Onica’s Announcements At A Glance series analyzes the latest AWS news and announcements, simplifying and explaining the significance for AWS consumers.
March Madness doesn’t hold a candle to the madness surrounding the dizzying pace of innovation at AWS! March continued the trend of a mix of service announcements, service enhancements, and a brand-new specialty certification entering general availability. The purpose of this blog is to curate some of the announcements that we believe should garner the attention of enterprises looking to transform themselves via the cloud. Links to more information for each are included, so let’s take a look!
AWS made two different announcements in March concerning Deep Learning tools on AWS.
The first is the availability of AWS Deep Learning Containers (AWS DL Containers). These containers are pre-installed with Deep Learning frameworks to make the provisioning of these frameworks seamless and simple using containers. No more having to navigate the difficulties of containerized DL frameworks such as mismatched software dependencies and version compatibility issues. You can deploy AWS DL Containers on Amazon Elastic Container Service for Kubernetes (Amazon EKS), self-managed Kubernetes, Amazon Elastic Container Service (Amazon ECS), and Amazon EC2. The container images are available in AWS Elastic Container Registry (ECR) and are free to use. You can learn more and get started with Deep Learning AMIs here.
Second is that the AWS Deep Learning AMIs are now available for Amazon Linux 2, in addition to Amazon Linux and Ubuntu. Amazon Linux 2 is the next generation of Amazon Linux AMIs. MXNet 1.4.0, Chainer 5.3.0, PyTorch 1.0.1, and TensorFlow 1.13.1 are all available for these AMIs and are custom-built from source. They are also tuned for high-performance for training data on EC2. The availability of Apache MXNet 1.4 brings performance tuning and adds Java bindings for inference, Julia bindings, experimental control flow operators, JVM memory management, and many more under-the-hood enhancements. You can learn more and get started with Deep Learning AMIs here.
AWS Config Remediations
AWS Config has been a cloud-native, super easy way to keep your AWS architecture under constant watch. It helps to assess, audit, evaluate, and report on changes in your environment. Historically, AWS Config has been more, “Observe and Report” than “Recognize and Remediate.” Not anymore! March brings the announcement that AWS Config is now capable of not only monitoring your AWS infrastructure but providing remediation actions for elements that are out of compliance.
Setting up the remediation actions is pretty easy. You can choose in the AWS Config console to apply a preset list of remediations. Some of the preset remediation help with common tasks such as identifying an Amazon S3 bucket that is inadvertently set to public and changing it back to private via AWS Config remediation. For those that need more customization, you can also choose to apply your own custom remediation actions using AWS Systems Manager Automation documents. Standard metering rates apply to record configuration items and per active AWS Config rules. To get started, you can log in to AWS Config in the console, or dive into the documentation found here.
Need help getting your AWS Config rules set up? Set up a Well-Architected Review with us to ensure your architecture is running at its best!
AWS Certified Machine Learning – Specialty Certification
The AWS Certified Machine Learning – Specialty Certification is now widely available! If you’ve paid any attention to LinkedIn you’ve no doubt seen the flurry of posts about people who have taken the Beta exam finally beginning to get their results. As of March 15th, the AWS Certified Machine Learning – Specialty Certification is available to schedule in your certification account. Here is a list of knowledge items for the exam that AWS recommends on their website:
· One to two years of experience developing, architecting, or running ML/deep learning workloads on the AWS Cloud
· The ability to express the intuition behind basic ML algorithms
· Experience performing basic hyperparameter optimization
· Experience with ML and deep learning frameworks
· The ability to follow model-training best practices
· The ability to follow deployment and operational best practices
This exam follows the same concepts as other AWS Specialty Certification exams in terms of time, question format, and cost to sit. For those who pride themselves in staying “all in AWS certified,” it looks like there is some work to do. This one isn’t for the faint of heart either. Below are the domains that the exam covers.
|Domain 1: Data Engineering||20%|
|Domain 2: Exploratory Data Analysis||24%|
|Domain 3: Modeling||36%|
|Domain 4: Machine Learning Implementation and Operations||20%|
Multi-Account Support for Direct Connect Gateway
This one is a big deal, and has long been a feature request from enterprise customers! AWS Direct Connect (DX) is an excellent choice for cloud connectivity for enterprises who have a business need for low latency, low jitter, high-bandwidth connection to the cloud that bypasses the open internet. Working with an AWS DX partner you can request a connection in the console and work with the partner to complete the connection. Direct Connect Gateway added the ability to have a DX connection to multiple VPCs cross-region within the same account. This added the flexibility of being able to leverage a single DX connection across VPCs for the enterprise, as long as they were in the same account.
March brings the announcement of Multi-Account Support for Direct Connect Gateways. This will allow added flexibility for enterprises to facilitate attachment from a single DX connection to VPCs across accounts. You can associate up to 10 Amazon VPCs from multiple accounts with a Direct Connect gateway. For the configuration to work the Amazon VPCs and the Direct Connect gateway must be owned by AWS Accounts that belong to the same AWS payer account ID. While there is no cost to use Multi-Account Support of DX Gateways, Data Transfer Out is billed to the AWS account that owns the private virtual interface as always. To learn more check out the AWS User Guide found here.
Performance Insights for Amazon RDS for MariaDB
At some point, almost everyone needs to debug a performance issue related to a database. Traditionally, this dark art has been left squarely in the domain of seasoned DBAs, and often required expensive and complicated tools. Performance Insights for Amazon RDS and Amazon Aurora makes it easy to identify performance issues in the database by both experts and novices alike via a simple to understand dashboard. Performance insights is also available on Amazon Aurora, and Amazon RDS for PostgreSQL, MySQL, SQL Server, and Oracle. Learn more about Performance Insights here
To follow these updates and gain insights on how they can impact your business, subscribe to our blog!
Ready to improve your AWS presence? Get in touch!