AWS Announces Amazon SageMaker Edge Manager

Recently AWS announced a new capability of Amazon SageMaker called Amazon SageMaker Edge Manager. This new capability in Amazon SageMaker makes it easy for customers to prepare, run, monitor, and update machine learning models on fleets of edge devices such as smart cameras, robots, and industrial machines.

Amazon SageMaker Edge Manager is one of the nine significant updates to the cloud-based machine learning platform Amazon SageMaker, announced during the annual re:Invent. With Edge Manager, the company delivers a solution for its customers to more effortless deploy and manage models at the edge.

Furthermore, Amazon SageMaker Edge Manager extends capabilities previously only available in the cloud by sampling models’ input and output data from edge devices and sending it to the cloud - allowing developers to continuously improve model quality by using Amazon SageMaker Model Monitor for drift detection, then relabel the data and retrain the models.

Developers can train or import a model in Amazon SageMaker, and subsequently let Amazon SageMaker Edge Manager optimize it for their hardware platform using Amazon SageMaker Neo – a service launched two years ago. This service converts models into an efficient standard format executed on the device by a low footprint runtime. Currently, Neo supports devices based on chips manufactured by Ambarella, ARM, Intel, NVIDIA, NXP, Qualcomm, TI, and Xilinx.

#edge #machine learning #aws #amazon sagemaker

What is GEEK

Buddha Community

AWS Announces Amazon SageMaker Edge Manager

AWS Announces Amazon SageMaker Edge Manager

Recently AWS announced a new capability of Amazon SageMaker called Amazon SageMaker Edge Manager. This new capability in Amazon SageMaker makes it easy for customers to prepare, run, monitor, and update machine learning models on fleets of edge devices such as smart cameras, robots, and industrial machines.

Amazon SageMaker Edge Manager is one of the nine significant updates to the cloud-based machine learning platform Amazon SageMaker, announced during the annual re:Invent. With Edge Manager, the company delivers a solution for its customers to more effortless deploy and manage models at the edge.

Furthermore, Amazon SageMaker Edge Manager extends capabilities previously only available in the cloud by sampling models’ input and output data from edge devices and sending it to the cloud - allowing developers to continuously improve model quality by using Amazon SageMaker Model Monitor for drift detection, then relabel the data and retrain the models.

Developers can train or import a model in Amazon SageMaker, and subsequently let Amazon SageMaker Edge Manager optimize it for their hardware platform using Amazon SageMaker Neo – a service launched two years ago. This service converts models into an efficient standard format executed on the device by a low footprint runtime. Currently, Neo supports devices based on chips manufactured by Ambarella, ARM, Intel, NVIDIA, NXP, Qualcomm, TI, and Xilinx.

#edge #machine learning #aws #amazon sagemaker

Jamison  Fisher

Jamison Fisher

1619416800

New – Profile Your Machine Learning Training Jobs With Amazon SageMaker Debugger

Today, I’m extremely happy to announce that Amazon SageMaker Debugger can now profile machine learning models, making it much easier to identify and fix training issues caused by hardware resource usage.

Despite its impressive performance on a wide range of business problems, machine learning (ML) remains a bit of a mysterious topic. Getting things right is an alchemy of science, craftsmanship (some would say wizardry), and sometimes luck. In particular, model training is a complex process whose outcome depends on the quality of your dataset, your algorithm, its parameters, and the infrastructure you’re training on.

As ML models become ever larger and more complex (I’m looking at you, deep learning), one growing issue is the amount of infrastructure required to train them. For instance, training BERT on the publicly available COCO dataset takes well over six hours on a single p3dn.24xlarge instance, even with its eight NVIDIA V100 GPUs. Some customers like autonomous vehicle companies deal with much larger datasets, and train object detection models for several days.

When a complex training job takes this long, the odds that something goes wrong and ruins it are pretty high, not only wasting time and money but also causing lots of frustration. Important work needs to be put on the back burner while you investigate, figure out the root cause, try to fix it, and then run your training job again. Often, you’ll have to iterate quite a few times to nail the problem.

#amazon machine learning #amazon sagemaker #announcements #artificial intelligence #aws re:invent #pytorch on aws #tensorflow on aws

Christa  Stehr

Christa Stehr

1598408880

How To Unite AWS KMS with Serverless Application Model (SAM)

The Basics

AWS KMS is a Key Management Service that let you create Cryptographic keys that you can use to encrypt and decrypt data and also other keys. You can read more about it here.

Important points about Keys

Please note that the customer master keys(CMK) generated can only be used to encrypt small amount of data like passwords, RSA key. You can use AWS KMS CMKs to generate, encrypt, and decrypt data keys. However, AWS KMS does not store, manage, or track your data keys, or perform cryptographic operations with data keys.

You must use and manage data keys outside of AWS KMS. KMS API uses AWS KMS CMK in the encryption operations and they cannot accept more than 4 KB (4096 bytes) of data. To encrypt application data, use the server-side encryption features of an AWS service, or a client-side encryption library, such as the AWS Encryption SDK or the Amazon S3 encryption client.

Scenario

We want to create signup and login forms for a website.

Passwords should be encrypted and stored in DynamoDB database.

What do we need?

  1. KMS key to encrypt and decrypt data
  2. DynamoDB table to store password.
  3. Lambda functions & APIs to process Login and Sign up forms.
  4. Sign up/ Login forms in HTML.

Lets Implement it as Serverless Application Model (SAM)!

Lets first create the Key that we will use to encrypt and decrypt password.

KmsKey:
    Type: AWS::KMS::Key
    Properties: 
      Description: CMK for encrypting and decrypting
      KeyPolicy:
        Version: '2012-10-17'
        Id: key-default-1
        Statement:
        - Sid: Enable IAM User Permissions
          Effect: Allow
          Principal:
            AWS: !Sub arn:aws:iam::${AWS::AccountId}:root
          Action: kms:*
          Resource: '*'
        - Sid: Allow administration of the key
          Effect: Allow
          Principal:
            AWS: !Sub arn:aws:iam::${AWS::AccountId}:user/${KeyAdmin}
          Action:
          - kms:Create*
          - kms:Describe*
          - kms:Enable*
          - kms:List*
          - kms:Put*
          - kms:Update*
          - kms:Revoke*
          - kms:Disable*
          - kms:Get*
          - kms:Delete*
          - kms:ScheduleKeyDeletion
          - kms:CancelKeyDeletion
          Resource: '*'
        - Sid: Allow use of the key
          Effect: Allow
          Principal:
            AWS: !Sub arn:aws:iam::${AWS::AccountId}:user/${KeyUser}
          Action:
          - kms:DescribeKey
          - kms:Encrypt
          - kms:Decrypt
          - kms:ReEncrypt*
          - kms:GenerateDataKey
          - kms:GenerateDataKeyWithoutPlaintext
          Resource: '*'

The important thing in above snippet is the KeyPolicy. KMS requires a Key Administrator and Key User. As a best practice your Key Administrator and Key User should be 2 separate user in your Organisation. We are allowing all permissions to the root users.

So if your key Administrator leaves the organisation, the root user will be able to delete this key. As you can see **KeyAdmin **can manage the key but not use it and KeyUser can only use the key. ${KeyAdmin} and **${KeyUser} **are parameters in the SAM template.

You would be asked to provide values for these parameters during SAM Deploy.

#aws #serverless #aws-sam #aws-key-management-service #aws-certification #aws-api-gateway #tutorial-for-beginners #aws-blogs

AWS Fargate for Amazon Elastic Kubernetes Service | Caylent

On-demand cloud computing brings new ways to ensure scalability and efficiency. Rather than pre-allocating and managing certain server resources or having to go through the usual process of setting up a cloud cluster, apps and microservices can now rely on on-demand serverless computing blocks designed to be efficient and highly optimized.

Amazon Elastic Kubernetes Service (EKS) already makes running Kubernetes on AWS very easy. Support for AWS Fargate, which introduces the on-demand serverless computing element to the environment, makes deploying Kubernetes pods even easier and more efficient. AWS Fargate offers a wide range of features that make managing clusters and pods intuitive.

Utilizing Fargate
As with many other AWS services, using Fargate to manage Kubernetes clusters is very easy to do. To integrate Fargate and run a cluster on top of it, you only need to add the command –fargate to the end of your eksctl command.

EKS automatically configures the cluster to run on Fargate. It creates a pod execution role so that pod creation and management can be automated in an on-demand environment. It also patches coredns so the cluster can run smoothly on Fargate.

A Fargate profile is automatically created by the command. You can choose to customize the profile later or configure namespaces yourself, but the default profile is suitable for a wide range of applications already, requiring no human input other than a namespace for the cluster.

There are some prerequisites to keep in mind though. For starters, Fargate requires eksctl version 0.20.0 or later. Fargate also comes with some limitations, starting with support for only a handful of regions. For example, Fargate doesn’t support stateful apps, DaemonSets or privileged containers at the moment. Check out this link for Fargate limitations for your consideration.

Support for conventional load balancing is also limited, which is why ALB Ingress Controller is recommended. At the time of this writing, Classic Load Balancers and Network Load Balancers are not supported yet.

However, you can still be very meticulous in how you manage your clusters, including using different clusters to separate trusted and untrusted workloads.

Everything else is straightforward. Once the cluster is created, you can begin specifying pod execution roles for Fargate. You have the ability to use IAM console to create a role and assign it to a Fargate cluster. Or you can also create IAM roles and Fargate profiles via Terraform.

#aws #blog #amazon eks #aws fargate #aws management console #aws services #kubernetes #kubernetes clusters #kubernetes deployment #kubernetes pods

Rory  West

Rory West

1621714500

Introducing AWS Systems Manager Change Manager

Because you are constantly listening to the feedback from your customer, you are iterating, innovating, and improving your applications and infrastructures. You continually modify your IT systems in the cloud. And let’s face it, changing something in a working system risks breaking things or introducing side effects that are sometimes unpredictable; it doesn’t matter how many tests you do. On the other hand, not making changes is stasis, followed by irrelevance, followed by death.

This is why organizations of all sizes and types have embraced a culture of controlling changes. Some organizations adopt change management processes such as the ones defined in ITIL v4. Some have adopted DevOps’ Continuous Deployment, or other methods. In any case, to support your change management processes, it is important to have tools.

Today, we are launching AWS Systems Manager Change Manager, a new change management capability for AWS Systems Manager. It simplifies the way ops engineers track, approve, and implement operational changes to their application configurations and infrastructures.

#announcements #aws re:invent #aws systems manager #aws