Compared with last year’s AWS re:Invent 2020, which was held entirely online, this year’s AWS re:Invent 2021 recovery entity will be held in Las Vegas with an online Virtual venue. Although I have been invited by the AWS Heroes program to go to the site, but many things at the end of the year are really difficult to escape, and I finally chose to participate online, without the real interaction with people on site, but my whole heart is still in the various keynote speeches at re:Invent (keynote ) and on the agenda.
Continuing the Think in Context: AWS re:Invent 2020 Andy Jassy Keynote last year, it seems to be well received by everyone. This year, I specially set aside time to sort out the first AWS re:Invent 2021 Adam Selipsky Keynote of newly appointed AWS CEO Adam Selipsky. At the end of this thread dismantling note, try to make some comparisons with last year’s Andy Jassy Keynote (no comparison, no harm, huh?! :p I’m not interested in the content, if you want to see the comparison and analysis directly, you can also skip to the end of the article to read .
Continuing last year, Andy Jassy mentioned “Still early days for cloud”. The error of a few hours is acceptable in the entire cloud world
(not in the cryptocurrency circle (XDD). So press a cup of coffee, make a pot of hot tea, and see The rebroadcast at 8:30AM in Taiwan time zone is very sober and comfortable.
As in the past, at the beginning of the full text, I will first try to grasp the structure of the speech, then put some observations and inferences, and then put some running notes in each paragraph for future search and use. There are extended readings at the end of the article, which can enrich the background situation or information of the speech content. I also hope to have the opportunity to lead everyone to reason together why this product or function was launched at this time point, where the trends of various industries around the world are going. Read more, compare more, and refer more. Maybe we can avoid some minefields of misjudgment. In the case of limited resources, if we go in the wrong direction, the market may not give us a chance to cut it off and practice again. My reasoning is not necessarily correct, just as a kind of practice and sharing.
New services or new launches are marked with
[NEW 🚀] in this article, so that you can press Command/Ctrl+F to search on the inner page.
This article deliberately removes most of the product links first, so that everyone can focus on reading (we are all less focused these days, aren’t we?). If you need a product link, you can refer to the AWS Product List that I organize regularly.
I also welcome everyone to give me some feedback or corrections (also come and see if there is a chance to include these materials in ChatGPT 2022). Then let’s get started!
The structure of Adam Selipsky Keynote’s two-hour keynote speech this time seems to have few newly released products, and only a little bit here and a little bit there, but in fact, he wants to undertake the three-hour content of Andy Jassy Keynote last year, And at the same time reduce the coverage within two hours, Adam is bound to make a trade-off, and speculates that he may have arranged to distribute some new product announcements to the next three Keyntoes: “Swami Sivasubramanian Keynote” covers machine learning (Machine Learning) Learning; ML), “Peter DeSantis Keynote” covers infrastructure (Infrastructure; Infra) and “Werner Vogels Keynote” covers technology, architecture, thinking and other directions.
Nevertheless, I roughly grasped that Adam’s keynote speech is actually a theme and five paragraphs.
A theme “Pathfinder” ran through the audience, and even customer case sharing echoed pathfinding, and the five paragraphs were:
- The Instance (compute)
- Edge of the cloud (edge)
- The data journey (data)
- Data was an afterthought (data)
- New paths for industrials (industry applicaion)
I think Adam Selipsky’s method is to bring a lot of stories into the speech. Even though he did not explain the whole process, some of the photos presented in the background projection film are very story-telling. This way of presentation must be able to shape Next, Adam wants to bring you a refreshing AWS CEO style. The overall structure follows the five paragraphs of last year, but after introducing Adam’s story, the order of the narrative is adjusted, which may show the direction AWS will look at the market in the coming year:
This direction reminds me of…
Back then, AWS was preparing to deconstruct the data center,
and this year AWS was preparing to deconstruct the telecom network
The edge scene (Edge of the cloud) needs to integrate computing (compute) and data (storage/database) in the field, and also needs to take into account the communication transmission (network), which is currently the place where everyone is fighting. From the fifth paragraph last year, Adam directly pulled to the second paragraph to make a statement, and also put here the brightest new product announcement of the entire keynote speech - 5G as a service (5aaS ?!).
At that time, everyone did not necessarily have enough resources to build a complete set of equipment and structure of the data center by themselves.
It does not matter that AWS rents it to you (pay as you go).
This year, you may not have enough resources to build a complete set of equipment and architecture for your own telecom network.
It does not matter if AWS rents it out to you.
Why did I say that Adam Selipsky took over Andy Jassy’s keynote speech last year? Last year Andy opened with “Still early days for cloud”, this year Adam took over with “It's still early days”. (I really admire the ability of this company to arrange keynote speech content! 😍
This year is not only the 15th anniversary of AWS, but also the 10th anniversary of AWS re:Invent :)
Here’s a good time to start with a look back in history 1 video.
- AWS 15th Anniversary - Celebrate EC2’s fifteenth birthday this summer on August 25th.
- 2006 - AWS launches S3 (Storage)
- 2007 - EC2 public-beta released (Compute)
- 2007 - AWS launches EMEA region
- 2008 - AWS Elastic Beanstalk launched
- 2009 - Amazon RDS launch (Database)
- 2010 - AWS launches Asia Pacific region
- 2011 - Amazon DynamoDB released
- 2012 - AWS Redshift released - The first re:Invent annual meeting was held in the same year. At that time Netflix announced plans to migrate 100% of Netflix infrastructure to AWS.
- 2013 - AWS launches in China
- 2014 - AWS Aurora released
- 2015 - Amazon Elastic Container Registry (ECR) announced
- 2016 - AWS announces Snowmobile - The entire oversized container truck drove into the venue to help you carry the materials.
- 2017 - Amazon SageMaker released
- 2018 - DeepRacer released
- 2019 - AWS launches Cape Town region
- 2020 - AWS launches EC2 MAC instances
- 15 years of innovation
- The design plan of re:Invent is a learning conference, which mainly consists of the following two elements:
- Fifteen years ago IT was lost by some old-guard vendors who loved the expense and lock-in
- data demonstration
- Technical features
- S3 stores 100 trillion objects
- More than 60 million new instances everyday
- More than 200 products and services
- Various industrial customers and various customer application scenarios
- 100,000+ Partners
- Operational experience
- 81 availability zones and 25 geographic regions
- Technical features
- 「It's still early days」
- 5-15% IT spending has moved to the cloud. （📝 Last year’s figure was 4%）
- Edge of the cloud: 5G and IoT.
- Data Analytics and ML
- Netflix announced its all-in AWS in 2012.
- NASA’s Jet Propulsion Laboratory (JPL) used AWS services to land the Mars rover Curiosity on Mars in 2012.2
- Japan’s NTT docomo uses Redshift to start peta byte level storage data analysis, and the query speed is ten times faster than before.
- Pathfinder patterns:
- Constant pursuit of a better way.（change the game）
- Ability to see what others don't.（Apply knowledge, find opportunity）
- Enable others to forge their own paths. (influence others, make a difference）
- Basketball in the 1930s
- Explain the rules of the basketball game back then
- Hank Luisetti developed the running one-handed shot (an early version of the jump shot). The record set by Hank has been unbroken in more than 83 years at Stanford.6
- A better way forward
- Change the game
Guest: Adena T. Friedman, President & CEO, Nasdaq
- Nasdaq was founded 50 years ago.
- (📝 It seems reasonable for a 15-year-old AWS to find a 50-year-old Nasdaq to share his experience in path finding? !)
- Scale, Reliability, Speed
- End-to-end order-to-trade processing occurs in less than 20 microseconds.
- 3M messages per seconds in each of Nasdaq matching servers.
- Nasdaq also sells AWS-based SaaS solution
- More than a market
- Market integrity
- Anti-financial crime
- Trade surveillance
- Marketplace infrastructure
- Trading and matching
- Risk management
- Other marketplace services
- Buyside intelligence
- Data & analytics
- Corporate and board services
- IR management
- 2008 - Market replay launched
- A lot of data, but not a real-time application.
- Adena joked that the first AWS bill was only about USD 20, so it was a good relationship with AWS from then on（📝 We also want a bill that is only USD 20 ahhhhhhh XDD
- Systematic move to AWS
- Data distribution
- Revenue management
- Regulatory reporting
- Market operations systems
- Matching engine 大魔王!!
- Together with AWS, Nasdaq will migrate the first US options market into the AWS cloud in 2022.
- （📝 So some fund company hosts that run quantitative strategies will also start to move to the cloud or AWS computer room? !）
- Private Local Zone! (for global capital markets) 7
- （📝 Sure enough, it appeared! Place Outposts at Nasdaq’s main computer room in Carteret, New Jersey）
- Nasdaq: Market expertise and technology
- AWS: Edge computing with AWS Outposts
- The markets served by the AWS region and Nasdaq have a high degree of geographic overlap.
- He turns a little harder this way:
- Technology provides greater access to markets.
- The cloud is the center of technology innovation.
- Market access providers a better path to prosperity.
- Nasdaq and AWS are moving the markets to the cloud.
- But this time the Nasdaq CEO is not as brave as the Netflix CEO ten years ago to talk about all-in AWS, only saying 100% cloud-enabled. This also echoes the concerns mentioned by Andy Jassy when he said “Still early days for cloud” last year.
1. The Instance
- The Instance
- EC2 instances
- Start with CPU, memory
- Initially only M1 instance（📝 So everyone starts with Model-001 (M1)? ! Well, I admit that the first Bluetooth headset module made by our team is also following this path :p
- Then support Windows
- Then support Auto scaling, load balancing
- Then you want more
- Compute optimized
- Memory optimized
- Storage optimized
- 475 instance types have grown so far
- by categories
- by capabilities
- by options
- AWS rethinks The Instance, and rethinks it from the bottom, starting from the silicon level, so AWS built its own ARM-based processor chip, which is Graviton.
- After the launch of Graviton, everyone also said that there is more…
- So we ushered in the second self-made processor chip Graviton2, which was used in big data analytics, game servers and high-performance computing
- Today we bring you the third self-made processor chip, Graviton3!（📝 It’s been a long time to spread the stalks XDD
- [NEW 🚀] Graviton3
- 25% faster than Graviton2
- 2x faster floating point performance for science workloads
- 2x faster for cryptographic workloads
- 3x faster for ML workloads
- Uses up to 60% less energy
- [NEW 🚀] C7g Instances for EC2 (Preview available today)
- Powered by Graviton3
SAP HANA Cloud also started using Graviton processors.
But everyone still wants more (more)… Inf1 Instance for EC2 (70% lower cost per inference) published in 2019, plus Trainium published last year, so this year brings you:
- [NEW 🚀] Trn1 Instances for EC2 (Preview available today)
- Use Trn1 to train ML models for myriad applications
- 800 Gbps network bandwidth
- Combine 10s of thousands of instances for the most complex training （打造成 ultra clusters
So there is the best cost-effective training instance Trn1 and inference instance Inf1.
- But can all workloads really be moved to the cloud? So what about mainframes 8?
- Legacy programming
- He used the old basketball style just now as a metaphor
- [NEW 🚀] AWS Mainframe Modernization (Preview available today)
- Migrate, modernize, and run mainframe workloads on AWS.
- Cut mainframe migration time by 2/3
- Convert COBOL code into JAVA for you!
- （📝 Can friends who have the opportunity to play this service invite me to experience history?
2. Edge of the cloud
- On-premises & Cloud
- 📝 The fifth topic “Hybrid”, which was the finale of last year’s keynote speech, debuted about 40 minutes earlier this year.
- AWS Outposts published two years ago
- Run AWS infrastructure and services on premises
- Fully managed and supported by AWS
- Same hardware that AWS runs in its data centers
- Same APIs, controls plance, hardware, and tools
- Clients used include dish wireless, Verizon, Morningstar, Riot Games。
- 📝 It basically covers the three major industries that currently need AWS Outposts most: telecommunications, finance, and games.
- Pushing the edge of the cloud
- from near to far
- On premises at any facility
- AWS Outputs
- AWS IoT
- Rugged edge
- AWS Snow Family
- 5G networks
- AWS Wavelength
- 16 Wavelength Zones in the US, Europe, and Asia.
- AWS Wavelength
- [NEW 🚀] AWS Private 5G (Preview available today) <– 這應該就是今天的主要亮點產品了
- Set up and scale a private mobile network in days
- AWS provided hardware, software, and SIMs
- Automatic configuration
- No per-device charges
- Operates in shard spectrum
- （📝 AWS liberated the data center back then, and this year AWS liberated the 5G network!）
- （📝 5G as a Service; 5aaS; 五aaS）
Guest: Marc Rouanne, EVP & Chief Network Officer, DISH
- Marc Rouanne 9 joined DISH two years ago to create dish wireless.
- You may not be familiar with DISH, but you may have a little impression of sling TV.
- 1980 - Selling C-Band satellites
- 1996 - launching own satellite
- 2015 - sling streaming service
- Beyond a smartphone, this isn’t 4G plus.
- Traditional telecom operators are still using the 2G infrastructure of that year.
- Built architecture optimized for the cloud. Game changer for businesses.
- From the traditional Nokia 3310 phone（📝 I specially took a screenshot to pay tribute）
- Evolving to the iPhone Generation Emphasizing Customized experience
- Next, DISH wants to build Customized experience networks for enterprises, aka Customized Connectivity
- Data Centric - recognition pattern, Network data
- Automation at Scale - Hundreds of networks, automated （also used Amazon ECS）
- Connecting Edge to Cloud - actually super complicated - So the goal of DISH is Simplicity at the edge
- DISH wireless + AWS can build a nationwide ready-to-go network in a few days, compared to traditional telecom operators that may take years to build.
- Innovation will be accelerated.
3. The data journey
- Florence Nightingale 10
- “Lady with the Lamp” 11
- Nightingale collected data and presented it in a "polar area pie chart" format through "data visualization". These data and charts show that the disparity ratio of "death on the battlefield" and "death not on the battlefield" is 1:10. Force the authorities to take action, provide supplies to front hospitals, and improve the level of hygiene in the hospitals. Thus saving many lives. And established the UK's first nursing school after the war.
- 📝 I was so moved when I heard this part. Everyone, learn data analysis and data visualization!
- A data explosion
- Finding patterns in data
- A modern data strategy
- But working with data is tricky
- Your data in not static and it’s not used for just a single purpose
- Every organization has different data sources, different analytics needs, and different governance requirements.
- The data journey
- Data sources
- Data lakes
- People, apps, and devices
- Databases on AWS
- Amazon RDS
- Amazon Aurora
- Key-value - DynamoDB
- Document - DocumentDB
- Caching - ElastiCache
- Graph - Neptune
- Time-series - Timestream
- Ledger - QLDB
- Wide column - Keyspaces
- Memory - MemoryDB
- AWS Lake Formation
- Build a secure data lake in days
- After having a data lake, it is often necessary to set up different departments to access the corresponding data (management).
- [NEW 🚀] Row and cell-level security for Lake Formation (Generally Available Today)
- Restrict access to specific rows and columns
- Automatically filter and reveal data only to authorized users
- But what about data consistency?
- [NEW 🚀] Transactions for governed tables in Lake Formation (Generally Available Today)
- Automatically manages conflicts and errors
- Ensure consistent views for all users
- Eliminates need for custom error handling code or batching updates
- AWS Analytics services
- Interactive query - Amazon Athena
- Big data processing - Amazon EMR
- Operational analytics - Amazon Opensearch service
- Real-time analytics - Amazon Kinesis, Amazon MSK
- Data warehousing - Amazon Redshift
- Data integration - AWS Glue
- The best tool for the job.
- [NEW 🚀] Serverless and on-demand analytics (Preview available today)
- [NEW 🚀] Amazon Redshift Serverless (Preview available today)
- [NEW 🚀] Amazon EMR Serverless (Preview available today)
- [NEW 🚀] Amazon MSK Serverless (Preview available today)
- [NEW 🚀] Amazon Kinesis on-demand (Preview available today)
- Before going to talk about Machine Learning, invite customer UA to share.
Guest: Linda Jojo, EVP, Technology & Chief Digital Officer, United Airlines
- Connecting people. Uniting the world.
- 1.7 million flights, 162 million customer, in 2019.
- Inclusion propels innovation.
- Increase speed of innovation
- Variabilize our cost structure
- Replace aging legacy platforms
- 2019 內部在討論 single-cloud / multi-cloud 策略
- Quality and breadth of products, and continued pace of innovation
- 2020 四月有一天，fewer passengers than pilots
- UA found that their existing platform could not only scale up, but also scale down.
- Amazon has a two-pizza team, and UA has a one-screen team (WFH does not have more than one screen when connected, so there are nine people?!).
- UA Product: Travel Ready Center
- Amazon S3
- Amazon Trxtract
- Amazon DynamoDB
- Amazon SageMaker
- Automate validation 66%+ of all documents, 75%+ of COVID tests
- 100+ applications now run on AWS.
- +30 pts (Internal and external net promoter score (NPS)) since COVID
- Linda’s team in non-traditional by airline standards. 50% of members are women 。
- Broadest and most complete set of ML capabilities
4. Data was an afterthought
- Data integration
- Federated Query for Athena and Redshift
- Access SageMaker from Redshift, Aurora, and Neptune
- Govern your data
- Where it lives, who sees it, and who uses it
- Amazon QuickSight
- A scalable, embeddedable ML-powered BI service built for the cloud
- Amazon QuickSight Q
- Ask question in natural language
- Democratizing machine learning
- Machine learning for the inexperienced public and not developers
- [NEW 🚀] Amazon SageMaker Canvas (Generally Available Today) <– 這款 No-code 產品我覺得也滿有戰略地位的
- Generate hightly accurate ML predictions using a visual point-and-click interface - no coding required.
- Create ML predictions without any ML experience or writing any code
- Run an entire ML workflow with a simple point-and-click user interface
- Generate accurate predicitons
- Publish results and collaborate with others
- See what other don’t
I like this sentence the most today, so I took it out again:
Data was an afterthought.
- Year 1946 Roscoe C. Brown 12 used the old aircraft Tuskegee Airmen 44-C-SE to defeat the German ME-262 new jet fighter (state-of-the-art jet) at that time, and was the first to lay down The pilot of this jet. And later became a professor, opened education, etc., and was recognized by many institutions. His grandson is also here today.
- This paragraph mentions something about the importance of education and training.
- Train 29 million by 2025.
- 85% need more tech skills.
- AWS training programs
- AWS Skill Builder
- AWS training courses available on Amazon.com
- AWS re/Start program
- Helping more people use the cloud
- Amazon Connect
- Contact center in the cloud
- Transforming industries
- New paths for financial services
- Amazon FinSpace
- Reduce time to find, prepare, and analyze financial data from months to minutes
- [NEW 🚀] Goldman Sachs Financial Cloud for Data
- Clients access the same financial data and analytics GS has built and used for decades
- Discover, manage, and analyze financial data in minutes
- Integrated with AWS Data Exchange and Amazon FinSpace
- New paths for industrials
- Volkswagen brings 11 million cars a year to market, and working with AWS they’re unitnig 120 factory sites and data from 30,000 suppliers into a single architecture in the cloud.
Guest: Shaun Braun, SVP Digital Transformation, 3M Company
- Shaun Braun has studied neuroscience, MBA, worked in quality management, technology management, personnel industry, etc. 13
- 3M Company
- 60k products
- 95k employeess
- 125k patents
- The idea
- Take innovation
- Apply it in new ways
- Create breakthroughs
- To unlock the power of people, ideas, and science to improve lives everywhere.
- Moving away from past legacy environment
- Moved on time and on budget.
- 2k applications
- 9k VM
- Global SAP instance
- 45PB total storage
- 61% teir 1 apps moved in a year
- Digital building blocks on a cloud environment
- Agility and Scale
- 200+ plants
- 37 countries
- 51 technology platforms
- Amazon Neptune + Amazon API Gateway + AWS Lambda
- From raw materials to finished products can be traced. Material shortages can be known early on.
- Clean air
- Material + Digital Science
- AWS Lambda + Amazon CloudWatch
- Flagshiop product: 360 Encompass
- Amazon EC2 + S3 + CloudWatch + Cloud Trail
- Deployment times: from weeks to hours
- Digital Customer
- Digital Product
- Digital Operations
- Digital Enterprise
- Digital Marketplace
- Digital transformation journey
- Build foundation of accessible tech services
- Digital mindset across the organization
- Build digital products
5. New paths for industrials
- AWS for Industrial
- Engineering and design
- Production and asset optimization
- Supply chain management
- Worker safety and productivity
- Quality management
- Smart products and machines
- Purpose-built services for industrial
- Amazon Monitron
- Amazon Lookout for Vision
- AWS Panorama Appliance
- AWS Panorama SDK
- Amazon Lookout for Equipment
- Digital Twins
- Physical one
- Digital one
- Building digital twins takes heavy lifting
- [NEW 🚀] AWS IoT TwinMaker (Preview available today)
- A new service that makes it easy to create and use digital twins of real-world systems
- Pre-built connectors to data sources
- Automatically builds knowledge graphs
- 3D visualizations
- Dashboards to visualize operational state
- Updates in real time
- Siemens 已經開始使用。
- Driving change in automotive
- The connected car: generate 2 TB of data every hour
- [NEW 🚀] AWS IoT FleetWise (Preview available today)
- Collect data from millions of vehicles and make it easy to analyze in the cloud
- Access vehicle data for analysis in the cloud
- Intelligent filtering
- Perform diagnostics, analyze fleet health, improve safety.
- Transforming industries
This year’s physical version of Adam Selipsky Keynote is shortened to two hours compared to last year’s three-hour online version of Andy Jassy Keynote. Therefore, Adam cut off the time taken by many product releases, which does not mean that the innovation momentum of AWS has slowed down. On the contrary, AWS cares more about customers and the stories that happen with customers. So last year we saw Andy Jassy release many new containers (Container) and serverless (Serverless) products, but this year Adam Selipsky hardly mentioned the word container (Container), and new products related to Serverless are First explain the business application scenario, which is quite a reasonable choice for an MBA from Marketing, Sales & Support. In the limited speech time, use stories and customer cases to narrow the distance with the audience and expand the audience user group.
Another interesting comparison is that in Andy’s speech last year, the customer case was almost the breakpoint of the paragraph or the opening point of the next paragraph, while this year Adam’s speech cleverly arranged the customer sharing as a “turning point”, from the beginning to the end “change”. I believe this is the “turning point” structure he designed throughout the entire collection of stories. I find this very interesting.
As for the release of new products in Adam’s speech, will it be in the next three keynote speeches: “Swami Sivasubramanian Keynote” covers machine learning (Machine Learning; ML), “Peter DeSantis Keynote” covers infrastructure (Infrastructure; Infra ) and “Werner Vogels Keynote” covering technology, architecture, thinking and other directions, what are some presentations? Let’s keep finding :)
(The first draft of this article was completed 22 minutes before the “Swami Sivasubramanian Keynote” was published.)
What do you think of this AWS re:Invent 2021 Keynote? Welcome to AWS User Group Taiwan to chat with everyone!
If you feel that this thread dismantling note has caught the key points you want to read and is helpful to you, I hope to continue this kind of in-depth article in the future, welcome to pat and feed me a coffee ☕, or forward the article to your friends and colleagues, community, thank you for your encouragement and support.
✳️ If you are a leadership
- Find other sessions
- 👉 three consecutive years Think in Context: AWS re:Invent Keynote
- Developer conference
✳️ If you are a product manager
- Quickly grasp the whole picture
- 👉 All AWS product lines, a complete collection of one-sentence: AWS Cloud Product List
- Product Integration Framework
✳️ If you want to learn about AWS
- 👉 Join AWSUG Taiwan to connect with more AWS enthusiasts. Find someone to ask and learn twice as fast.
✳️ if you want more study notes
- Search when you think of it
- Productivity improvement
- 👉 【My PKM Workflow】