Accelerate Outposts Implementation for Low Latency Crypto Applications

In the ever-evolving landscape of cloud computing, the demand for ultra-low latency solutions has never been higher. AWS Outposts has been at the forefront of bringing cloud capabilities to on-premises environments, and now, with the introduction of specialized Amazon EC2 instances featuring accelerated networking, the bar for performance has been raised even higher.

The second-generation Outposts rack announced on April 29th now provides additional low latency, local data processing, or data residency needs, such as game servers for multi-player online games, crypto transaction data, medical records, industrial and manufacturing control systems, telecom Business Support Systems (BSS), and edge inference of a variety of machine learning (ML) models. Customers can now take advantage of the latest generation of processors and more advanced configurations of Outposts racks to support faster processing, higher memory capacity, and increased network bandwidth.

First, you can now scale your compute resources independently from your networking infrastructure, giving you more flexibility and cost efficiency as your workloads grow. Second, we’ve built in network resilience from the ground up, with the network rack automatically handling device failures to keep your systems running smoothly. Third, connecting to your on-premises environment and AWS Regions is now a breeze – you can configure everything from IP addresses to VLAN and BGP settings through straightforward APIs or our updated console interface.

Leading this new capability includes the bmn-sf2e instances, engineered for ultra-low latency with deterministic performance. These cutting-edge instances are powered by Intel's latest Sapphire Rapids processors (4th Gen Xeon Scalable), delivering an impressive 3.9 GHz sustained performance across all cores. With a generous memory allocation of 8GB of RAM for every CPU core, these instances are primed for demanding workloads. The bmn-sf2e instances are equipped with AMD Solarflare X2522 network cards, connecting directly to top-of-rack switches for optimal network performance.

For financial services customers, particularly those in capital markets, these instances offer game-changing capabilities. They provide deterministic networking through native Layer 2 (L2) multicast, precision time protocol (PTP), and equal cable lengths. This enables customers to meet stringent regulatory requirements around fair trading and equal access while seamlessly integrating with existing trading infrastructure.

When deploying these high-performance instances on AWS Outpost, proper configuration and power planning are critical success factors. In this article, we'll explore the key considerations for optimizing your Outpost setup to harness the full potential of these ultra-low latency instances. From form factor selection to network configuration, power requirements, and cooling considerations, we'll cover the essential aspects that contribute to a successful deployment.

Let's dive into what you need to consider when configuring your AWS Outpost for ultra-low latency applications. This information will be critical to accelerate the ordering process and provide the necessary information to your data center team.

Network configuration is crucial for achieving ultra-low latency with AWS Outpost. Begin by carefully planning and selecting high-bandwidth, dedicated connections between your Outpost and local network infrastructure. This preparation is essential for the upcoming Outpost Site Assessment Validation (SAV) stage, which typically occurs a few weeks after submitting your Outpost order. During the SAV, AWS representatives will conduct an on-site review of your environment. This critical step involves verifying AWS connectivity and assessing your specific connection types for Outpost compatibility, ensuring a smooth deployment and optimal performance.

Power selection and connector types:

• Uplink speed (Gbps) Choices: 10G, 40G, or 100G Choices in AWS Console
• Uplink count: FNE to Customer Switches
• Fiber type: MULTI MODE or SINGLE MODE
• Optical standard: 40GBASE-SR
• Connector Type LC MMF or SMF Choices

Second, proper power planning is essential for maintaining the reliability and performance of your Outpost deployment. Begin by calculating the total power consumption of your chosen server configurations by collaborating with your account team, factoring in both typical and peak usage scenarios. Add a safety margin to account for potential future expansion. Implement N+1 power redundancy at minimum, meaning you have at least one more power supply than necessary to run your equipment. This ensures continuous operation even if a single power supply fails. Choose Power Distribution Units (PDUs) that can handle your peak load requirements with room for growth. Consider implementing intelligent PDUs that allow for remote power monitoring and management. Uninterruptible Power Supply (UPS) systems are crucial for maintaining power during brief outages and providing time for graceful shutdowns during extended power losses. Size your UPS to support critical loads for a duration that aligns with your business continuity plans.

Power selection and connector types:

• Power draw (kVA): 15KVa or 30KVA+ Choices in AWS Console
• Power phase: SINGLE PHASE (Typically: 6 Drops 3 primary and 3 redundant for compute and 2 pairs for for Network rack), Redundant Single Phase, 3 Phase, or Redundant 3Phase Choices AWS Console
• Power connector: L6_30P (Most Common for Single Phase)
• Power feed drop: ABOVE RACK or BELOW RACK in Post Order SAV review stage

Cooling Considerations:

Effective cooling is crucial for maintaining optimal performance and longevity of your Outpost hardware. Begin by calculating the total heat output of your server configuration and ensure your cooling system can handle this load plus a safety margin. Consider implementing a hot aisle/cold aisle configuration to maximize cooling efficiency. Monitor ambient temperature and humidity levels continuously, as these factors can significantly impact hardware performance and lifespan. Implement environmental sensors throughout your rack and data center to provide real-time monitoring and alerting. Plan for proper airflow management within your rack setup, using blanking panels to prevent hot air recirculation and ensuring adequate space between equipment for ventilation. Consider implementing liquid cooling solutions for high-density configurations or in environments where traditional air cooling may be insufficient.

• Ensure adequate cooling capacity for your chosen configuration
• Monitor ambient temperature and humidity levels
• Plan for proper airflow management in your rack setup

Timeline Considerations and Steps to Ordering:

1. Initial SAV submission after order in Console: 1-2 weeks
2. Documentation review: 1-2 weeks
3. Virtual validation: 1 week
4. On-site validation (if needed): 1-2 days
5. Remediation period: Varies based on requirements
6. Final approval for shipping: 1 week
7. Shipping West Coast to East Coast: Typically 5 days

The SAV process is crucial for ensuring a successful Outpost deployment and ongoing operation. It helps identify potential issues early and ensures proper planning for a smooth implementation.

To learn more, visit the AWS Outposts racks product page and user guide. You can also review this blog post or talk to an Outposts expert if you are ready to discuss your on-premises needs.