Amazon EC2 FAQs

Q: What is Amazon Elastic Compute Cloud (Amazon EC2)?

Amazon Elastic Compute Cloud (Amazon EC2) is a web service that provides resizable compute capacity in the cloud. It is designed to make web-scale computing easier for developers.

Q: What can I do with Amazon EC2?

Just as Amazon Simple Storage Service (Amazon S3) enables storage in the cloud, Amazon EC2 enables “compute” in the cloud. Amazon EC2’s simple web service interface allows you to obtain and configure capacity with minimal friction. It provides you with complete control of your computing resources and lets you run on Amazon’s proven computing environment. Amazon EC2 reduces the time required to obtain and boot new server instances to minutes, allowing you to quickly scale capacity, both up and down, as your computing requirements change. Amazon EC2 changes the economics of computing by allowing you to pay only for capacity that you actually use.

Q: How can I get started with Amazon EC2?

To sign up for Amazon EC2, click the “Sign up for This Web Service” button on the Amazon EC2 detail page. You must have an Amazon Web Services account to access this service; if you do not already have one, you will be prompted to create one when you begin the Amazon EC2 sign-up process. After signing up, please refer to the Amazon EC2 documentation, which includes our Getting Started Guide.

Q: Why am I asked to verify my phone number when signing up for Amazon EC2?

Amazon EC2 registration requires you to have a valid phone number and email address on file with AWS in case we ever need to contact you. Verifying your phone number takes only a couple of minutes and involves receiving a phone call during the registration process and entering a PIN number using the phone key pad.

Q: What can developers now do that they could not before?

Until now, small developers did not have the capital to acquire massive compute resources and ensure they had the capacity they needed to handle unexpected spikes in load. Amazon EC2 enables any developer to leverage Amazon’s own benefits of massive scale with no up-front investment or performance compromises. Developers are now free to innovate knowing that no matter how successful their businesses become, it will be inexpensive and simple to ensure they have the compute capacity they need to meet their business requirements.

The “Elastic” nature of the service allows developers to instantly scale to meet spikes in traffic or demand. When computing requirements unexpectedly change (up or down), Amazon EC2 can instantly respond, meaning that developers have the ability to control how many resources are in use at any given point in time. In contrast, traditional hosting services generally provide a fixed number of resources for a fixed amount of time, meaning that users have a limited ability to easily respond when their usage is rapidly changing, unpredictable, or is known to experience large peaks at various intervals.

Q: How do I run systems in the Amazon EC2 environment?

Once you have set up your account and select or create your AMIs, you are ready to boot your instance. You can start your AMI on any number of On-Demand instances by using the RunInstances API call. You simply need to indicate how many instances you wish to launch. If you wish to run more than 20 On-Demand instances, complete the Amazon EC2 instance request form.

If Amazon EC2 is able to fulfill your request, RunInstances will return success, and we will start launching your instances. You can check on the status of your instances using the DescribeInstances API call. You can also programmatically terminate any number of your instances using the TerminateInstances API call.

If you have a running instance using an Amazon EBS boot partition, you can also use the StopInstances API call to release the compute resources but preserve the data on the boot partition. You can use the StartInstances API when you are ready to restart the associated instance with the Amazon EBS boot partition.

In addition, you have the option to use Spot Instances to reduce your computing costs when you have flexibility in when your applications can run. Read more about Spot Instances for a more detailed explanation on how Spot Instances work.

If you prefer, you can also perform all these actions from the AWS Management Console or through the command line using our command line tools, which have been implemented with this web service API.

Q: What is the difference between using the local instance store and Amazon Elastic Block Store (Amazon EBS) for the root device?

When you launch your Amazon EC2 instances you have the ability to store your root device data on Amazon EBS or the local instance store. By using Amazon EBS, data on the root device will persist independently from the lifetime of the instance. This enables you to stop and restart the instance at a subsequent time, which is similar to shutting down your laptop and restarting it when you need it again.

Alternatively, the local instance store only persists during the life of the instance. This is an inexpensive way to launch instances where data is not stored to the root device. For example, some customers use this option to run large web sites where each instance is a clone to handle web traffic.

Q: How quickly will systems be running?

It typically takes less than 10 minutes from the issue of the RunInstances call to the point where all requested instances begin their boot sequences. This time depends on a number of factors including: the size of your AMI, the number of instances you are launching, and how recently you have launched that AMI. Images launched for the first time may take slightly longer to boot.

Q: How do I load and store my systems with Amazon EC2?

Amazon EC2 allows you to set up and configure everything about your instances from your operating system up to your applications. An Amazon Machine Image (AMI) is simply a packaged-up environment that includes all the necessary bits to set up and boot your instance. Your AMIs are your unit of deployment. You might have just one AMI or you might compose your system out of several building block AMIs (e.g., webservers, appservers, and databases). Amazon EC2 provides a number of tools to make creating an AMI easy. Once you create a custom AMI, you will need to bundle it. If you are bundling an image with a root device backed by Amazon EBS, you can simply use the bundle command in the AWS Management Console. If you are bundling an image with a boot partition on the instance store, then you will need to use the AMI Tools to upload it to Amazon S3. Amazon EC2 uses Amazon EBS and Amazon S3 to provide reliable, scalable storage of your AMIs so that we can boot them when you ask us to do so.

Or, if you want, you don’t have to set up your own AMI from scratch. You can choose from a number of globally available AMIs that provide useful instances. For example, if you just want a simple Linux server, you can choose one of the standard Linux distribution AMIs.

Q: How do I access my systems?

The RunInstances call that initiates execution of your application stack will return a set of DNS names, one for each system that is being booted. This name can be used to access the system exactly as you would if it were in your own data center. You own that machine while your operating system stack is executing on it.

Q: Is Amazon EC2 used in conjunction with Amazon S3?

Yes, Amazon EC2 is used jointly with Amazon S3 for instances with root devices backed by local instance storage. By using Amazon S3, developers have access to the same highly scalable, reliable, fast, inexpensive data storage infrastructure that Amazon uses to run its own global network of web sites. In order to execute systems in the Amazon EC2 environment, developers use the tools provided to load their AMIs into Amazon S3 and to move them between Amazon S3 and Amazon EC2. See How do I load and store my systems with Amazon EC2? for more information about AMIs.

We expect developers to find the combination of Amazon EC2 and Amazon S3 to be very useful. Amazon EC2 provides cheap, scalable compute in the cloud while Amazon S3 allows users to store their data reliably.

Q: How many instances can I run in Amazon EC2?

You are limited to running On-Demand Instances per your vCPU-based On-Demand Instance limit, purchasing 20 Reserved Instances, and requesting Spot Instances per your dynamic Spot limit per region. New AWS accounts may start with limits that are lower than the limits described here.

If you need more instances, complete the Amazon EC2 limit increase request form with your use case, and your limit increase will be considered. Limit increases are tied to the region they were requested for.

Q: Are there any limitations in sending email from Amazon EC2 instances?

Yes. In order to maintain the quality of Amazon EC2 addresses for sending email, we enforce default limits on the amount of email that can be sent from EC2 accounts. If you wish to send larger amounts of email from EC2, you can apply to have these limits removed from your account by filling out this form.

Q: How quickly can I scale my capacity both up and down?

Amazon EC2 provides a truly elastic computing environment. Amazon EC2 enables you to increase or decrease capacity within minutes, not hours or days. You can commission one, hundreds or even thousands of server instances simultaneously. When you need more instances, you simply call RunInstances, and Amazon EC2 will typically set up your new instances in a matter of minutes. Of course, because this is all controlled with web service APIs, your application can automatically scale itself up and down depending on its needs.

Q: What operating system environments are supported?

Amazon EC2 currently supports a variety of operating systems including: Amazon Linux, Ubuntu, Windows Server, Red Hat Enterprise Linux, SUSE Linux Enterprise Server, openSUSE Leap, Fedora, Fedora CoreOS, Debian, CentOS, Gentoo Linux, Oracle Linux, and FreeBSD. We are looking for ways to expand it to other platforms.

Q: Does Amazon EC2 use ECC memory?

In our experience, ECC memory is necessary for server infrastructure, and all the hardware underlying Amazon EC2 uses ECC memory.

Q: How is this service different than a plain hosting service?

Traditional hosting services generally provide a pre-configured resource for a fixed amount of time and at a predetermined cost. Amazon EC2 differs fundamentally in the flexibility, control and significant cost savings it offers developers, allowing them to treat Amazon EC2 as their own personal data center with the benefit of Amazon.com’s robust infrastructure.

When computing requirements unexpectedly change (up or down), Amazon EC2 can instantly respond, meaning that developers have the ability to control how many resources are in use at any given point in time. In contrast, traditional hosting services generally provide a fixed number of resources for a fixed amount of time, meaning that users have a limited ability to easily respond when their usage is rapidly changing, unpredictable, or is known to experience large peaks at various intervals.

Secondly, many hosting services don’t provide full control over the compute resources being provided. Using Amazon EC2, developers can choose not only to initiate or shut down instances at any time, they can completely customize the configuration of their instances to suit their needs – and change it at any time. Most hosting services cater more towards groups of users with similar system requirements, and so offer limited ability to change these.

Finally, with Amazon EC2 developers enjoy the benefit of paying only for their actual resource consumption – and at very low rates. Most hosting services require users to pay a fixed, up-front fee irrespective of their actual computing power used, and so users risk overbuying resources to compensate for the inability to quickly scale up resources within a short time frame. 

Q: What is changing?

Amazon EC2 is transitioning On-Demand Instance limits from the current instance count-based limits to the new vCPU-based limits to simplify the limit management experience for AWS customers. Usage toward the vCPU-based limit is measured in terms of number of vCPUs (virtual central processing units) for the Amazon EC2 Instance Types to launch any combination of instance types that meet your application needs.

Q: What are vCPU-based limits?

You are limited to running one or more On-Demand Instances in an AWS account, and Amazon EC2 measures usage towards each limit based on the total number of vCPUs (virtual central processing unit) that are assigned to the running On-Demand instances in your AWS account. The following table shows the number of vCPUs for each instance size. The vCPU mapping for some instance types may differ; see Amazon EC2 Instance Types for details.

Q: How many On-Demand instances can I run in Amazon EC2?

There are five vCPU-based instance limits, each defines the amount of capacity you can use of a given instance family. All usage of instances in a given family, regardless of generation, size, or configuration variant (e.g. disk, processor type), will accrue towards the family’s total vCPU limit, listed in the table below. New AWS accounts may start with limits that are lower than the limits described here.

Q: Are these On-Demand Instance vCPU-based limits regional?

Yes, the On-Demand Instance limits for an AWS account are set on a per-region basis.

Q: Will these limits change over time?

Yes, limits can change over time. Amazon EC2 is constantly monitoring your usage within each region and your limits are raised automatically based on your use of EC2.

Q: How can I request a limit increase?

Even though EC2 automatically increases your On-Demand Instance limits based on your usage, if needed you can request a limit increases from the Limits Page on Amazon EC2 console, the Amazon EC2 service page on the Service Quotas console, or the Service Quotas API/CLI.

Q: How can I calculate my new vCPU limit?

You can find the vCPU mapping for each of the Amazon EC2 Instance Types or use the simplified vCPU Calculator to compute the total vCPU limit requirements for your AWS account.

Q: Do vCPU limits apply when purchasing Reserved Instances or requesting Spot Instances?

No, the vCPU-based limits only apply to running On-Demand instances.

Q: How can I view my current On-Demand Instance limits?

You can find your current On-Demand Instance limits on the EC2 Service Limits page in the Amazon EC2 console, or from the Service Quotas console and APIs.

Q: Will this affect running instances?

No, opting into vCPU-based limits will not affect any running instances.

Q: Can I still launch the same number of instances?

Yes, the vCPU-based instance limits allow you to launch at least the same number of instances as count-based instance limits.

Q: Will I be able to view instance usage against these limits?

With the Amazon CloudWatch metrics integration, you can view EC2 usage against limits in the Service Quotas console. Service Quotas also enables customers to use CloudWatch for configuring alarms to warn customers of approaching limits. In addition, you can continue to track and inspect your instance usage in Trusted Advisor and Limit Monitor.

Q: Will I still be able to use the DescribeAccountAttributes API?

With the vCPU limits, we no longer have total instance limits governing the usage. Hence the DescribeAccountAttributes API will no longer return the max-instances value. Instead you can now use the Service Quotas APIs to retrieve information about EC2 limits. You can find more information about the Service Quotas APIs in the AWS documentation.

Q: Will the vCPU limits have an impact on my monthly bill?

No. EC2 usage is still calculated either by the hour or the second, depending on which AMI you're running and the instance type and size you’ve launched.

Q: Will vCPU limits be available in all Regions?

vCPU-based instance limits are available in all commercial AWS Regions except the AWS China (Beijing and Ningxia) Regions and they are available in AWS GovCloud (US).

Q. What is changing?

Starting Jan-27 2020, Amazon Elastic Compute Cloud (EC2) will begin rolling out a change to restrict email traffic over port 25 by default to protect customers and other recipients from spam and email abuse. Port 25 is typically used as the default SMTP port to send emails. AWS accounts that have requested and had Port 25 throttles removed in the past will not be impacted by this change.

Q. I have a valid use-case for sending emails to port 25 from EC2. How can I have these port 25 restrictions removed?

If you have a valid use-case for sending emails to port 25 (SMTP) from EC2, please submit a Request to Remove Email Sending Limitations to have these restrictions lifted. You can alternately send emails using a different port, or leverage an existing authenticated email relay service such as Amazon Simple Email Service (SES).

Q. What does your Amazon EC2 Service Level Agreement guarantee?

Our SLA guarantees a Monthly Uptime Percentage of at least 99.99% for Amazon EC2 and Amazon EBS within a Region.

Q. How do I know if I qualify for a SLA Service Credit?

You are eligible for a SLA credit for either Amazon EC2 or Amazon EBS (whichever was Unavailable, or both if both were Unavailable) if the Region that you are operating in has an Monthly Uptime Percentage of less than 99.99% during any monthly billing cycle. For full details on all of the terms and conditions of the SLA, as well as details on how to submit a claim, please see http://aws.amazon.com/ec2/sla/

Q: What are Accelerated Computing instances?

Accelerated Computing instance family is a family of instances which use hardware accelerators, or co-processors, to perform some functions, such as floating-point number calculation and graphics processing, more efficiently than is possible in software running on CPUs. Amazon EC2 provides three types of Accelerated Computing instances – GPU compute instances for general-purpose computing, GPU graphics instances for graphics intensive applications, and FPGA programmable hardware compute instances for advanced scientific workloads.

Q. When should I use GPU Graphics and Compute instances?

GPU instances work best for applications with massive parallelism such as workloads using thousands of threads. Graphics processing is an example with huge computational requirements, where each of the tasks is relatively small, the set of operations performed form a pipeline, and the throughput of this pipeline is more important than the latency of the individual operations. To be able build applications that exploit this level of parallelism, one needs GPU device specific knowledge by understanding how to program against various graphics APIs (DirectX, OpenGL) or GPU compute programming models (CUDA, OpenCL).

Q: How are P3 instances different from G3 instances?

P3 instances are the next-generation of EC2 general-purpose GPU computing instances, powered by up to 8 of the latest-generation NVIDIA Tesla V100 GPUs. These new instances significantly improve performance and scalability, and add many new features, including new Streaming Multiprocessor (SM) architecture for machine learning (ML)/deep learning (DL) performance optimization, second-generation NVIDIA NVLink high-speed GPU interconnect, and highly tuned HBM2 memory for higher-efficiency.

G3 instances use NVIDIA Tesla M60 GPUs and provide a high-performance platform for graphics applications using DirectX or OpenGL. NVIDIA Tesla M60 GPUs support NVIDIA GRID Virtual Workstation features, and H.265 (HEVC) hardware encoding. Each M60 GPU in G3 instances supports 4 monitors with resolutions up to 4096x2160, and is licensed to use NVIDIA GRID Virtual Workstation for one Concurrent Connected User. Example applications of G3 instances include 3D visualizations, graphics-intensive remote workstation, 3D rendering, application streaming, video encoding, and other server-side graphics workloads.

Q: What are the benefits of NVIDIA Volta GV100 GPUs?

The new NVIDIA Tesla V100 accelerator incorporates the powerful new Volta GV100 GPU. GV100 not only builds upon the advances of its predecessor, the Pascal GP100 GPU, it significantly improves performance and scalability, and adds many new features that improve programmability. These advances will supercharge HPC, data center, supercomputer, and deep learning systems and applications.

Q: Who will benefit from P3 instances?

P3 instances with their high computational performance will benefit users in artificial intelligence (AI), machine learning (ML), deep learning (DL) and high performance computing (HPC) applications. Users includes data scientists, data architects, data analysts, scientific researchers, ML engineers, IT managers and software developers. Key industries include transportation, energy/oil & gas, financial services (banking, insurance), healthcare, pharmaceutical, sciences, IT, retail, manufacturing, high-tech, transportation, government, academia, among many others.

Q: What are some key use cases of P3 instances?

P3 instance use GPUs to accelerate numerous deep learning systems and applications including autonomous vehicle platforms, speech, image, and text recognition systems, intelligent video analytics, molecular simulations, drug discovery, disease diagnosis, weather forecasting, big data analytics, financial modeling, robotics, factory automation, real-time language translation, online search optimizations, and personalized user recommendations, to name just a few.

Q: Why should customers use GPU-powered Amazon P3 instances for AI/ML and HPC?

GPU-based compute instances provide greater throughput and performance because they are designed for massively parallel processing using thousands of specialized cores per GPU, versus CPUs offering sequential processing with a few cores. In addition, developers have built hundreds of GPU-optimized scientific HPC applications such as quantum chemistry, molecular dynamics, meteorology, among many others. Research indicates that over 70% of the most popular HPC applications provide built-in support for GPUs.

Q: Will P3 instances support EC2 Classic networking and Amazon VPC?

P3 instances will support VPC only.

Q: How are G3 instances different from P2 instances?

G3 instances use NVIDIA Tesla M60 GPUs and provide a high-performance platform for graphics applications using DirectX or OpenGL. NVIDIA Tesla M60 GPUs support NVIDIA GRID Virtual Workstation features, and H.265 (HEVC) hardware encoding. Each M60 GPU in G3 instances supports 4 monitors with resolutions up to 4096x2160, and is licensed to use NVIDIA GRID Virtual Workstation for one Concurrent Connected User. Example applications of G3 instances include 3D visualizations, graphics-intensive remote workstation, 3D rendering, application streaming, video encoding, and other server-side graphics workloads.

P2 instances use NVIDIA Tesla K80 GPUs and are designed for general purpose GPU computing using the CUDA or OpenCL programming models. P2 instances provide customers with high bandwidth 25 Gbps networking, powerful single and double precision floating-point capabilities, and error-correcting code (ECC) memory, making them ideal for deep learning, high performance databases, computational fluid dynamics, computational finance, seismic analysis, molecular modeling, genomics, rendering, and other server-side GPU compute workloads.

Q: How are P3 instances different from P2 instances?

P3 Instances are the next-generation of EC2 general-purpose GPU computing instances, powered by up to 8 of the latest-generation NVIDIA Volta GV100 GPUs. These new instances significantly improve performance and scalability and add many new features, including new Streaming Multiprocessor (SM) architecture, optimized for machine learning (ML)/deep learning (DL) performance, second-generation NVIDIA NVLink high-speed GPU interconnect, and highly tuned HBM2 memory for higher-efficiency.

P2 instances use NVIDIA Tesla K80 GPUs and are designed for general purpose GPU computing using the CUDA or OpenCL programming models. P2 instances provide customers with high bandwidth 25 Gbps networking, powerful single and double precision floating-point capabilities, and error-correcting code (ECC) memory.

Q: What APIs and programming models are supported by GPU Graphics and Compute instances?

P3 instances support CUDA 9 and OpenCL, P2 instances support CUDA 8 and OpenCL 1.2 and G3 instances support DirectX 12, OpenGL 4.5, CUDA 8, and OpenCL 1.2.

Q: Where do I get NVIDIA drivers for P3 and G3 instances?

There are two methods by which NVIDIA drivers may be obtained. There are listings on the AWS Marketplace which offer Amazon Linux AMIs and Windows Server AMIs with the NVIDIA drivers pre-installed. You may also launch 64-bit, HVM AMIs and install the drivers yourself. You must visit the NVIDIA driver website and search for the NVIDIA Tesla V100 for P3, NVIDIA Tesla K80 for P2, and NVIDIA Tesla M60 for G3 instances.

Q: Which AMIs can I use with P3, P2 and G3 instances?

You can currently use Windows Server, SUSE Enterprise Linux, Ubuntu, and Amazon Linux AMIs on P2 and G3 instances. P3 instances only support HVM AMIs. If you want to launch AMIs with operating systems not listed here, contact AWS Customer Support with your request or reach out through EC2 Forums.

Q: Does the use of G2 and G3 instances require third-party licenses?

Aside from the NVIDIA drivers and GRID SDK, the use of G2 and G3 instances does not necessarily require any third-party licenses. However, you are responsible for determining whether your content or technology used on G2 and G3 instances requires any additional licensing. For example, if you are streaming content you may need licenses for some or all of that content. If you are using third-party technology such as operating systems, audio and/or video encoders, and decoders from Microsoft, Thomson, Fraunhofer IIS, Sisvel S.p.A., MPEG-LA, and Coding Technologies, please consult these providers to determine if a license is required. For example, if you leverage the on-board h.264 video encoder on the NVIDIA GRID GPU you should reach out to MPEG-LA for guidance, and if you use mp3 technology you should contact Thomson for guidance.

Q: Why am I not getting NVIDIA GRID features on G3 instances using the driver downloaded from NVIDIA website?

The NVIDIA Tesla M60 GPU used in G3 instances requires a special NVIDIA GRID driver to enable all advanced graphics features, and 4 monitors support with resolution up to 4096x2160. You need to use an AMI with NVIDIA GRID driver pre-installed, or download and install the NVIDIA GRID driver following the AWS documentation.

Q: Why am I unable to see the GPU when using Microsoft Remote Desktop?

When using Remote Desktop, GPUs using the WDDM driver model are replaced with a non-accelerated Remote Desktop display driver. In order to access your GPU hardware, you need to utilize a different remote access tool, such as VNC.

Q: What is Amazon EC2 F1?

Amazon EC2 F1 is a compute instance with programmable hardware you can use for application acceleration. The new F1 instance type provides a high performance, easy to access FPGA for developing and deploying custom hardware accelerations.

Q: What are FPGAs and why do I need them?

FPGAs are programmable integrated circuits that you can configure using software. By using FPGAs you can accelerate your applications up to 30x when compared with servers that use CPUs alone. And, FPGAs are reprogrammable, so you get the flexibility to update and optimize your hardware acceleration without having to redesign the hardware.

Q: How does F1 compare with traditional FPGA solutions?

F1 is an AWS instance with programmable hardware for application acceleration. With F1, you have access to FPGA hardware in a few simple clicks, reducing the time and cost of full-cycle FPGA development and scale deployment from months or years to days. While FPGA technology has been available for decades, adoption of application acceleration has struggled to be successful in both the development of accelerators and the business model of selling custom hardware for traditional enterprises, due to time and cost in development infrastructure, hardware design, and at-scale deployment. With this offering, customers avoid the undifferentiated heavy lifting associated with developing FPGAs in on-premises data centers.

Q: What is an Amazon FPGA Image (AFI)?

The design that you create to program your FPGA is called an Amazon FPGA Image (AFI). AWS provides a service to register, manage, copy, query, and delete AFIs. After an AFI is created, it can be loaded on a running F1 instance. You can load multiple AFIs to the same F1 instance, and can switch between AFIs in runtime without reboot. This lets you quickly test and run multiple hardware accelerations in rapid sequence. You can also offer to other customers on the AWS Marketplace a combination of your FPGA acceleration and an AMI with custom software or AFI drivers.

Q: How do I list my hardware acceleration on the AWS Marketplace?

You would develop your AFI and the software drivers/tools to use this AFI. You would then package these software tools/drivers into an Amazon Machine Image (AMI) in an encrypted format. AWS manages all AFIs in the encrypted format you provide to maintain the security of your code. To sell a product in the AWS Marketplace, you or your company must sign up to be an AWS Marketplace reseller, you would then submit your AMI ID and the AFI ID(s) intended to be packaged in a single product. AWS Marketplace will take care of cloning the AMI and AFI(s) to create a product, and associate a product code to these artifacts, such that any end-user subscribing to this product code would have access to this AMI and the AFI(s).

Q: What is available with F1 instances?

For developers, AWS is providing a Hardware Development Kit (HDK) to help accelerate development cycles, a FPGA Developer AMI for development in the cloud, an SDK for AMIs running the F1 instance, and a set of APIs to register, manage, copy, query, and delete AFIs. Both developers and customers have access to the AWS Marketplace where AFIs can be listed and purchased for use in application accelerations.

Q: Do I need to be a FPGA expert to use an F1 instance?

AWS customers subscribing to a F1-optimized AMI from AWS Marketplace do not need to know anything about FPGAs to take advantage of the accelerations provided by the F1 instance and the AWS Marketplace. Simply subscribe to an F1-optimized AMI from the AWS Marketplace with an acceleration that matches the workload. The AMI contains all the software necessary for using the FPGA acceleration. Customers need only write software to the specific API for that accelerator and start using the accelerator.

Q: I’m a FPGA developer, how do I get started with F1 instances?

Developers can get started on the F1 instance by creating an AWS account and downloading the AWS Hardware Development Kit (HDK). The HDK includes documentation on F1, internal FPGA interfaces, and compiler scripts for generating AFI. Developers can start writing their FPGA code to the documented interfaces included in the HDK to create their acceleration function. Developers can launch AWS instances with the FPGA Developer AMI. This AMI includes the development tools needed to compile and simulate the FPGA code. The Developer AMI is best run on the latest C5, M5, or R4 instances. Developers should have experience in the programming languages used for creating FPGA code (i.e. Verilog or VHDL) and an understanding of the operation they wish to accelerate.

Q: I’m not an FPGA developer, how do I get started with F1 instances?

Customers can get started with F1 instances by selecting an accelerator from the AWS Marketplace, provided by AWS Marketplace sellers, and launching an F1 instance with that AMI. The AMI includes all of the software and APIs for that accelerator. AWS manages programming the FPGA with the AFI for that accelerator. Customers do not need any FPGA experience or knowledge to use these accelerators. They can work completely at the software API level for that accelerator.

Q: Does AWS provide a developer kit?

Yes. The Hardware Development Kit (HDK) includes simulation tools and simulation models for developers to simulate, debug, build, and register their acceleration code. The HDK includes code samples, compile scripts, debug interfaces, and many other tools you will need to develop the FPGA code for your F1 instances. You can use the HDK either in an AWS provided AMI, or in your on-premises development environment. These models and scripts are available publically with an AWS account.

Q: Can I use the HDK in my on-premises development environment?

Yes. You can use the Hardware Development Kit HDK either in an AWS-provided AMI, or in your on-premises development environment.

Q: Can I add an FPGA to any EC2 instance type?

No. F1 instances comes in two instance sizes f1.2xlarge, f1.4xlarge, and f1.16 xlarge.

Q: How do I use the Inferentia chip in Inf1 instances?

You can start your workflow by building and training your model in one of the popular ML frameworks such as TensorFlow, PyTorch, or MXNet using GPU instances such as P3 or P3dn. Once the model is trained to your required accuracy, you can use the ML framework’s API to invoke Neuron, a software development kit for Inferentia, to compile the model for execution on Inferentia chips, load it in to Inferentia’s memory, and then execute inference calls. In order to get started quickly, you can use AWS Deep Learning AMIs that come pre-installed with ML frameworks and the Neuron SDK. For a fully managed experience you will be able to use Amazon SageMaker which will enable you to seamlessly deploy your trained models on Inf1 instances.

Q: When would I use Inf1 vs. C5 vs. G4 instances for inference?

Customers running machine learning models that are sensitive to inference latency and throughput can use Inf1 instances for high-performance cost-effective inference. For those ML models that are less sensitive to inference latency and throughput, customers can use EC2 C5 instances and utilize the AVX-512/VNNI instruction set. For ML models that require access to NVIDIA’s CUDA, CuDNN or TensorRT libraries, we recommend using G4 instances.

Q: When should I choose Elastic Inference (EI) for inference vs Amazon EC2 Inf1 instances?

There are two cases where developers would choose EI over Inf1 instances: (1) if you need different CPU and memory sizes than what Inf1 offers, then you can use EI to attach acceleration to the EC2 instance with the right mix of CPU and memory for your application (2) if your performance requirements are significantly lower than what the smallest Inf1 instance provides, then using EI could be a more cost effective choice. For example, if you only need 5 TOPS, enough for processing up to 6 concurrent video streams, then using the smallest slice of EI with a C5.large instance could be up to 50% cheaper than using the smallest size of an Inf1 instance.

Q: What ML models types and operators are supported by EC2 Inf1 instances using the Inferentia chip?

Inferentia chips support the commonly used machine learning models such as single shot detector (SSD) and ResNet for image recognition/classification and Transformer and BERT for natural language processing and translation and many others. A list of supported operators can be found on GitHub.

Q: How do I take advantage of AWS Inferentia’s NeuronCore Pipeline capability to lower latency?

Inf1 instances with multiple Inferentia chips, such as Inf1.6xlarge or Inf1.24xlarge, support a fast chip-to-chip interconnect. Using the Neuron Processing Pipeline capability, you can split your model and load it to local cache memory across multiple chips. The Neuron compiler uses ahead-of-time (AOT) compilation technique to analyze the input model and compile it to fit across the on-chip memory of single or multiple Inferentia chips. Doing so enables the Neuron Cores to have high-speed access to models and not require access to off-chip memory, keeping latency bounded while increasing the overall inference throughput.

Q: What is the difference between AWS Neuron and Amazon SageMaker Neo?

AWS Neuron is a specialized SDK for AWS Inferentia chips that optimizes the machine learning inference performance of Inferentia chips. It consists of a compiler, run-time, and profiling tools for AWS Inferentia and is required to run inference workloads on EC2 Inf1 instances. On the other hand, Amazon SageMaker Neo is a hardware agnostic service that consists of a compiler and run-time that enables developers to train machine learning models once, and run them on many different hardware platforms.  

Q. When should I use Compute Optimized instances?

Compute Optimized instances are designed for applications that benefit from high compute power. These applications include compute-intensive applications like high-performance web servers, high-performance computing (HPC), scientific modelling, distributed analytics and machine learning inference.

Q: What are Amazon EC2 C6g instances?

Amazon EC2 C6g instances are the next-generation of compute-optimized instances powered by Arm-based AWS Graviton2 Processors. C6g instances deliver up to 40% better price performance over C5 instances. They are built on the AWS Nitro System, a combination of dedicated hardware and Nitro hypervisor.

Q: What are some of the ideal use cases for C6g instances?

C6g instances deliver significant price performance benefits for compute-intensive workloads such as high performance computing (HPC), batch processing, ad serving, video encoding, gaming, scientific modelling, distributed analytics, and CPU-based machine learning inference. Customers deploying applications built on open source software across the C instance family will find the C6g instances an appealing option to realize the best price performance within the instance family. Arm developers can also build their applications directly on native Arm hardware as opposed to cross-compilation or emulation.

Q: What are the various storage options available on C6g instances?

C6g instances are EBS-optimized by default and offer up to 19,000 Mbps of dedicated EBS bandwidth to both encrypted and unencrypted EBS volumes. C6g instances only support Non-Volatile Memory Express (NVMe) interface to access EBS storage volumes. Additionally, options with local NVMe instance storage are also available through the C6gd instance types.

Q: Which network interface is supported on C6g instances?

C6g instances support ENA based Enhanced Networking. With ENA, C6g instances can deliver up to 25 Gbps of network bandwidth between instances when launched within a Placement Group.

Q: Will customers need to modify their applications and workloads to be able to run on the C6g instances?

The changes required are dependent on the application. Customers running applications built on open source software will find that the Arm ecosystem is well developed and already likely supports their applications. Most Linux distributions as well as containers (Docker, Kubernetes, Amazon ECS, Amazon EKS, Amazon ECR) support the Arm architecture. Customers will find Arm versions of commonly used software packages available for installation through the same mechanisms that they currently use. Applications that are based on interpreted languages (such as Java, Node, Python) not reliant on native CPU instruction sets should run with minimal to no changes. Applications developed using compiled languages (C, C++, GoLang) will need to be re-compiled to generate Arm binaries. The Arm architecture is well supported in these popular programming languages and modern code usually requires a simple ‘Make’ command. Refer to the Getting Started guide on Github for more details.

Q: Will there be more compute choices offered with the C6 instance families?

Yes, we plan to offer Intel and AMD CPU powered instances in the future as part of the C6 instance families.

Q. Can I launch C4 instances as Amazon EBS-optimized instances?

Each C4 instance type is EBS-optimized by default. C4 instances 500 Mbps to 4,000 Mbps to EBS above and beyond the general-purpose network throughput provided to the instance. Since this feature is always enabled on C4 instances, launching a C4 instance explicitly as EBS-optimized will not affect the instance's behavior.

Q. How can I use the processor state control feature available on the c4.8xlarge instance?

The c4.8xlarge instance type provides the ability for an operating system to control processor C-states and P-states. This feature is currently available only on Linux instances. You may want to change C-state or P-state settings to increase processor performance consistency, reduce latency, or tune your instance for a specific workload. By default, Amazon Linux provides the highest-performance configuration that is optimal for most customer workloads; however, if your application would benefit from lower latency at the cost of higher single- or dual-core frequencies, or from lower-frequency sustained performance as opposed to bursty Turbo Boost frequencies, then you should consider experimenting with the C-state or P-state configuration options that are available to these instances. For additional information on this feature, see the Amazon EC2 User Guide section on Processor State Control.

Q. Which instances are available within Compute Optimized instances category?

C6g instances: Amazon EC2 C6g instances are powered by Arm-based AWS Graviton2 processors. They deliver up to 40% better price performance over current generation C5 instances and are ideal for running advanced compute-intensive workloads. This includes workloads such as high performance computing (HPC), batch processing, ad serving, video encoding, gaming, scientific modelling, distributed analytics, and CPU-based machine learning inference. 

C5 instances: C5 instances are the latest generation of EC2 Compute Optimized instances. C5 instances are based on Intel Xeon Platinum processors, part of the Intel Xeon Scalable (codenamed Skylake-SP or Cascade Lake) processor family, are available in 9 sizes, and offer up to 96 vCPUs and 192 GiB memory. C5 instances deliver 25% improvement in price/performance compared to C4 instances. The C5d instances have local NVMe storage for workloads that require very low latency and storage access with high random read and write IOPS ability.

C5a instances: C5a instances deliver leading x86 price-performance for a broad set of compute-intensive workloads including batch processing, distributed analytics, data transformations, log analysis, and web applications. C5a instances feature 2nd Gen 3.3GHz AMD EPYC processors with up to 96 vCPUs and up to 192 GiB of memory. The C5ad instances have local NVMe storage for workloads that require very low latency and storage access with high random read and write IOPS ability 

C5n instances: C5n instances are ideal for applications requiring high network bandwidth and packet rate. The C5n instances are ideal for applications like HPC, data lakes, network appliances as well as applications that require inter-node communication and the Message Passing Interface (MPI). C5n offer a choice of Intel Xeon Platinum 3.0 GHz processors with up to 72 vCPUs and 192GiB of Memory.

C4 instances: C4 instances are based on Intel Xeon E5-2666 v3 (codenamed Haswell) processors. C4 instances are available in 5 sizes and offer up to 36 vCPUs and 60 GiB memory.

Q. Should I move my workloads from C3 or C4 instances to C5 instances?

The generational improvement in CPU performance and lower price of C5 instances, which combined result in a 25% price/performance improvement relative to C4 instances, benefit a broad spectrum of workloads that currently run on C3 or C4 instances. For floating point intensive applications, Intel AVX-512 enables significant improvements in delivered TFLOPS by effectively extracting data level parallelism. Customers looking for absolute performance for graphics rendering and HPC workloads that can be accelerated with GPUs or FPGAs should also evaluate other instance families in the Amazon EC2 portfolio that include those resources to find the ideal instance for their workload.

Q. Which operating systems/AMIs are supported on C5 Instances?

EBS backed HVM AMIs with support for ENA networking and booting from NVMe-based storage can be used with C5 instances. The following AMIs are supported on C5:

• Amazon Linux 2014.03 or newer
  
• Ubuntu 14.04 or newer
• SUSE Linux Enterprise Server 12 or newer
• Red Hat Enterprise Linux 7.4 or newer
• CentOS 7 or newer
• Windows Server 2008 R2
• Windows Server 2012
• Windows Server 2012 R2
• Windows Server 2016
• FreeBSD 11.1-RELEASE
  

For optimal local NVMe-based SSD storage performance on C5d, Linux kernel version 4.9+ is recommended.

Q. What are the storage options available to C5 customers?

C5 instances use EBS volumes for storage, are EBS-optimized by default, and offer up to 9 Gbps throughput to both encrypted and unencrypted EBS volumes. C5 instances access EBS volumes via PCI attached NVM Express (NVMe) interfaces. NVMe is an efficient and scalable storage interface commonly used for flash based SSDs such as local NVMe storage provided with I3 and I3en instances. Though the NVMe interface may provide lower latency compared to Xen paravirtualized block devices, when used to access EBS volumes the volume type, size, and provisioned IOPS (if applicable) will determine the overall latency and throughput characteristics of the volume. When NVMe is used to provide EBS volumes, they are attached and detached by PCI hotplug.

Q. What network interface is supported on C5 instances?

C5 instances use the Elastic Network Adapter (ENA) for networking and enable Enhanced Networking by default. With ENA, C5 instances can utilize up to 25 Gbps of network bandwidth.

Q. Which storage interface is supported on C5 instances?

C5 instances will support only NVMe EBS device model. EBS volumes attached to C5 instances will appear as NVMe devices. NVMe is a modern storage interface that provides latency reduction and results in increased disk I/O and throughput.

Q. How many EBS volumes can be attached to C5 instances?

C5 instances support a maximum for 27 EBS volumes for all Operating systems. The limit is shared with ENI attachments which can be found here http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using-eni.html. For example: since every instance has at least 1 ENI, if you have 3 additional ENI attachments on the c4.2xlarge, you can attach 24 EBS volumes to that instance.

Q. What is the underlying hypervisor on C5 instances?

C5 instances use a new EC2 hypervisor that is based on core KVM technology.

Q: Why does the total memory reported by the operating system not match the advertised memory of the C5 instance type?

In C5, portions of the total memory for an instance are reserved from use by the Operating System including areas used by the virtual BIOS for things like ACPI tables and for devices like the virtual video RAM.

Q: What are Amazon EC2 T4g instances?

Amazon EC2 T4g instances are the next-generation of general purpose burstable instances powered by Arm-based AWS Graviton2 processors. T4g instances deliver up to 40% better price performance over T3 instances. They are built on the AWS Nitro System, a combination of dedicated hardware and Nitro hypervisor.

Q: What are some of the ideal use cases for T4g instances?

T4g instances deliver up to 40% better price performance over T3 instances for a wide variety of burstable general purpose workloads such as micro-services, low-latency interactive applications, small and medium databases, virtual desktops, development environments, code repositories, and business-critical applications. Customers deploying applications built on open source software across the T instance family will find the T4g instances an appealing option to realize the best price performance within the instance family. Arm developers can also build their applications directly on native Arm hardware as opposed to cross-compilation or emulation.

Q. Who is eligible for T4g free-trial?

All existing and new customers with an AWS account can take advantage of the T4g free-trial. T4g free-trial is available for limited time, until Dec 31, 2020. The start and end time of free-trial are as per Coordinated Universal Time (UTC). T4g free-trial will be available in addition to the existing AWS Free-Tier on t2.micro/t3.micro. Customers who have exhausted their t2.micro (or t3.micro, depending on the region) free-tier usage can still benefit from T4g free-trial.

Q. How can customers get access to the T4g free-trial?

Until Dec 31, 2020, all AWS customers will be automatically enrolled in the T4g free trial.
During the free-trial period, customers who run a t4g.micro instance will automatically get 750 free hours per month deducted from their bill during each month. The 750 hours will be calculated in aggregate across all 7 launch regions.

Q. What is the regional availability of T4g free-trial?

T4g Free-trial is currently available across US East (N. Virginia, Ohio), US West (Oregon), Asia Pacific (Tokyo, Mumbai), Europe (Frankfurt, Ireland). Customers will be able to run free t4g.micro across one or multiple regions of choice from a single cumulative bucket of 750 free hours per month till Dec 31st 2020. For example, a customer can run t4g.micro in Oregon for 300 hours for a month and run another t4g.micro in Tokyo for 450 hours that same month. This would add up to 750 hours/month of free-trial limit.

Q. Is there an additional charge for running specific AMIs under the T4g free-trial?

Under t4g.micro free-trial, there will be no AMI charge for Amazon Linux 2, RHEL, and SUSE Linux AMIs that are available through EC2 Console Quickstart for the first 750 free hours/month. After 750 free hours /month, regular OD prices including AMI charge (if any) will be applicable. The applicable software fees for AWS Marketplace offers with AMI fulfillment options is not included in the free trial, only the t4g.micro infrastructure cost is included and covered under the free-trial.

Q. How will the t4g.micro free trial be reflected on my AWS bill?

T4g free-trial has a monthly billing cycle that starts in the 1st of every month and ends on the last day of that month. Under the T4g free-trial billing plan, customers using t4g.micro will see a $0 line item on their bill under on-demand pricing plan for the first 750 aggregate hours of usage for every month during the free-trial period. Customers can start anytime during the free-trial period and get 750 free hours for the remainder of that month. Any unused hours from the previous month will not be carried over. Customers can launch multiple t4g.micro instances under the free-trial. Customers will be notified automatically via email using AWS Budgets when their aggregate monthly usage reaches 85% of 750 free hours. When the aggregate instance usage exceeds 750 hours for the monthly billing cycle, customers will be charged based on regular on-demand pricing for the exceeded hours that month. For customers with a Compute Savings Plan or T4g Instance Savings Plan, Savings Plan (SV) discount will be applied to on-demand pricing for hours beyond the 750 free trial hours. One consideration is if customers have purchased T4g Reserved Instance (RI) plan, RI plan applies first to any usage on an hourly basis. For any remaining usage after RI plan has been applied, the free-trial billing plan takes effect.

Q. If customers sign-up for consolidated billing (i.e. a single payer account), can they get the T4g Free Trial for each account that is tied to the payer account?

No, customers who use consolidated billing to consolidate payment across multiple accounts will have access to one free trial per Organization. Each payer account gets a total aggregate of 750 free hours a month. More details on consolidated billing can be found here.

Q. Will customers get charged for surplus CPU credits as a part of T4g free-trial?

Customers will not have to pay for surplus CPU credits when they exceed the instances allocated credits during the 750 free hours of T4g free-trial program. At the end of free-trial or for any usage beyond the 750 free hours per month, all regular billing charges including surplus credits charges will apply. For details on how CPU credits work, please refer to documentation.

Q. At the end of free trial, how will a customer be billed for t4g.micro instance?

Starting Jan 1, 2021, customers running on t4g.micro will be automatically switched from the free-trial plan to on demand pricing plan (or RI/SV plan if purchased) and accumulated credits will be set to zero. Customers will receive email notification 7 days prior to end of free-trial period automatically that the free-trial period will be ending in 7 days. Starting Jan 1st 2021, if RI plan is purchased, the RI plans will apply. Otherwise, customers will be charged regular on-demand pricing for t4g.micro. For customer that have the T4g Instance Savings Plan or a Compute Savings Plan, t4g.micro instance billing will apply Savings Plan discount on their OD pricing.

Q: What are Amazon EC2 M6g instances?

Amazon EC2 M6g instances are the next-generation of general-purpose instances powered by Arm-based AWS Graviton2 Processors. M6g instances deliver up to 40% better price/performance over M5 instances. They are built on the AWS Nitro System, a combination of dedicated hardware and Nitro hypervisor.

Q: What are the specifications of the new AWS Graviton2 Processors?

The AWS Graviton2 processors deliver up to 7x performance, 4x the number of compute cores, 2x larger caches, 5x faster memory, and 50% faster per core encryption performance than first generation AWS Graviton processors. Each core of the AWS Graviton2 processor is a single-threaded vCPU. These processors also offer always-on fully encrypted DRAM memory, hardware acceleration for compression workloads, dedicated engines per vCPU that double the floating-point performance for workloads such as video encoding, and instructions for int8/fp16 CPU-based machine learning inference acceleration. The CPUs are built utilizing 64-bit Arm Neoverse cores and custom silicon designed by AWS on the advanced 7 nm manufacturing technology.

Q: Is memory encryption supported by AWS Graviton2 processors?

AWS Graviton2 processors support always-on 256-bit memory encryption to further enhance security. Encryption keys are securely generated within the host system, do not leave the host system, and are irrecoverably destroyed when the host is rebooted or powered down. Memory encryption does not support integration with AWS KMS system and customers cannot bring their own keys.

Q: What are some of the ideal use cases for M6g instances?

M6g instances deliver significant performance and price performance benefits for a broad spectrum of general-purpose workloads such as application servers, gaming servers, microservices, mid-size databases, and caching fleets. Customers deploying applications built on open source software across the M instance family will find the M6g instances an appealing option to realize the best price-performance within the instance family. Arm developers can also build their applications directly on native Arm hardware as opposed to cross-compilation or emulation.

Q: What are the various storage options available on M6g instances?

M6g instances are EBS-optimized by default and offer up to 19,000 Mbps of dedicated EBS bandwidth to both encrypted and unencrypted EBS volumes. M6g instances only support Non-Volatile Memory Express (NVMe) interface to access EBS storage volumes. Additionally, options with local NVMe instance storage are also available through the M6gd instance types.

Q: Which network interface is supported on M6g instances?

M6g instances support ENA based Enhanced Networking. With ENA, M6g instances can deliver up to 25 Gbps of network bandwidth between instances when launched within a Placement Group.

Q: Will customers need to modify their applications and workloads to be able to run on the M6g instances?

The changes required are dependent on the application. Customers running applications built on open source software will find that the Arm ecosystem is well developed and already likely supports their applications. Most Linux distributions as well as containers (Docker, Kubernetes, Amazon ECS, Amazon EKS, Amazon ECR) support the Arm architecture. Customers will find Arm versions of commonly used software packages available for installation through the same mechanisms that they currently use. Applications that are based on interpreted languages (such as Java, Node, Python) not reliant on native CPU instruction sets should run with minimal to no changes. Applications developed using compiled languages (C, C++, GoLang) will need to be re-compiled to generate Arm binaries. The Arm architecture is well supported in these popular programming languages and modern code usually requires a simple ‘Make’ command. Refer to the Getting Started guide on Github for more details.

Q: Will there be more compute choices offered with the M6 instance families?

Yes, we plan to offer Intel and AMD CPU powered instances in the future as part of the M6 instance families.

Q: What are Amazon EC2 A1 instances?

Amazon EC2 A1 instances are general purpose instances powered by the first-generation AWS Graviton Processors that are custom designed by AWS.

Q: What are the specifications of the first-generation AWS Graviton Processors? 

AWS Graviton processors are custom designed by AWS utilizing Amazon’s extensive expertise in building platform solutions for cloud applications running at scale. These processors are based on the 64-bit Arm instruction set and feature Arm Neoverse cores as well as custom silicon designed by AWS. The cores operate at a frequency of 2.3 GHz.

Q: When should I use A1 instances?

A1 instances deliver significant cost savings for scale-out workloads that can fit within the available memory footprint. A1 instances are ideal for scale-out applications such as web servers, containerized microservices, and data/log processing. These instances will also appeal to developers, enthusiasts, and educators across the Arm developer community.

Q: Will customers have to modify applications and workloads to be able to run on the A1 instances?

The changes required are dependent on the application. Applications based on interpreted or run-time compiled languages (e.g. Python, Java, PHP, Node.js) should run without modifications. Other applications may need to be recompiled and those that don't rely on x86 instructions will generally build with minimal to no changes.

Q: Which operating systems/AMIs are supported on A1 Instances?

The following AMIs are supported on A1 instances: Amazon Linux 2, Ubuntu 16.04.4 or newer, Red Hat Enterprise Linux (RHEL) 7.6 or newer, SUSE Linux Enterprise Server 15 or newer. Additional AMI support for Fedora, Debian, NGINX Plus are also available through community AMIs and the AWS Marketplace. . EBS backed HVM AMIs launched on A1 instances require NVMe and ENA drivers installed at instance launch.

Q: Are there specific AMI requirements to run on M6g and A1 instances?

You will need to use the “arm64” AMIs with the M6g and A1 instances. x86 AMIs are not compatible with M6g and A1 instances.

Q: When should customers use A1 instances versus the new M6g instances?

A1 instances continue to offer significant cost benefits for scale-out workloads that can run on multiple smaller cores and fit within the available memory footprint. The new M6g instances are a good fit for a broad spectrum of applications that require more compute, memory, networking resources and/or can benefit from scaling up across platform capabilities. M6g instances will deliver the best price-performance within the instance family for these applications. M6g supports up to 16xlarge instance size (A1 supports up to 4xlarge), 4GB of memory per vCPU (A1 supports 2GB memory per vCPU), and up to 25 Gbps of networking bandwidth (A1 supports up to 10 Gbps).

Q: What are the various storage options available to A1 customers?

A1 instances are EBS-optimized by default and offer up to 3,500 Mbps of dedicated EBS bandwidth to both encrypted and unencrypted EBS volumes. A1 instances only support Non-Volatile Memory Express (NVMe) interface to access EBS storage volumes. A1 instances will not support the blkfront interface.

Q: Which network interface is supported on A1 instances?

A1 instances support ENA based Enhanced Networking. With ENA, A1 instances can deliver up to 10 Gbps of network bandwidth between instances when launched within a Placement Group.

Q: Do A1 instances support the AWS Nitro System?

Yes, A1 instances are powered by the AWS Nitro System, a combination of dedicated hardware and Nitro hypervisor.

Q: Why does the total memory reported by Linux not match the advertised memory of the A1 instance type?

In A1 instances, portions of the total memory for an instance are reserved from use by the operating system including areas used by the virtual UEFI for things like ACPI tables.

Q: What are the key use cases for Amazon EC2 M5 Instances?

M5 instances offer a good choice for running development and test environments, web, mobile and gaming applications, analytics applications, and business critical applications including ERP, HR, CRM, and collaboration apps. Customers who are interested in running their data intensive workloads (e.g. HPC, or SOLR clusters) on instances with a higher memory footprint will also find M5 to be a good fit. Workloads that heavily use single and double precision floating point operations and vector processing such as video processing workloads and need higher memory can benefit substantially from the AVX-512 instructions that M5 supports.

Q: Why should customers choose EC2 M5 Instances over EC2 M4 Instances?

Compared with EC2 M4 Instances, the new EC2 M5 Instances deliver customers greater compute and storage performance, larger instance sizes for less cost, consistency and security. The biggest benefit of EC2 M5 Instances is based on its usage of the latest generation of Intel Xeon Scalable processors (Skylake-SP or Cascade Lake), which deliver up to 20% improvement in price/performance compared to M4. With AVX-512 support in M5 vs. the older AVX2 in M4, customers will gain 2x higher performance in workloads requiring floating point operations. M5 instances offer up to 25 Gbps of network bandwidth and up to 10 Gbps of dedicated bandwidth to Amazon EBS. M5 instances also feature significantly higher networking and Amazon EBS performance on smaller instance sizes with EBS burst capability.

Q: How does support for Intel AVX-512 benefit EC2 M5 and M5d Instance customers?

Intel Advanced Vector Extension 512 (AVX-512) is a set of new CPU instructions available on the latest Intel Xeon Scalable processor family, that can accelerate performance for workloads and usages such as scientific simulations, financial analytics, artificial intelligence, machine learning/deep learning, 3D modeling and analysis, image and video processing, cryptography and data compression, among others. Intel AVX-512 offers exceptional processing of encryption algorithms, helping to reduce the performance overhead for cryptography, which means EC2 M5 and M5d customers can deploy more secure data and services into distributed environments without compromising performance.

Q: What are the various processor options available to M5 customers?

The M5 and M5d instance types use a 3.1 GHz Intel Xeon Platinum 8000 series processor. The M5a and M5ad instance types use a 2.5 GHz AMD EPYC 7000 series processor.

Q: What are the various storage options available to M5 customers?

The M5 and M5a instance types leverage EBS volumes for storage. The M5d and M5ad instance types support up to 3.6TB (4 x 900GB) of local NVMe storage.

Q: When should I use the different M5 instance types?

Customers should consider using the M5a and M5ad instance types if they are looking to save money on price when their workloads do not fully utilize the compute resources of their chosen instance, resulting in them paying for performance that they don’t actually need. For workloads that require the highest processor performance or high floating-point performance capabilities, including vectorized computing with AVX-512 instructions, then we suggest you use the M5 or M5d instance types.

Q: Which network interface is supported on M5 instances?

M5, M5a, M5d, and M5ad instances support only ENA based Enhanced Networking and will not support netback. With ENA, M5 and M5d instances can deliver up to 25 Gbps of network bandwidth between instances and the M5a and M5ad instance types can support up to 20Gbps of network bandwidth between instances.

Q. Which operating systems/AMIs are supported on M5 Instances?

EBS backed HVM AMIs with support for ENA networking and booting from NVMe-based storage can be used with M5 instances. The following AMIs are supported on M5, M5a, M5ad, and M5d:

• Amazon Linux 2014.03 or newer
• Ubuntu 14.04 or newer
• SUSE Linux Enterprise Server 12 or newer
• Red Hat Enterprise Linux 7.4 or newer
• CentOS 7 or newer
• Windows Server 2008 R2
• Windows Server 2012
• Windows Server 2012 R2
• Windows Server 2016
• FreeBSD 11.1-RELEASE

For optimal local NVMe-based SSD storage performance on M5d, Linux kernel version 4.9+ is recommended.