fat-tree network topology

E679889

interconnection network network topology

A fat-tree network topology is a hierarchical, tree-like interconnection structure for parallel and distributed systems that increases link bandwidth toward the root to avoid bottlenecks and provide high bisection bandwidth and scalability.

Try in SPARQL Jump to: Surface forms Disambiguation Statements Elicitation Referenced by

All labels observed (3)

Label	Occurrences
Clos topology	1
fat-tree data structure in computer science	1
fat-tree network topology canonical	1

How this entity was disambiguated

This entity first appeared as the object of triple T7666043 — resolving that mention is where its identity was fixed. The disambiguator weighed these candidate entities and picked the highlighted one (or “None”, minting a new entity). This is how homonymy is resolved: the same surface form can point to different entities.

NED1 Entity disambiguation (via context triple) gpt-5-mini-2025-08-07

Target entity: fat-tree network topology
Context triple: [Charles E. Leiserson, notableConcept, fat-tree network topology]

A. Network-in-Network architecture
Network-in-Network architecture is a convolutional neural network design that replaces traditional linear convolution layers with micro multilayer perceptrons (MLPs) to enhance feature abstraction and model expressiveness.
B. Equinix Fabric
Equinix Fabric is a software-defined interconnection service that enables private, on-demand connectivity between enterprises, cloud providers, and network services within Equinix’s global data center ecosystem.
C. OSA-Express networking
OSA-Express networking is IBM’s high-speed, integrated network adapter technology for mainframe systems, providing advanced Ethernet and IP connectivity for IBM System z environments.
D. Next Generation Network architectures
Next Generation Network architectures are advanced telecommunications frameworks that integrate voice, data, and multimedia services over a unified, packet-based IP infrastructure to enable flexible, scalable, and service-agnostic communication.
E. AVE network
The AVE network is Spain’s high-speed rail system that connects major cities across the country with fast, long-distance train services.
F. None of above. chosen
G. Unsure - the case is ambiguous/there is not enough information to decide.

NED2 Entity disambiguation (via description) gpt-5-mini-2025-08-07

Target entity: fat-tree network topology
Target entity description: A fat-tree network topology is a hierarchical, tree-like interconnection structure for parallel and distributed systems that increases link bandwidth toward the root to avoid bottlenecks and provide high bisection bandwidth and scalability.

A. Network-in-Network architecture
Network-in-Network architecture is a convolutional neural network design that replaces traditional linear convolution layers with micro multilayer perceptrons (MLPs) to enhance feature abstraction and model expressiveness.
B. Equinix Fabric
Equinix Fabric is a software-defined interconnection service that enables private, on-demand connectivity between enterprises, cloud providers, and network services within Equinix’s global data center ecosystem.
C. OSA-Express networking
OSA-Express networking is IBM’s high-speed, integrated network adapter technology for mainframe systems, providing advanced Ethernet and IP connectivity for IBM System z environments.
D. Next Generation Network architectures
Next Generation Network architectures are advanced telecommunications frameworks that integrate voice, data, and multimedia services over a unified, packet-based IP infrastructure to enable flexible, scalable, and service-agnostic communication.
E. AVE network
The AVE network is Spain’s high-speed rail system that connects major cities across the country with fast, long-distance train services.
F. None of above. chosen

Statements (50)

Predicate	Object
instanceOf	interconnection network ⓘ network topology ⓘ
basedOn	tree topology ⓘ
canBeImplementedWith	commodity Ethernet switches ⓘ multi-stage switch networks ⓘ
commonlyUsedFor	cloud data center fabrics ⓘ cluster interconnects ⓘ
connects	aggregation switches to core switches ⓘ edge switches to aggregation switches ⓘ end hosts to edge switches ⓘ
differsFrom	Clos network ⓘ linked to: Clos networks bus topology ⓘ mesh topology ⓘ ring topology ⓘ simple tree topology ⓘ
enables	full bisection bandwidth with sufficient switch ports ⓘ
hasAdvantage	can be built from identical switch building blocks ⓘ good scalability with number of switches ⓘ
hasCharacteristic	higher link capacity near the root ⓘ multiple equal-cost paths between hosts ⓘ
hasDesignGoal	avoid bottlenecks near the root ⓘ provide high aggregate bandwidth ⓘ support large-scale system expansion ⓘ
hasLevel	aggregation layer ⓘ core layer ⓘ edge layer ⓘ
hasLimitation	increased cabling complexity ⓘ requires careful routing and traffic engineering ⓘ
hasProperty	high bisection bandwidth ⓘ multi-path connectivity ⓘ non-blocking under certain traffic patterns ⓘ regular ⓘ scalable ⓘ symmetrical ⓘ
hasStructure	hierarchical ⓘ tree-like ⓘ
increasesBandwidthToward	root of the tree ⓘ
introducedBy	Charles E. Leiserson ⓘ
introducedIn	1985 ⓘ
isInspiredBy	fat-tree data structure in computer science ⓘ linked to: fat-tree network topology
oftenUses	ECMP routing ⓘ layer-3 IP routing in data centers ⓘ
originallyProposedFor	VLSI networks ⓘ
requires	routing protocols that exploit path diversity ⓘ
supports	fault tolerance via path redundancy ⓘ load balancing across multiple paths ⓘ
usedIn	data center networks ⓘ distributed systems ⓘ high-performance computing clusters ⓘ parallel computing systems ⓘ

How these facts were elicited

Referenced by (3)

Full triples — surface form annotated when it differs from this entity's canonical label.

Charles E. Leiserson → notableConcept → fat-tree network topology ⓘ

Cisco FabricPath → supportsTopology → Clos topology ⓘ

linked to: fat-tree network topology

fat-tree network topology → isInspiredBy → fat-tree data structure in computer science ⓘ