(Image: Kristin Wilson)

Leostream’s Remote Desktop Access Platform has new features that aim to improve high-performance computing (HPC) and complex imaging and video workloads on Amazon Elastic Compute Cloud (EC2). A further goal of these updates is to deliver secure, effective, cost-efficient resources to users of virtual machines.

Amazon EC2 instances supply users with on-demand, scalable computing capacity in the AWS Cloud. They are configured by the user in terms of CPU, memory, storage (temporary and persistent), networking capacity and graphics hardware, and are secured using public and private key pairs.

HPC Compute on EC2

Amazon EC2 has advantages for HPC compute and graphics rendering. It gives access to highly available, flexible cloud resources that are scalable and redundant. With EC2, organisations pay for computing as an operating expense (OPEX) that can be monitored and controlled, instead of a considerable up-front capital expense. AWS also supplies EC2 with NICE DCV, a GPU-accelerated display protocol, free of charge, which means users do not have to pay for third-party display protocol licenses.

To those advantages, Leostream adds more performance, security measures and controls on costs by simplifying end-user machine management – even for large user pools, large datasets and environments combining EC2 and on-premises resources. The Leostream Gateway is used to connect securely to EC2 without relying on VPNs, which are often expensive and slow down performance. The gateway also gives users access to the NICE DCV protocol to deliver GUIs to end users.

With identity management and authentication based on Zero-Trust Network Access (ZTNA), Leostream creates a zero-trust architecture around access to HPC workloads, ensuring the security of data and assets. Organisations gain visibility into the resources accessed by users and devices as Leostream tracks end-user connections, source locations, volumes/machines accessed and session durations for security audit purposes.

Remote Workforce

Since HPC environments are often used by a transient remote workforce, such as temporary contractors and project-based workers who only need a machine for the project, the Leostream platform can set up scheduled and scalable workloads. This approach simplifies the task of moving workers to another project, or granting them access to different resources.

Example HPC deployment with Leostream in AWS EC2.

Because HPC and other GPU-accelerated compute tasks like nonlinear editing are generally only active while a user is accessing the machine, Leostream monitors the user's session to power down machines when not in use. Automated control and power state management avoids unnecessary EC2 compute expenses due to user error, helping to further manage costs. For efficiency, Leostream supports architectures that limit data egress to computational resources nearer to the data, which results in better application performance and controls data transfer costs.

Mix and Manage

"Integrating EC2 and Leostream into a single, custom system is a way to provision high powered end-user machines in front of the flexibility and resiliency of cloud or hybrid cloud," said Leostream CEO Karen Gondoly. "Giving remote access to a secure, elastic HPC architecture with built-in policies and features to keep costs down is a potential step forward for industries engaged in resource-intensive tasks like media, scientific research, energy and finance."

Among remote desktop access systems, one of Leostream’s differentiators is its ability to mix and manage on-premises and cloud-based hosting environments. A unified admin console controls the platform, simplifying the management of users, cloud desktops and IT assets with real-time dashboard visibility on usage and environment details, helping to make informed decisions.  leostream.com