US 7,925,711 B1 · Centralized adaptive network memory engine

Abstract

There is a constant battle to break even between continuing improvements in DRAM capacities and the growing memory demands of large-memory high-performance applications. Performance of such applications degrades quickly once the system hits the physical memory limit and starts swapping to the local disk. We present the design, implementation and evaluation of Anemone—an Adaptive Network Memory Engine—that virtualizes the collective unused memory of multiple machines across a gigabit Ethernet LAN, without requiring any modifications to the either the large memory applications or the Linux kernel. We have implemented a working prototype of Anemone and evaluated it using real-world unmodified applications such as ray-tracing and large in-memory sorting. Our results with the Anemone prototype show that unmodified single-process applications execute 2 to 3 times faster and multiple concurrent processes execute 6 to 7.7 times faster, when compared to disk based paging. The Anemone prototype reduces page-fault latencies by a factor of 19.6—from an average of 9.8 ms with disk based paging to 500 μs with Anemone. Most importantly, Anemone provides a virtualized low-latency access to potentially “unlimited” network memory resources.

Claims10

Dependent claims

A distributed remote memory access system, comprising:

(1) a client communication network interface to a packet data communication network subject to packet loss, which is adapted to operate without communication of an acknowledgement packet in response to a successful communication of a data packet,
(2) a client having a client memory in adapted to store memory pages, and
(3) a reliable client remote memory access protocol, adapted to:
when a memory page is to be evicted from said client memory, communicate said memory page to a cooperating server through the client communication network interface network, and after receipt of an acknowledgment of successful communication of the memory page, evict said memory page from said client memory, and
when said evicted memory page is required at said client memory, communicate through said client communication network interface, destined to the cooperating server, a request for said evicted memory page and control retrieval of said evicted memory page from the cooperating server to said client memory.

2 more elements

Dependencies are inferred from the claim text.

Centralized adaptive network memory engine

Life of this patent

Centralized adaptive network memory engine

Life of this patent

Abstract

Claims10