As AWS Premier Consulting Partners, the engineering team here at Flux7 works day-in and day-out with a wide variety of AWS services. Whether we’re helping our enterprise customers speed time to market, grow security, or leverage high performance computing for competitive advantage, our goal is to use the right tool for the right job. Ironically, that was the theme of this year’s re:Invent keynotes, in which Andy Jassy and Werner Vogels (true to form) announced many new AWS products and services. In today’s blog, our engineers are weighing in on which announcements they’re most excited about and why.
- Instance Hibernation is now possible! This new feature allows you to launch EC2 instances, set them up as desired, and then hibernate them, which means you can wake them from their hibernated state when you need them. While in hibernation, the in-memory state of the instance, and the private and elastic IP addresses are stored, allowing it to pick up exactly where it left off. The main benefit here is the savings this will bring to AWS users, in terms of time spent booting OSs, apps, and other components -- and cost as you’ll no longer need to over-provision in case you require incremental capacity very quickly.
- For microservices, you can now visualize your setup and its health using AWS Cloud Map. This new service tracks your application components, their locations, attributes and health status for you. With a simply query to AWS Cloud Map, your application can discover the locations of its dependencies. While it’s challenging to track microservice dependencies, AWS Cloud Map benefits us by allowing applications to scale dynamically and connect to upstream services directly, without having to worry about app resources, even if they are built using containers that scale dynamically.
- Lambda Layers help create reusable components for serverless apps. Introduced by Werner Vogels on Thursday, Lambda Layers extend the Lambda execution environment with any binaries, dependencies or runtimes. According to Mr. Vogels, Lambda functions in a serverless application typically share common dependencies and now with layers, AWS operators can centrally manage common components across multiple functions resulting in better code reuse.
Continuing on the Lambda thread, our DevOps consulting team was also excited to see Application Load Balancer (ALB) support for Lambda, with users able to now invoke Lambda functions to serve HTTP requests. This is exciting because we can now access our serverless apps from any HTTP client, including web browsers whereas before we had to rely on proxy solutions to do so. This liberates us to place our apps behind the load balancer and we can now decide where things run without having to recode for our customers.
- WebSocket support for API Gateway! A widely-requested feature, it allows builders to create real-time, two-way communication applications. This is really exciting as it means we’ll no longer have to use work-arounds like using the IoT service to send messages over websockets. We look forward to trying it out once it’s released.
- Our engineers thought that Amazon’s focus on saving storage costs with two new offerings was noteworthy. Specifically, our customers will be able to save with S3 intelligent tiering which uses a Machine Learning model with a customer’s specific access patterns to their objects. Based on this learning, the new service will automatically move objects to a cooler or warmer tier based on what makes most sense.
In addition, AWS announced Glacier Deep Archive which got a resounding ‘woot’ from our engineering team. Anyone who manages tape today knows how much of a headache maintaining it can be. With Glacier Deep Archive, customers can store low access data more cost efficiently in the cloud ($1 US TB/month) and, with much greater accessibility.
- There were several database-related announcements this week, which was great to see as interest among our customers in this area (especially in data lakes) is growing. Specifically, our team’s interest was piqued around the ability of DynamoDB to now predict capacity so you don’t have to. Called Amazon DynamoDB On- Demand, it offers pay-per-request pricing for read and write requests so that you only pay for what you use. In this way we no longer have to provision for worst-case scenario-level capacity, thus making it much easier to balance cost and performance.
- For disaster recovery use cases, we’re looking forward to Amazon Aurora Global Database. The new database offering allows recovery from region-wide outages by spanning multiple AWS regions. And, provides fast replication to enable low-latency global reads.
- The new Amazon TimeStream is a good example of having the right tool for the right job. Custom-built specifically for collecting, storing, and processing time-series data Amazon Timestream can process the trillions of events that are generated every day by IoT applications, server and network logs and more with up to one thousand times faster query performance than a general-purpose database. Currently in preview mode, we look forward to trying this out for ourselves!
- Amazon Quantum Ledger Database, or QLDB, was introduced by Andy Jassy on Wednesday as a fully managed ledger database with a central trusted authority. To answer the age-old question, ‘what changed?’ we can now rely on QLDB to track each and every application data change and maintain a complete and verifiable history of changes over time. Obvious, but important, is that the change history in QLDB is immutable, and with cryptography built in, we can easily see if there have been any unintended modifications to an app’s data. Also in preview, we look forward to sharing our in-depth view of the new QLDB in the near future.
While there was plenty more news that our team was excited to share, please join us again next week as we wrap up our AWS re:Invent news analysis with coverage of security-related announcements, a bonanza of shared file system news, and more.
At Flux7, our business model is steeped in helping enterprises modernize their IT systems through our proprietary assessment, implementation services and knowledge transfer that help drive business success. As testament to that, this is the fifth year in a row that Flux7 or its customers have been featured at AWS re:invent. Watch our customer, Toyota Research Institute, present on how they are advancing the field of autonomous vehicle development using AWS and distributed deep learning.