Programmable NIC and Optical Switching Fabric for Data Center Topologies
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data center network designs face bottlenecks due to increased traffic volumes, leading to higher latency and server hops, and are limited by the number of ports in existing switches, which are expensive to scale and inefficient in virtualized environments.
Innovation Solution
A server cluster architecture with a programmable network interface card (NIC) and optical port routing enclosure that enables high-speed communication between servers using a distributed switch design, minimizing latency and server hops by providing a flexible and scalable network topology with a hybrid software-hardware approach.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional tree topology with routers and switches is used, then network coverage is achieved, but latency and number of hops increase
Solution Approach 1:
The patent segments the network into rack-level domains where servers within the same rack communicate directly via optical connections, bypassing the need for hierarchical routing through multiple switches and routers. This segmentation reduces the number of hops and latency for intra-rack communication.
Solution Approach 2:
The patent introduces an optical switching fabric as an intermediary that enables direct optical paths between servers across different racks without traversing traditional electrical switches and routers. This intermediary optical infrastructure reduces latency and hop count while maintaining network coverage.
2Quantity of substance
If more ports are added to switches to handle increased traffic, then network capacity increases, but cost and complexity increase
Solution Approach 1:
The patent replaces traditional electrical switching mechanisms with optical switching technology. Optical switches provide higher port density and bandwidth capacity without proportionally increasing complexity or cost, as optical infrastructure can handle multiple wavelengths and channels simultaneously.
Solution Approach 2:
The patent adds a spatial dimension to network capacity by implementing a 3D optical switching fabric that provides multiple dimensional paths for traffic flow. This allows network capacity to scale without linearly increasing switch port counts, as traffic can be routed through multiple spatial dimensions simultaneously.
3Productivity
If cluster size increases to accommodate more servers, then processing capacity increases, but latency and congestion increase
Solution Approach 1:
The patent segments the large cluster into smaller rack-level domains with direct optical connectivity. Servers within each segment can communicate with low latency, while the optical switching fabric provides high-capacity inter-segment communication, allowing the overall cluster to scale without proportionally increasing latency.
Solution Approach 2:
The patent creates a universal optical infrastructure that serves multiple functions simultaneously: intra-rack communication, inter-rack communication, and high-capacity backbone transport. This multi-functional optical network allows the cluster to scale in processing capacity without adding separate communication infrastructures that would increase latency.
Data Source
AI summary
A network interface card (NIC) and a method for stablishing a connection between virtual machines of a network. The NIC includes: a programmable switching ASIC (application-specific integrated circuit), a central processing unit (CPU), multiple Ethernet controllers, and multiple on-board transceivers functioning as external ports. The switching ASIC functions as a switch that manipulates data traffic within the NIC including by switching the data traffic between and among the CPU, the Ethernet controllers, and the on-board transceivers. The method includes: installing rules that route a Synchronize (SYN) packet from a source virtual machine (VM) through a software engine, appending a signed cookie to the SYN packet; verifying that a policy represented by the signed cookie appended to the SYN packet matches a policy of a destination VM; and returning the SYN packet to the source VM which establishes a connection between the source VM and the destination VM.