Network-Accessible Memory Caching via Intermediary Logic

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Network storage systems face issues with access latency, limited memory addressability, and challenges in sharing state between failover partners and acquiring snapshots due to latency and synchronization problems.

Innovation Solution

A network-accessible memory (NAM) system comprising a substrate with mounted RAM, network interface circuitry, and logic circuitry, which operates as a memory controller to control access to the RAM via a network interface, providing a cost-effective solution for caching, SAN acceleration, and improved state sharing in cluster-failover configurations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If caching is used to reduce access latency, then access speed is improved, but device complexity and cost increase

Engineering Contradiction:
Improveaccess speedVSAvoiddevice complexity
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent introduces a network-accessible memory device as an intermediary component between the storage server and mass storage devices. This mediator provides caching functionality by intercepting read requests, checking for cached data in its memory, and serving data directly from cache when available, thereby reducing access latency without significantly complicating the existing storage architecture.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The network-accessible memory device uses volatile memory (such as DRAM) that is relatively inexpensive per gigabyte compared to persistent storage. While the memory is volatile and requires continuous power, its low cost and high speed make it an economical choice for caching frequently accessed data, accepting the trade-off of data loss on power failure in exchange for performance improvement.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

2Quantity of substance

If larger cache capacity is provided, then more data can be cached, but cost increases

Engineering Contradiction:
Improvecache capacityVSAvoidcost
Core Design Contradiction:
Quantity of substanceVSEase of manufacture

Solution Approach 1:

The patent changes the capacity parameter of the network-accessible memory device to match the size of the mass storage devices it caches, rather than using traditional small cache sizes. This parameter change enables caching of entire disk images or large volumes of data, significantly increasing cache capacity. The cost is managed by using cost-effective DRAM technology and scaling capacity according to specific caching needs rather than providing universal large capacity.

Inventive Principle:
Principle #35Parameter changes

3Loss of time

If network-accessible memory is used, then access latency is reduced, but memory addressability limitations remain

Engineering Contradiction:
Improveaccess latencyVSAvoidmemory addressability
Core Design Contradiction:
Loss of timeVSAdaptability or versatility

Solution Approach 1:

The patent segments the storage system into distinct components: the network-accessible memory device handling caching operations independently, and the mass storage devices handling persistent storage. This segmentation allows the memory device to operate with its own address space and memory management, avoiding the addressability limitations of the host system while still providing transparent caching services through network protocols.

Inventive Principle:
Principle #1Segmentation

4Reliability

If state sharing between failover partners is implemented, then reliability is improved, but synchronization latency increases

Engineering Contradiction:
ImprovereliabilityVSAvoidsynchronization latency
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The network-accessible memory device creates and maintains copies of data for failover partners. When one partner fails, the other can immediately access the cached copy from the memory device without waiting for synchronization over the network. This copying approach provides immediate failover capability while minimizing synchronization latency, as the memory device can serve multiple partners simultaneously with low-latency memory access.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS8316074B2Network-accessible memory
Publication Date: 2012.11.20 NETAPP INC
  • US8316074B2 patent drawing
  • US8316074B2 patent drawing
  • US8316074B2 patent drawing

AI summary

A network-accessible memory (NAM) element comprises a substrate, a memory mounted to the substrate, network interface circuitry and logic circuitry. The network interface circuitry implements a network physical layer and is mounted to the substrate. The logic circuitry is mounted to the substrate and is coupled to the network interface circuitry and the memory, and is configured to operate as a memory controller to control access to the memory by a host processor which is external to the NAM element through the network interface circuitry. The NAM element can be interconnected with other NAM elements to form a larger NAM storage device.