Application Device Queues (ADQ) is a queuing and steering technology. By deploying Intel® Ethernet 800 Series Network Adapters with ADQ, organizations can improve predictability and performance to deliver a better customer experience and consistently meet SLAs.
ADQ now supports a broader range of workloads and environments: containerized, bare metal, virtualized. ADQ is also easier to configure and deploy.
Managing latency, predictability and throughput on containerized workloads or CDNs can be challenging. Learn how the latest features in ADQ can help.
Twitter and Intel Collaborate on Pelikan Cache Acceleration with ADQ for Dramatic Tail Latency Improvement
When maintaining performance during peak usage periods, response times are limited by the slowest outliers (tail latency). Twitter and Intel collaborated to achieve an up to 10X tail latency improvement by accelerating Twitter’s Pelikan Cache framework with ADQ.1 The reduction in tail latency enables Twitter to provide stricter latency targets and give customers access to more data.
Scalable, Low-Latency Storage Using ADQ and LightOS
Through a strategic collaboration, Intel and Lightbits Labs deliver greater performance, flexibility, and lower TCO for NVMe over Fabrics (NVMe-oF)-based disaggregated storage. Initial results: NVMe/TCP with a LightOS cluster demonstrates up to 30% predictability increase, up to 50% latency reduction and up to 70% IOPS improvement when using ADQ vs. without ADQ.2
Aerospike: First Commercial Database to Support Intel® Ethernet 800 Series with ADQ
Get useful insights from big data in real time. Based on Aerospike’s testing in the Intel Labs, ADQ can result in greater than 75% improved throughput and, most importantly, greater than 45% increased response-time predictability for the Aerospike 4.7 or later next-generation NoSQL database running on a dual-socket server based on 2nd Gen Intel® Xeon® Scalable processors, configured with two Intel® Ethernet 800 Series network adapters with ADQ technology compared to the previously optimized approach of NUMA pinning.