AWS EC2 Instances
Amazon EC2 is service that provides scalable compute capacity in the cloud. In other words, the EC2 service of AWS gives us a virtual private machine in the cloud.
COMPUTE capacity refers to the Power required to satisfy your workload in the cloud. This computer power given by EC2 service in terms of EC2 Instances is scalable and dynamic. If your workload requires a small computer power, let us say Hosting a small website, a small EC2 instance capacity is enough. On the go, if your website visitors/usage is increased obviously you require a high computing power. These cases are handled by EC2 by scaling up or scaling down its compute power. That’s we say EC2 is scalable.
There are two key things to consider while launching an EC2 instance on AWS:
- The amount of virtual hardware dedicated to the instance(Instance Types)
- The software loaded on the
The Instance types define the virtual hardware of the EC2 Instances. There are various Instance types available which are mainly defined by these dimensions.
- vCPU (Virtual CPU)
- RAM (Memory)
- Hard Disk (Virtual Storage and Type)
- Network performance
AWS EC2 Instances Life Cycle:
Instance types are grouped into Families based on the above dimensions.
- General purpose – T2,T3,M4,M5
- Computer Optimized – C4,C5
- Storage Optimized – H1, I3, D2
- Memory Optimized – R4,R5, X1, X1e, Z1d
- Accelerated Computing – P2, P3, G3, F1
General Purpose — For workloads ranging from low, mid-level processing. Compute-optimized—For workloads requiring significant processing Memory optimized—For memory-intensive workloads
Storage optimized—For workloads requiring high amounts of fast SSD storage GPU-based instances—Intended for graphics and general-purpose GPU compute workloads.
General Purpose -T2 Instance
T2 instances are the best burstable instances which can burst its performance beyond the baseline performance.T2 unlimited instances can sustain high CPU performance without incurring additional charges. The baseline performance is governed by CPU Credits. T 2 instances receive CPU Credits continuously at a set rate depending on the instance size, accumulating CPU Credits when they are idle, and consuming CPU credits when they are active.
- Low cost
- High-frequency Xenon processors
- Burstable CPU are the features of the T2 instance
Commonly T2 Instances are used widely in web applications, websites, Microservices, Test and staging environments. The following table gives the Instance types on the General-Purpose T2 category.
General Purpose – T3 Instance
T3 instances are the next generation burstable instances which can burst as long as required.T 3 instances are designed for the applications with moderate CPU usage which is having tempo rary spikes in it. T3 instances accumulate CPU credits when a workload is operating below bas eline threshold. Each earned CPU credit provides the T3 instance the opportunity to burst wit h the performance of a full CPU core for one minute when needed.
- Burstable CPU
- Unlimited mode by
T3 instances are best suited for micro-services, low-latency interactive applications, small and medium databases, virtual desktops, development environments, code repositories, and business-critical applications. The following table gives the Instance types on the General-Purpose T3category.
General Purpose – M4 Instance
M4 instances are the good fit for many applications as it gives a balance of Compute, memory And networking resources.
- 3/2.4 GHz Intel Xeon processors ( Broadwell & Haswell)
- Good Network
Data processing, backend servers, databases etc.
General Purpose – M5 Instance
M5 is the latest generation general purpose machines. And the features are more similar to that of M4 instances.
- 5 GHz Intel Xeon® Platinum 8175 processors with the advanced instruction set.
- Maximum of 25Gbps bandwidth
- Larger sizes than M4
- Local Nvme SSD’s
Databases, data processing tasks, caching fleets, Backend servers, Microsoft SharePoint and other enterprise applications.
Compute Optimized – C5 Instance
When we want to run a high compute-intensive workload then EC2-Compute optimized instances are the best choice. Cost effective and high performance.
- 0 GHz Intel Xeon Platinum processors with Intel Advanced Vector Extension instruction set.
- Up to 5 GHz per core using Intel Turbo Boost Technology
- Storage offered via EBS or NVMe SSD’s
- Maximum of 25Gbps bandwidth using ENA
Web servers, Batch processing, distributed analytics, high-performance computing, machine l earning, highly scalable multiplayer gaming, and video encoding
Compute Optimized– C4 Instance
Designed for compute-intensive workloads and deliver very cost-effective high-performance machines in the cloud.
- High-frequency Intel Xeon E5-2666 v3 (Haswell) processors optimized specifically for EC2
- EBS-optimized storage for good performance
- Higher networking performance
Web servers, Batch processing, distributed analytics, high-performance computing, machine learning, highly scalable multiplayer gaming, and video encoding
Memory Optimized – R4 Instance
R4 Instances are designed for memory-intensive workloads
- High-Frequency Intel Xeon (Broadwell) processors
- DDR4 Memory
- Enhanced Networking
Big data analytics, Real-time Processing of unstructured data, Hadoop and spark clusters, data
mining and other enterprise level applications.
Memory Optimized– R5 Instance
R5 well suits for memory intensive applications which provides an additional memory of 5% than R4 instance. R5 instances deliver with an increased CPU performance than R4.
- Up to 768 GiB of memory per instance
- Skylake processors with turbo speed
- Nitro system
Real-time big data analytics, High-performance databases, distributed webscale in-memory caches, In-memory databases, and other enterprise applications.
Memory Optimized – X1,X1e Instance
Memory-optimized X1, X1e instances gives the lowest price per GB RAM among the EC2 instances and are optimized for large-scale enterprise-class in-memory applications.
- High-frequency Intel Xeon (Haswell) processors
- Lowest price for RAM
- Up to 3TB DRAM-based memory in X1e and 2TB offered in X1 based machines
- SSD instance storage for temporary block-level storage
- EBS-optimized by default at no additional cost
In-memory databases, big data processing engines, high-performance computing. Certified by SAP to run the SAP workloads.
Memory Optimized – Z1d Instance
z1d instances offers both high compute capacity and a high memory. It provides the fastest sustained high core frequency of about 4.0 GHz range.
- Xenon processor with sustained all core frequency of up to 0 GHz
- 8 Tb of storage
- Memory up to 384 GB
- Nitro system, which eliminates all virtualization overhead
- With z1d instances, local NVMe-based SSDs are physically connected to the host server an d provide block-level storage that is coupled to the lifetime of the z1d instance
Good for electronic design automation (EDA) and relational database workloads.
Storage Optimized – D2
These D2 instances offer low price, high disk throughput and it can feature up to 48 TB of Hard Disk-based local storage.
- Intel Xeon (Haswell) processors that provide greater
- Hard Disk-based storage(HDD)
- High disk throughput
- Support for Enhanced Networking
Distributed computing systems like Hadoop, MapReduce, data processing applications like apache Kafka flume, Parallel processing data warehouses.
Storage Optimized – H1
These H1 instance category can offer up to 16 TB of HDD based systems which delivers high disk throughput.
Powered by 2.3 GHz Intel® Xeon® processors (codenamed Broadwell) Up to 16TB of HDD storage
High disk throughput Networking Up to 25 Gbps(ENA) Use cases:
Distributed computing systems like Hadoop, MapReduce, data processing applications, Parallel
processing data warehouses.
Storage Optimized – I3
I3 Instances are built for low latency, very high random I/O performance, high sequential read throughput and provide high IOPS at a low cost. It also offers bare metal for a non-virtualized environment.
- High-Frequency Intel Xeon (Broadwell) Processors with the base frequency of 3 GHz
- Up to 25 Gbps of network bandwidth
- Elastic Network Adapter (ENA)-based Enhanced Networking
- High Random I/O performance and High Sequential Read Throughput
- Support for bare metal Instance(i3.metal)
Suitable for NoSQL databases, In-memory Databases, Scale-out transactional databases, Data warehouse and analytical workloads.
Accelerated Computing – P3, P2
P3, P2 instances are intended for general-purpose GPU compute applications.
- High-performance NVIDIA GPUs with parallel
- Supports Nvlink, GPUDirect™ for peer-to-peer GPU communications
- High-frequency Intel Xeon (Broadwell) processors
- Elastic Network Adapter (ENA)-based Enhanced Networking(25 Gbps bandwidth)
Machine/Deep learning, high-performance computing, computational fluid dynamics, computational finance, seismic analysis, speech recognition, autonomous vehicles, drug discovery.
Accelerated Computing – G3
Optimized for Graphic intensive applications
- High-frequency Intel Xeon (Broadwell) processors
- NVIDIA Tesla GPUs with parallel processing
- Enables NVIDIA GRID Virtual Application capabilities for application virtualization software like Citrix XenApp Essentials and VMware Horizon
- Enhanced Networking using the Elastic Network Adapter (ENA) with 25 Gbps of aggregate
network bandwidth within a Placement Group
3D visualizations, graphics-intensive remote workstation, 3D rendering, application streaming, video encoding, and other server-side graphics workloads.
Accelerated Computing – F1
Customizable hardware acceleration with field programmable gate arrays (FPGAs).
- Xilinx FPGAs
- 64 GiB of ECC-protected memory on 4x DDR4
- Dedicated PCI-Express x16 interface
- Approximately 5 million logic elements
- Approximately 6,800 Digital Signal Processing (DSP) engines
- FPGA Developer AMI
Genomics research, financial analytics, real-time video processing, big data search, and analysis and security.