Platform. Through out the whole session, they will drill into details on what they have in Google Cloud Platform (aka GCP).
First of all, they are having a very clear directions on how/what/where to grow for GCP - globally availability. And therefor they are introducing this platform and sharing the knowledge of this platform loudly - everything to & from cloud - the data, storage, data analysis and machine learning services.
Notes: Google having a console named as Cloud Shell for command lines operation - this tools is coming with pre-installed libraries for users to interact with Google Cloud Platform.
Compute & Storage
Even there are products that look similar, but they are actually targeting different demands. below is the advised storage solution based on a business's needs.
Cloud Data Store - Cloud Server
From Google Cloud console, you can simply create a Virtual Machine with any size & memories in different location you wants - a 30GB for SEA, a 30GB for London, database/application distributed across different regions will help in reducing latency & increase the global availability. Users can create and remove them within seconds.
Use cases: startup, app engine applications.
Cloud Data Store - Cloud filesystem storage
Similar as Amazon cloud file storage, we will be able to create bucket and store related items/objects into the bucket - this is up to us on how do we categorize buckets. And they are having a console that you may copy files from a bucket/machine into another bucket by using commands.
Use cases: media files, backups
Cloud SQL
A fully-managed database service (MYSQL or Postgress) that supports backups, fast connection & worldwide connection, and Google security hosted on GCP. HIghly accessible from any applications.
use cases: users credentials
Bigtable
Google Bigtable is a scalable NoSQL database service which designed to handle massive workloads with low latency and high throughput. It is supporting volume up to petabytes of data. From the demo, the instructor manage to pull a 5gb of data within 7seconds.
use cases: Advertising u, IoT (https://internetofthingsagenda.techtarget.com/definition/Internet-of-Things-IoT)
Cloud Spanner
A relational database service that support horizontal scaling (rows,regions,continents) with transactional consistency. It could handle schema changes without downtime across the databases distributed to different regions.
Use cases: High I/O,global consistency required, Logistics And Transportations, Global Call Centers
BigQuery
Serverless, scablable data warehouse, designed to focus on real time data analytics, support up to petabytes of data. BigQuery will be able to process data in Cloud Storage, BigTable and even spreadsheets in Google Drive without a human consolidation required. BigTable also a foundation for Machine Learning and Artificial Intelligence with integrating to CloudML Engine and TensorFlow for user to train a model from a structured data - basically they are providing the abilities to learn & analyze data, and transform data that help for Machine Learning. With security concerns, BigQuery support a access control through Google Cloud IAM (Reference: https://cloud.google.com/iam/)
Use cases: Data Warehousing, Real-time inventory Management System
Machine Learning
After everything ready in cloud, what's next? Use them. Experience increase the accuracy - this is why ML requires massive compute.
The session did introduce Google ML REST APIs:
Ending
When data store solution on cloud, the text things is "Development". So Google claims that with the above solution, as a business will only needs to focus in Programming.
Focus on insight, not infrastructure, because they handling everything.
Others
Our neighbor is a person in-charging on standardize business applications & hardware, shared that their company is using multiple cloud services in the same time for different purposes. And most of the time their web development team will be asking for a test environment for certain project, and Gan's team will be arranging a cloud data storage for their development & remove after deployment - this reduce the server demand, maintenance onto a physical server, time consuming arrangement to setup a test database only for development purpose.
For opening, WEGO team been invited to share their products and their experience in using Google Cloud Platform for their solution. WEGO - a travel search websites, is using GCP for data analysis, machine learning for demand prediction, searching and comparing results, calculate the rates of "you might also interest to..." from airlines, hotel and travel agency. They doing these through data, but they clarified - it's should be from CLEAN DATA. With clean data they believe it will increase the accuracy of the demand predictions.
Reference of WEGO engineers websites: https://geeks.wego.com/
Referece:
Conference Document: https://drive.google.com/file/d/12I2lqVr4SAWSGF-XYKMjs96UTVLfczKk/view
蠻有意思的一個seminar，這類型的seminar其實還蠻不錯的。我們很努力的想說8.30去然後拿Free Gift，怎知8.35抵達的時候Free Gift已經派完了，而且已經看到非常多的人正在交談和吃早餐。這，10點開始的seminar,8.30開始的registration,8.35分竟然就擠滿了人是怎麼回事。Google的安排其實是讓我們有機會跟不同的人交流，他們明顯的安置了幾個站點，讓有興趣的人上前聚起來問起來，看著看著，感覺這根本是IT怪獸們的聚會。有的躲在一旁自己拼命的吃，有的站在一旁自己看著書，有的竟然打開電腦工作，有的一班玩鬧著，然後有許多的在交換名片。
出去走走看看還是需要的，這世界多大呀看了才提醒自己多渺小，該努力的。
想說，他們的食物真的超好吃！
然後以下是大長篇，formal式報告。
Event Date: 17 April 2018 08:30 AM - 05:30 PM
Location: Sands Expo and Convention Centre
Attendee: Yin² & Jessie
This is an event hosted by Google - Google Cloud Platform. Through out the whole session, they will drill into details on what they have in Google Cloud Platform (aka GCP).
First of all, they are having a very clear directions on how/what/where to grow for GCP - globally availability. And therefor they are introducing this platform and sharing the knowledge of this platform loudly - everything to & from cloud - the data, storage, data analysis and machine learning services.
Notes: Google having a console named as Cloud Shell for command lines operation - this tools is coming with pre-installed libraries for users to interact with Google Cloud Platform.
Compute & Storage
Even there are products that look similar, but they are actually targeting different demands. below is the advised storage solution based on a business's needs.
Cloud Data Store - Cloud Server
From Google Cloud console, you can simply create a Virtual Machine with any size & memories in different location you wants - a 30GB for SEA, a 30GB for London, database/application distributed across different regions will help in reducing latency & increase the global availability. Users can create and remove them within seconds.
Use cases: startup, app engine applications.
Cloud Data Store - Cloud filesystem storage
Similar as Amazon cloud file storage, we will be able to create bucket and store related items/objects into the bucket - this is up to us on how do we categorize buckets. And they are having a console that you may copy files from a bucket/machine into another bucket by using commands.
Use cases: media files, backups
Cloud SQL
A fully-managed database service (MYSQL or Postgress) that supports backups, fast connection & worldwide connection, and Google security hosted on GCP. HIghly accessible from any applications.
use cases: users credentials
Bigtable
Google Bigtable is a scalable NoSQL database service which designed to handle massive workloads with low latency and high throughput. It is supporting volume up to petabytes of data. From the demo, the instructor manage to pull a 5gb of data within 7seconds.
use cases: Advertising u, IoT (https://internetofthingsagenda.techtarget.com/definition/Internet-of-Things-IoT)
Cloud Spanner
A relational database service that support horizontal scaling (rows,regions,continents) with transactional consistency. It could handle schema changes without downtime across the databases distributed to different regions.
Use cases: High I/O,global consistency required, Logistics And Transportations, Global Call Centers
BigQuery
Serverless, scablable data warehouse, designed to focus on real time data analytics, support up to petabytes of data. BigQuery will be able to process data in Cloud Storage, BigTable and even spreadsheets in Google Drive without a human consolidation required. BigTable also a foundation for Machine Learning and Artificial Intelligence with integrating to CloudML Engine and TensorFlow for user to train a model from a structured data - basically they are providing the abilities to learn & analyze data, and transform data that help for Machine Learning. With security concerns, BigQuery support a access control through Google Cloud IAM (Reference: https://cloud.google.com/iam/)
Use cases: Data Warehousing, Real-time inventory Management System
Machine Learning
After everything ready in cloud, what's next? Use them. Experience increase the accuracy - this is why ML requires massive compute.
The session did introduce Google ML REST APIs:
- Use your own data to train models
TensorFlow & Cloud Machine Learning Engine allows you to create, and test new machine learning methods. A testing tools to try machine learning data flow: http://playground.tensorflow.org - Machine Learning API
Cloud Vision API, SpeechAPI, Natural LanguageAPI, Translation API and Video Intelligence.
Example (i) use Natural Language API to gather the customer feedbacks and to analyze sentiments of their users.
Example (ii) use Vision and Translation API to translate a text from Images.
Example (iii) use Speech API to convert audio to text
Ending
When data store solution on cloud, the text things is "Development". So Google claims that with the above solution, as a business will only needs to focus in Programming.
Focus on insight, not infrastructure, because they handling everything.
Others
Our neighbor is a person in-charging on standardize business applications & hardware, shared that their company is using multiple cloud services in the same time for different purposes. And most of the time their web development team will be asking for a test environment for certain project, and Gan's team will be arranging a cloud data storage for their development & remove after deployment - this reduce the server demand, maintenance onto a physical server, time consuming arrangement to setup a test database only for development purpose.
For opening, WEGO team been invited to share their products and their experience in using Google Cloud Platform for their solution. WEGO - a travel search websites, is using GCP for data analysis, machine learning for demand prediction, searching and comparing results, calculate the rates of "you might also interest to..." from airlines, hotel and travel agency. They doing these through data, but they clarified - it's should be from CLEAN DATA. With clean data they believe it will increase the accuracy of the demand predictions.Reference of WEGO engineers websites: https://geeks.wego.com/
Referece:
Conference Document: https://drive.google.com/file/d/12I2lqVr4SAWSGF-XYKMjs96UTVLfczKk/view
Tags:
betterme
