More and more companies are basing a large part or all of their applications on containerization. For example, Google, Spotify and Netflix have announced that their platforms have evolved in recent years to a container-based model.
Given the large volume of containers they handle, it is crucial to have an orchestrator that can control the management and operational tasks on a day-to-day basis. Some of these companies, such as Google, have been pioneers and have invented and evolved solutions such as Kubernetes.
But not everything ends with Kubernetes. Netflix chose to create its own development solution, Titus.[i] Titus is based on Amazon EC2 instances and integrates in a cloud-native path to Amazon Web Services (AWS) to run hundreds of thousands of containers that deliver productive workloads to the world leader in streaming services. AWS is evolving its container support and adoption plans by expanding the scope of one of its core solutions in the current portfolio, Amazon Elastic Container Service (ECS). Launched in 2015, ECS has a similar architecture as Kubernetes but is more closely aligned with AWS building blocks and terminology.
Scale Servies Globally, Leverage Equinix Fabric™ for Secure, Private Interconnection
Connect to the world’s largest ecosystems of interconnected partners and providers to accelerate your organization's digital transformation. Equinix international business exchange (IBX®) data centers leverage Equinix Fabric™ to gain control over IT cost and performance issues.Read More
Amazon ECS has a similar architecture as Kubernetes but is more closely aligned with AWS building blocks and terminology.
At the last Reinvent 2020, Amazon introduced ECS Anywhere, a feature that could revolutionize the use of containers in on-premises infrastructures. Amazon ECS Anywhere is game-changing and confirms that AWS is attending the demands of the community. Amazon ECS Anywhere simplifies managing containers on EC2 instances at any scale.
Up until now, Amazon ECS brought users a fully-managed service that takes care of the control plane (layer with the cluster control components) and let’s you choose two options for the data plane layer that provide the follwoing resources:
- Amazon EC2 instances: provides full control over the instances and access to the underlying host.
- AWS Fargate: Allows developers to abstract from instances and only define the necessary resources for the execution of tasks. AWS is already in charge of managing the computation layer. AWS Fargate is the serverless flavour to run containers on Amazon ECS.
But now, Amazon ECS Anywhere opens up a new way to execute workloads. Developers can now deploy cloud native workloads to an on-premises infrastructure. This means that our physical servers, virtual machines or even a laptop or a Raspberry Pi computer will be able to run containers, taking the operation and management from the control plane running on Amazon ECS. And very importantly, these resources are integrated within the ecosystem of AWS services with products such as: Cloudwatch, IAM, Amazon SQS, etc.
Amazon ECS Anywhere use cases
Here are a couple of use cases wehre Amazon ECS Anywhere provides added value to container options that are alredy in place:
- Taking advantage of capacity that companies currently have in their data center that is not being usedThere are resources in the cloud as well as other resources on-premies that can be used on other platforms. Developers can use that capacity in a “cloud-native” way with Amazon ECS Anywhere.
- Managing infrastructureIn some countries or industries, companies need to manage their own infrastructure. AWS Outposts[ii] in this case is not an option because this infrastructure is managed by AWS. In this case, using Amazon ECS Anywhere allows you to take advantage of the simplicity of ECS (i.e., forget the operations in control plane) in the same time that you are managing your resources (data plane) in your own data center, whle you are complying with the regulations of your country/sector.
Amazon ECS Anywhere requirements
The only requirement that Amazon ECS Anywhere has is that two agents must be installed:
- The first one is the SSM agent that allows you to register the resource to communicate with the control plane that is hosted in the cloud in a secure way. In this way, the servers will become manageable instances from AWS.
- The second is the ECS agent itself that will register you within the cluster you are going to operate. After that registration, when you launch a new task from ECS, you will have the option to select External, which means that the execution of that container will be carried out in your local infrastructure.
The importance of optimal communication
The communication between the control plane running in the cloud and your on-premises infrastructure becomes even more important now that you are going to extend the use of containers. Not only because of the permanent communication between the Amazon ECS Anywhere core and the agents, but also because the integration of workloads that are going to be run on-premises with AWS services is one of the great benefits of operating with Amazon ECS Anywhere.
Equinix Fabric is the perfect choice to guarantee performance stability while leveraging private and secure connectivity to quickly deploy hybrid multicloud infrastructures. It connects private IT infrastructure to AWS within minutes over high-speed, low-latency virtualized connections. Although it does not appear as an explicit requirement, Equini Fabric private interconnection is one way to offer the same user experience by deploying on-prem containers that interact with others in the cloud or even ECS services.
In addition, Equinix offers two more solutions that add value to the new Amazon ECS Anywhere service:
- Equinix Metal: If you need to manage computing in a data center where you currently have or will have a presence, Equinix Metal offers the best solution to get physical servers up and running in minutes as an on-demand service.
- Network Edge: you can use virtual network services from leading providers via Network Edge servcies from Equinix to route and secure your traffic between AWS and your on-premise servers. You can use Equinix Edge when enabling new connections to the cloud that need to be agile and on-demand, without deploying additional hardware network applications. Network Edge saves time and money when deploying virtual devices such as SS-WAN gateways and cloud routers.
The following diagram shows an application created by Amazon where the value of extending infrastructure on Platform Equinix and AWS is evident. Without going into detail about the flow of the application or workload, we can see how there is some data stored in on-premises repositories that are moved from one source to a destination due to a call from an ECS task executed in the Equinix International Business Exchange™ (IBX®) data center.
The task executed from ECS External interacts with the other tasks and the AWS services (Amazon SQS, Amazon Cloudwatch) through a private connection using Equinix Fabric, creating a virtual circuit with AWS Direct Connect and securing private access using AWS PrivateLink. In addition, security control over the task and services is added at all times using AWS IAM. Network Edge servcies can also interact with the connection depending on the type of interconnection and your business needs.
Equinix Fabric is the perfect choice to guarantee performance stability while leveraging private and secure connectivity to quickly deploy hybrid multicloud infrastructures.
In short, we are seeing how big players are betting on evolving their services toward hybrid multicloud architectures and Amazon ECS Anywhere is a clear example.
Learn more by reading the Equinix Fabric data sheet.
Or you can contact Equinix if you need to learn how we can help you optimize Amazon ECS Anywhere and other services.
The task executed from ECS External interacts with the other tasks and the AWS services through a private connection using Equinix Fabric.