AWS-HPC-Lens pdf free download

AWS-HPC-Lens pdf free download.High Performance Computing Lens.
With HTC, the loss of one node or job doesn’t delay the entire calculation. The lost work can be picked up later or, often, even omitted altogether. The nodes involved in the calculation can vary’ in specification and power.
A suitable architecture for an HTC workload has the following considerations:
Network: Because parallel processes in HTC do not interact with each other, the feasibility or performance of the workloads is not sensitive to the bandwidth and latency capabilities of the network. Therefore, network requirements for HTC workloads are minimal. You should not use placement groups for this scenario because they’ would provide no performance gain (due to little to no communication between nodes) and potentially weaken resiliency.
Storage: HTC workloads have varying storage requirements. Storage requirements are usually driven by the dataset size and the performance requirements for moving, reading, and writing data. Occasionally, it’s helpful to use a shared file system (for example, EFS or NFS) between the nodes. High Performance File Systems are also an option, if desired. Cloud Native applications optimize the use of Amazon 83 object storage in conjunction with local storage.
• Compute: Each application is different, but in general, the application’s memory-to-compute ratio drives the underlying EC2 instance type. Some applications are optimized to take advantage of graphics processing units (GPUs) or field-programmable gate array (FPGA) accelerators on EC2 instances.
• Deployment: HTC simulations often run across many—sometimes millions—of compute instances. Due to their loosely coupled nature, simulations can be deployed across Availability Zones without sacrificing performance. HTC simulations can be deployed with end-toend solutions such as AWS Batch and CfnCluster, or through solutions based on AWS services such as Amazon Simple Queue Service (Amazon SQS), Auto Scaling, and AWS Lambda.
There are four example architectures to consider as a starting point for design patterns for HTC applications:
• Batch
• Queue
• Traditional
• Serverless
Batch-Based Architecture
AWS Batch is a fully managed service that enables you to run large-scale compute workloads on the cloud without having to provision resources or manage schedulers.3 AWS Batch enables developers, scientists, and engineers to easily and efficiently run hundreds of thousands of batch computing jobs on AWS. AWS Batch dynamically provisions the optimal quantity and type of compute resources (for example, CPU or memory-optimized instances) based on the volume and specified resource requirements of the batch jobs submitted. It plans, schedules, and executes your batch computing workloads across the full range of AWS compute services and features, such as Amazon EC24 and Spot Instances.5 Without the need to install and manage batch computing software or server clusters that you use to run your jobs, you can focus on analyzing results and gaining new insights.With AWS Batch, you package the code for your batch jobs, specify their dependencies, and submit your batch jobs using the AWS Management Console, CLIs, or SDKs. You can specify execution parameters and job dependencies and integrate with a broad range of popular batch computing workflow engines and languages (for example, Pegasus WMS, Luigi, and AWS Step Functions). AWS Batch provides default job queues and compute environment definitions that enable you to get started quickly. Reference Arcl hitectureAWS-HPC-Lens  pdf download.