AWS CEO Andy Jassy & CTO Werner Vogels Headline Premiere Cloud Industry Event
Last month AWS held the 7th annual re:Invent conference in Las Vegas with 50,000 attendees &100,000+ more tuning in on live stream. The conference featured over 2,100 break-out sessions, seminars, & presentations. Twelve & a half years after the launch of AWS, the company has millions of active hosting customers in startup companies, enterprise corporations, retail, small business, & the public sector. Major startups who build on AWS are AirBNB, Pinterest, & Stripe. In financial services, AWS powers Goldman Sachs, Capital One, & Barclays in cloud data center operations. AWS is the preferred service provider for Fortune 500 companies like J&J, GE, Shell, BP, Halliburton, Netflix, Disney, Fox & Expedia. AWS dominates every vertical in enterprise with a $27 billion revenue run rate and 46% YoY growth. AWS currently holds a 52% total share of the cloud market worldwide with main competitors Google, Microsoft, IBM, Oracle, & Ali Baba. AWS also inspires a broad ecosystem which includes companies like Infosys, Acquia, Adobe, Accenture, WorkDay, VMware, & Rackspace. Customers want access to widest range of tools for software development when building web/mobile applications, where AWS has the most options for programmers. Industry competitors are currently chasing & imitating AWS cloud hosting services w/out being able to offer the same platform capabilities.
AWS Platform Diversity: Compute, Database, Containers, ML, Serverless, ARM, & GPU
AWS CEO Andy Jassy led the re:Invent 2018 Keynote Address by highlighting the vast platform diversity that AWS provides for customers. For security, AWS has over 117 encryption products available for companies to choose from according to their cloud data center requirements. AWS has 11 major database options with additional migration services for companies converting legacy operations or software. AWS has unique platforms for computation, containers, serverless, machine learning, ARM servers, & GPU processing. The AWS managed Kubernetes service (EKS) has seen huge growth in the last year. The AWS Lambda platform has expanded to integrate with 47 different microservices.
In storage, AWS provides block storage & data transfer rates that lead the industry. AWS has the most volume types for storage on hosting plans with the ability to elastically change the size of each volume. There are currently 11 ways of data transfer from on-premises data center operations to AWS like DirectConnect, FireHose, SnowBall, etc. AWS DataSynch benchmarks at 10x faster than open source tools. Amazon has added secure FTP to S3 object storage which now supports over 10,000 customer data lakes with cross-region replication in availability zones.
Amazon S3 also offers batch operations on object storage via APIs and Glacier storage for archival rates. Jassy announced a new Machine Learning object storage option based on access patterns for intelligent tiering, as well as a new Glacier Deep Archive service that is cheaper than storage tape drives that are used on-premises in enterprise data centers - at less than 1/10th of 1 cent per GB per month. This service offers data recovery in hours instead of weeks.
The Amazon Elastic File System (EFS) supports data stored across three availability zones to scale up/down elastically according to web traffic demand. This service has seen major uptake in enterprise on Linux, so AWS announced a new Windows option as well. AWS now runs 57.7% of Windows IaaS instances compared to 30% for Microsoft Azure worldwide. Learn More.
AWS Inferentia: ML Inference Chip for TensorFlow, MXNet, & PyTorch
Another major announcement from AWS at re:Invent 2018 was the news that Amazon will introduce their own machine learning chip to compete with Google & IBM in TPU development. AWS has primarily offered GPU & ML cloud hosting services based on Nvidia hardware previously. According to the AWS press release:
"AWS Inferentia is a machine learning inference chip designed to deliver high performance at low cost. AWS Inferentia will support the TensorFlow, Apache MXNet, and PyTorch deep learning frameworks, as well as models that use the ONNX format... AWS Inferentia provides high throughput, low latency inference performance at an extremely low cost. Each chip provides hundreds of TOPS (tera operations per second) of inference throughput to allow complex models to make fast predictions."
Amazon is seeking to reduce costs for AI/ML apps at high demand by introducing the dedicated inference chip. According to internal research, a dedicated chip like the Inferentia can reduce costs as much as 90% over standard web servers or 75% vs. GPU-powered hardware. The AWS Inferentia platform will be inter-operable with Amazon EC2 and AWS SageMaker standards.
Amazon Aurora: The Foundation for Cloud-Native DB Innovation
AWS CTO Werner Vogels introduced his AWS re:Invent 2018 keynote by reviewing how Amazon.com's challenges in managing data center operation for ecommerce in Christmas season sales influenced the design of new databases & cell based architecture systems. Amazon Aurora represents scaled-out distributed architecture that is the fastest growing service in AWS history. AWS Aurora can support internet workloads no other database can support.
The Amazon Aurora Database Platform:
- Highly reliable and resilient to complex failure scenarios
- Very fast time to repair in extreme failure cases
- Decouple into a highly scalable and reliable storage service
- Make the storage service intelligent by being DB record aware
- Reduce write latency reinventing what gets writing
- Reduce read latency by avoiding quorum reads
DynamoDB spearheaded the non-relational database movement for web/mobile app developers. The four pillars of DynamoDB are:
- Performance at Scale
- Comprehensive Security
- Built for modern applications
DynamoDB has a cell-based architecture with grid structure to process shards. Adaptive burst capacity allows DynamoDB to grow with automatic re-sharding.
AWS Purpose-Built Database Platforms:
- Amazon Aurora
- Amazon RDS
- Amazon DynamoDB
- Amazon ElastiCache
- Amazon Neptune
- Amazon Timestream
- Amazon QLDB
Database migration services from Oracle are available for live applications and using a backfill process for gradual conversion. DynamoDB is compatible with Amazon S3 block storage for performance, machine learning, & cost benefits.
AWS re:Invent 2018 - Keynote with CEO Andy Jassy
AWS Firecracker: "At re:Invent 2018, Amazon made a major contribution to the open source community through Firecracker – a virtual machine manager that can run serverless workloads such as functions at lightning speed. It’s based on the same technology that powers AWS Lambda and AWS Fargate services... Firecracker has the potential to disrupt the container ecosystem. It combines the security of virtual machines with the speed of containers to deliver unique execution environment well-suited for serverless technology. Firecracker effectively blurs the line between VMs and containers by bringing best of both worlds." Learn More About AWS re:Invent 2018.
AWS re:Invent 2018 - Keynote with CTO Dr. Werner Vogels
AWS Lambda: "As Amazon CTO Werner Vogels pointed out in his keynote on Thursday, developers debate about tools and everyone has their own idea of what tools they bring to the task every day. For starters, they decided to please the language folks introducing support for new languages. Those developers who use Ruby can now use Ruby Support for AWS Lambda... If C++ is your thing, AWS announced C++ Lambda Runtime. If neither of those match your programming language tastes, AWS opened it up for just about any language with the new Lambda Runtime API... For folks who have different tastes in IDEs (integrated development environments), AWS announced Lambda support for several popular ones including PyCharm and IntelliJ in preview and Visual Studio." Learn More About AWS re:Invent 2018.
Data Analytics & Java Apps: "Amazon Kinesis makes it easy to collect, process, and analyze real-time, streaming data so you can get timely insights and react quickly to new information. Amazon Kinesis Data Analytics... promises to help you process data in real-time using standard SQL. AWS is introducing support for Java in Amazon Kinesis Data Analytics. What does this mean exactly? In short, you can now use your own Java code to create powerful real-time applications that process streaming data." Learn More About AWS re:Invent 2018.
New Product Launches: Landing Zones, Control Tower, Elastic Inference, & Lake Formation
In current business management, companies have multiple teams with different microservice requirements, so AWS is introducing Landing Zones with multi-account management tools. AWS Control Tower is a new product for federated access with AWS Single Sign-On for identity management. AWS Lake Formation is a new service that speeds up the processing of data for machine learning. Amazon QLDB is a fully managed ledger database with a central trusted authority. Amazon Elastic Inference adds GPU acceleration to any Amazon EC2 instance for faster inference at up to 75% lower costs. This multi-front approach to platform innovation highlighted the launch of new products at AWS re:Invent 2018 which leads the industry in scale.