Amazon Elastic Compute Cloud (EC2) I4g, Im4gn, and Is4gen instances are next-generation, storage-optimized instances designed to run applications that require high throughput and low-latency access to large amounts of data on local SSD storage. They are powered by AWS Graviton2 processors and provide up to 30 TB of local NVME storage using AWS Nitro SSDs. AWS Nitro System SSDs are custom built by AWS and offer high I/O performance, low latency, minimal latency variability, and security with always-on encryption.
I4g instances offer similar memory and storage ratios to existing I4i instances and are optimized for workloads with small to medium sized datasets that perform a high mix of random read/write and require very low I/O latency, such as databases and real-time analytics. I4g instances deliver the best compute price performance for a storage optimized instance, and best storage performance per TB for a Graviton-based storage instance. Amazon EC2 I4g instances deliver up to 15% better compute performance compared to similar storage-optimized instances.
Im4gn and Is4gen instances are optimized for workloads that manage larger datasets and need a higher storage density per vCPU such as large-scale online analytical processing (OLAP), distributed file systems and Apache Hadoop distributed file systems. Im4gn and Is4gen instances provide up to 56% lower cost per TB than I4g. Is4gen provides highest density per vCPU of SSD storage and lowest cost per TB in Amazon EC2. In combination with highest network throughput per vCPU among storage-optimized instances, Im4gn and Is4gen instances are suitable for distributed file systems and distributed databases that need to rapidly move data across nodes in a cluster.
High performance for storage-intensive workloads
I4g instances deliver the best compute price performance for a storage-optimized instance, and the best storage performance per TB for a Graviton-based storage instance.
Low-cost SSD storage
Im4gn and Is4gen instances based on AWS Nitro SSD offer up to 30TB of local NVMe storage and provide up to 45% and 60% lower cost per TB respectively, compared to I4i instances. Is4gen instances provide the lowest cost per TB of SSD storage in Amazon EC2.
I4g, Im4gn, and Is4gen instances are powered by AWS Graviton2 processors, feature AWS Nitro SSDs, and are built on the AWS Nitro System. AWS Graviton2 processors feature always-on 256-bit DRAM encryption. The data on the AWS Nitro SSDs is always encrypted using ephemeral keys for maximum security. The AWS Nitro System features the AWS Nitro Security Chip, which provides the most secure cloud platform with a minimized attack surface and prohibited administrative access to prevent human error and tampering. These instances also support encrypted Amazon Elastic Block Store (EBS) volumes by default.
Maximized resource efficiency
I4g, Im4gn, and Is4gen instances are built on the AWS Nitro System, a combination of dedicated hardware and lightweight hypervisor that delivers practically all of the compute and memory resources of the host hardware to your instances for better overall performance.
Flexibility and choice
With I4g, Im4gn, and Is4gen instances, you can optimize your infrastructure based on your workload needs for the best price performance or for the lowest cost per TB of storage in Amazon EC2. I4g, Im4gn, and Is4gen instances add to the broadest and deepest portfolio of instances in the cloud and provide you with six different instances sizes that feature varying amounts of vCPU, memory, networking, and storage.
Extensive software support
AWS Graviton-based instances are supported by popular Linux operating systems, including Amazon Linux 2, Red Hat Enterprise Linux, SUSE, and Ubuntu. Many popular applications and services for security, monitoring and management, containers, and CI/CD from AWS and software partners also support AWS Graviton-based instances. The AWS Graviton Ready program provides customers with certified solutions from partner software vendors that can be used on Graviton-based instances.
Powered by AWS Graviton2 processors
AWS Graviton2 Processors are custom designed by AWS and based on 64-bit Arm Neoverse cores for optimized performance and cost. AWS Graviton2 Processors deliver 7x more performance, 4x more compute cores, 5x faster memory, and 2x larger caches versus first-generation AWS Graviton Processors.
High-performance storage and networking
I4g, Im4gn, and Is4gen instances feature AWS Nitro SSDs that are custom built by AWS to provide up to 60% lower I/O latency and up to 75% reduced latency variability compared to third-generation storage-optimized instances for workloads sensitive to I/O performance. Im4gn instances provide up to 100 Gbps networking bandwidth. I4g and Im4gn instances offer support for Elastic Fabric Adapter (EFA).
Built on AWS Nitro System
The AWS Nitro System is a rich collection of building blocks that offloads many of the traditional virtualization functions to dedicated hardware and software to deliver high performance, high availability, and high security while also reducing virtualization overhead.
Featuring AWS Nitro SSD
AWS Nitro SSDs build on the AWS silicon innovation with the AWS Nitro System and are custom-designed to deliver the best storage performance for your I/O intensive workloads running in Amazon EC2. Traditionally, SSDs maximize the peak read and write I/O performance. AWS Nitro SSDs are architected to minimize latency and latency variability of I/O intensive workloads, such as SQL and NoSQL databases, search engines, distributed file systems, and data analytics that continually read and write from the SSDs in a sustained manner, for fast and more predictable performance. AWS Nitro SSDs deliver up to 60% lower storage I/O latency and up to 75% reduced storage I/O latency variability compared to third generation EC2 storage-optimized instances. Data on the AWS Nitro SSDs is always encrypted using ephemeral keys. AWS Nitro SSDs also help AWS innovate faster by providing firmware updates at cloud scale and speed. AWS Nitro SSDs support Torn Write Prevention (TWP), which improves database performance and service level agreements (SLAs) without compromising data resiliencies, for workloads using MySQL and MariaDB database.
“Elastic is the leading data analytics platform for Observability, Security, and Search powered applications. At Elastic, we tested the i4g instances, the latest addition to the storage-optimized instances family, powered by AWS Graviton2 processors. Across the benchmarks that represent the performance of common use-case patterns of customers using Elasticsearch, we saw up to 35% better price-performance compared to the previous generation I3 instances. We are excited about the introduction of i4g instances which promises to provide improved price-performance for Elasticsearch users deploying on AWS.”
Yuvraj Gupta, Principal Product Manager - Elastic
As the only unified customer experience management platform (Unified-CXM), Sprinklr helps the world's largest and most valuable companies make their customers happier across 30+ digital channels—using industry-leading AI to create insight-driven strategies and better customer experiences.
“Sprinklr provides a unified customer experience management (Unified-CXM) platform that combines different applications for marketing, advertising, research, customer care, sales, and social media engagement. As a cloud-first company, Sprinklr is always looking to improve efficiency and optimize its cost savings. Sprinklr’s datasets require very low I/O latency and need high performance. After testing out, Amazon EC2 I4g instances, Sprinklr is seeing on average a 12% improvement in overall performance and up to 27% in some cases versus Amazon EC2 I4i instances. Sprinklr is looking forward to taking advantage of the compute price performance of Amazon EC2 I4g instances to provide faster results for their customers while optimizing costs.”
Abhay Bansal Senior Vice President of Engineering - Sprinklr
“Aerospike is excited to have their customers use the recently launched Amazon EC2 I4g storage optimized instance family. Our customers make use of the existing i3 and i4i storage optimized instance families, which deliver high performance local NVMe storage and come with the disk to memory ratios that work well for our Hybrid Memory ArchitectureTM. Aerospike, supports the AWS Graviton processor platform. In our testing of the Graviton2, for CPU bound workloads, we found a 1.6x improvement in price performance. We’re excited by the launch of I4g instances and look forward to have customers obtain the price performance savings.”
Srini Srinivasan, Chief Technology Officer and Founder - Aerospike
Splunk is the world's first Data-to-Everything Platform designed to remove the barriers between data and action, so that everyone thrives in the Data Age.
“We run C/C++ based workloads for indexing and searching event data. Our workload is CPU-bound and benefits from high-capacity and low-latency SSD storage. When evaluating the new Im4gn/Is4gen instances powered by AWS Graviton2, we observed an up to 50% decrease in search runtime compared to I3/I3en instances, which we currently use. This makes Im4gn and Is4gn instances a great choice for running our storage-intensive workloads with significant price performance benefits.”
Brad Murphy, Vice President, Cloud Operations & Infrastructure - Splunk
“As part of our continued effort to improve performance and reduce our TCO to maximize cost savings for our customers, we've actively benchmarked our OpenSearch workloads on AWS Graviton2-based Is4gen instances. These tests illustrated that with smaller clusters of Is4gen instances we were able to achieve optimal performance while yielding a 25% cost savings, compared to i3en instances. We are looking forward to moving all of our OpenSearch production workloads to Is4gen instances while working to benchmark additional workloads including our Kubernetes microservices and Kafka components.”
Jonah Kowall, CTO - Logz.io
"Honeycomb develops an observability platform that enables engineering teams to visualize, analyze, and improve cloud application quality and performance. We were able to validate our Kafka workload against preview instances of the im4gn instance family. Without needing to modify our Confluent Platform configuration, we have seen a +43% price-performance improvement on im4gn.4xlarge for our workload over equivalent i3en.6xlarge instances. We're excited by the guaranteed high-network throughput and higher-performance cores, allowing us to better match the shape of our workload to available instance families."
Emily Nakashima, VP Engineering - honeycomb.io
Lightbits makes high-performance data storage simple, agile, and cost-efficient.
"We are excited to partner with Amazon Web Services to bring the full benefits of our software-defined data storage solution to AWS Graviton. Porting to Graviton2-based instances was much easier than we anticipated when taking into account the highly-optimized nature of Lightbits software. Early testing shows that Lightbits on Graviton2-based Is4gen instances can deliver up to 20% price performance improvement as compared to x86-based I3en instances. With AWS Graviton, Lightbits can deliver high performance, resilient data storage that is easy to use and provides greater efficiency to its customers."
Muli Ben-Yehuda, Co-Founder and Chief Scientist at Lightbits Labs
Instance Size vCPU Memory (GiB) Instance Storage (GB) Network Bandwidth (Gbps) EBS Bandwidth (Gbps) i4g.large 2 16 1 x 468 GB = 468 GB Up to 10 Up to 10 i4g.xlarge 4 32 1 x 937 GB = 937 GB Up to 10 Up to 10 i4g.2xlarge 8 64 1 x 1,875 GB = 1,875 GB Up to 12 Up to 10 i4g.4xlarge 16 128 1 x 3,750 GB = 3,750 GB Up to 25 Up to 10 i4g.8xlarge 32 256 2 x 3,750 GB = 7,500 GB 18.75 10 i4g.16xlarge 64 512 4 x 3,750 GB = 15,000 GB 37.5 20
Instance Size vCPU Memory (GiB) Instance Storage (GB) Network Bandwidth (Gbps) EBS Bandwidth (Gbps) Im4gn.large 2 8 1 x 937 NVMe SSD Up to 25 Up to 9.5 Im4gn.xlarge 4 16 1 x 1875 NVMe SSD Up to 25 Up to 9.5 Im4gn.2xlarge 8 32 1 x 3750 NVMe SSD Up to 25 Up to 9.5 Im4gn.4xlarge 16 64 1 x 7500 NVMe SSD 25 9.5 Im4gn.8xlarge 32 128 2 x 7500 NVMe SSD 50 19 Im4gn.16xlarge 64 256 4 x 7500 NVMe SSD 100 38
Instance Size vCPU Memory (GiB) Instance Storage (GB) Network Bandwidth (Gbps) EBS Bandwidth (Gbps) Is4gen.medium 1 6 1 x 937 NVMe SSD Up to 25 Up to 9.5 Is4gen.large 2 12 1 x 1875 NVMe SSD Up to 25 Up to 9.5 Is4gen.xlarge 4 24 1 x 3750 NVMe SSD Up to 25 Up to 9.5 Is4gen.2xlarge 8 48 1 x 7500 NVMe SSD Up to 25 Up to 9.5 Is4gen.4xlarge 16 96 2 x 7500 NVMe SSD 25 9.5 Is4gen.8xlarge 32 192 4 x 7500 NVMe SSD 50 19