Reference

Glossary

Every technology the articles mention, in plain words, with the standard or vendor document that defines it. In an article, click any dotted-underlined term to see its definition.

8

800 V DC

Distributing power as 800 V direct current from the facility to the rack instead of AC plus 54 V busbars: less current, less copper, fewer conversion steps. NVIDIA plans it for 1 MW racks from 2027.

Defined in NVIDIA technical blog.

A

Adaptive routing

Switches choose a packet's next hop by current load rather than a fixed hash. Used in InfiniBand and in AI Ethernet fabrics such as Spectrum-X.

See also: ECMP, Packet spraying

ADD-PATH

A BGP extension that lets a router advertise several paths for the same prefix, each with a path identifier, instead of only its best one.

Defined in RFC 7911. See also: Route reflector, ECMP

AI factory

A data centre built to run AI training and inference as one large machine, judged on tokens or trained models produced per watt and per dollar. NVIDIA's name for it.

See also: Back-end network, PUE

Aliasing

Remote leaves load-balance to every leaf of an Ethernet segment, even to a leaf that never learned the host's MAC, using the per-EVI type-1 routes.

Defined in RFC 7432 section 8.4. See also: Ethernet segment

All-reduce

Every GPU contributes a buffer and every GPU receives the combined result (for example the sum of gradients). The main collective of data-parallel training.

See also: Collective, SHARP

Anycast gateway

Every leaf answers for the subnet's default gateway with the same IP and MAC, so hosts route at their first hop wherever they are, and keep the gateway when they move. Cumulus calls it VRR.

See also: Symmetric IRB, MAC Mobility

Anycast RP

Several routers share one RP address; each source and receiver uses the nearest, and the RPs share source information with MSDP or PIM so an RP failure goes unnoticed.

Defined in RFC 4610, RFC 3446. See also: RP, MSDP

ARP suppression

A leaf answers ARP requests itself from the IP-to-MAC bindings it learned through EVPN, instead of flooding them across the fabric. It cannot answer for a host it has never learned.

Defined in RFC 9161. See also: EVPN, Ingress replication

ASM

Any-Source Multicast: receivers ask for a group without naming its sources, so the network needs an RP to connect them.

Defined in RFC 7761. See also: RP, SSM, SPT

ASN

Autonomous System Number: the identifier of a BGP domain. Fabrics use private ranges (64512-65534, or 4200000000-4294967294 for 4-byte ASNs).

Defined in RFC 6996. See also: eBGP

B

Back-end network

The scale-out fabric that connects every GPU's own NIC (400G or 800G) between racks and carries the training collectives. Usually RoCEv2 or InfiniBand, built 1:1.

See also: Front-end network, Scale-up, RoCEv2

Best-external

A backup border keeps advertising its best eBGP path to iBGP peers even while it prefers a path learned over iBGP, so the rest of the network knows the backup exit before the primary fails.

Defined in draft-ietf-idr-best-external. See also: BGP PIC, ADD-PATH

BFD

Bidirectional Forwarding Detection: a light hello protocol that detects a dead path in milliseconds and tells BGP, instead of waiting for BGP's own hold timer.

Defined in RFC 5880. See also: eBGP

BGP PIC

Prefix Independent Convergence: a router installs a backup path in its forwarding table next to the best one, so when the best next hop disappears it switches once for every prefix behind it, without waiting for BGP.

See also: Best-external, ADD-PATH

BGP unnumbered

BGP sessions over IPv6 link-local addresses, discovered with router advertisements, carrying IPv4 routes with an IPv6 next hop. No IPv4 address on any fabric link.

Defined in RFC 8950. See also: eBGP

BIDIR

Bidirectional PIM: one (*,G) tree per group carries traffic both up towards the RP and down to receivers, with no per-source state. A designated forwarder per link keeps it loop-free.

Defined in RFC 5015. See also: PIM, RP

BMC

Baseboard Management Controller: a small computer on the server board for remote power, console and sensors, reached over the out-of-band management network.

Bus bandwidth

The nccl-tests figure that makes collective results comparable with link speed: algorithm bandwidth (size / time) times an operation factor, 2(n-1)/n for all-reduce.

Defined in nccl-tests performance notes. See also: All-reduce, NCCL

C

CDU

Coolant Distribution Unit: pumps the rack-side coolant loop and exchanges its heat with the facility water, keeping the two loops separate. Sits in the rack or in the row.

See also: Direct-to-chip cooling

Checkpoint

A saved copy of the model state during training, so a job can restart from it after a failure. Writing one is a large burst of storage traffic.

See also: Parallel file system

Clos

The leaf-spine topology: every leaf connects to every spine, so any two servers are the same number of hops apart and capacity grows by adding spines.

Defined in RFC 7938. See also: Oversubscription, Rail-optimized

Cluster ID

The identifier of a route-reflection cluster, by default the RR's router ID. Each RR adds it to a route's CLUSTER_LIST when reflecting, and drops routes whose CLUSTER_LIST already holds it.

Defined in RFC 4456. See also: Route reflector, ORIGINATOR_ID

CNP

Congestion Notification Packet: sent by a RoCEv2 receiver back to the sender when it gets ECN-marked packets; the sender's NIC reduces its rate.

See also: DCQCN, ECN

Collective

An operation where many GPUs exchange data together, such as all-reduce, all-gather or all-to-all. Training waits for the slowest participant.

See also: All-reduce, NCCL

CPO

Co-packaged optics: the optical engines sit on the switch package instead of in pluggable modules, cutting power per port (NVIDIA claims about 9 W against 30 W) at the cost of field replaceability. Lasers stay in separate replaceable units.

Defined in NVIDIA technical blog.

Credit-based flow control

The receiver tells the sender how much buffer it has (credits); the sender transmits only that much. Lossless by design, as in InfiniBand. Ultra Ethernet defines an optional Ethernet version (CBFC).

See also: InfiniBand, PFC

D

DCQCN

The congestion control of RoCEv2 NICs: switches mark with ECN, the receiving NIC returns a CNP, the sending NIC cuts its rate and then recovers. Implemented in NIC hardware.

Defined in Zhu et al., SIGCOMM 2015. See also: ECN, CNP, PFC

DCTCP

Data Center TCP: TCP that reduces its window in proportion to the fraction of ECN-marked packets, keeping switch queues short.

Defined in RFC 8257. See also: ECN, DCQCN

Designated forwarder

The one leaf of an Ethernet segment elected to deliver flooded traffic to the multihomed device, so it receives one copy, not one per leaf.

Defined in RFC 7432, RFC 9785. See also: Ethernet segment, Local bias

Direct-to-chip cooling

Liquid flows through cold plates mounted on the GPUs and CPUs, taking most of the heat away; fans cool what is left. The standard for 100 kW+ GPU racks.

See also: CDU, Immersion cooling

DPU

Data Processing Unit: a NIC with its own processors that offloads networking, storage and security from the host (NVIDIA BlueField). A SuperNIC is a NIC built for the AI back-end network.

See also: RDMA

DSCP

Differentiated Services Code Point: six bits in the IP header that classify traffic. Lossless RoCE is usually mapped to a priority by DSCP (NVIDIA's default is 26).

See also: PFC, ECN

E

eBGP

External BGP runs between different autonomous systems; internal BGP (iBGP) inside one. Data center fabrics usually run eBGP between every leaf and spine.

Defined in RFC 7938. See also: ASN, BGP unnumbered

ECMP

Equal-Cost Multi-Path: a switch spreads flows over several equal next hops, usually by hashing packet headers, so each flow stays on one path.

See also: Clos, Packet spraying

ECN

Explicit Congestion Notification: a switch marks packets (two bits in the IP header) instead of dropping them when a queue grows, so senders slow down before buffers overflow.

Defined in RFC 3168. See also: DCQCN, CNP, WRED

Ethernet segment

The set of links connecting one device to several leaves. Its 10-byte identifier, the ESI, is shared by every leaf on the segment; that is how EVPN multihoming replaces MLAG.

All-active multihoming uses type-1 routes (mass withdrawal and aliasing) and type-4 routes (DF election). Split horizon over VXLAN uses local bias.

Defined in RFC 7432, RFC 8365. See also: Designated forwarder, Aliasing, Local bias, MLAG

EVPN

Ethernet VPN: a BGP address family that advertises where MAC addresses, IP addresses and subnets live, so switches learn them from BGP instead of by flooding.

Used as the control plane for VXLAN fabrics. Its route types: type-1 (Ethernet auto-discovery, for multihoming), type-2 (MAC/IP), type-3 (flood list), type-4 (Ethernet segment), type-5 (IP prefix).

Defined in RFC 7432, RFC 8365. See also: VXLAN, Route target, Ethernet segment

EVPN route types

Type-1 Ethernet auto-discovery (multihoming), type-2 MAC/IP advertisement, type-3 inclusive multicast (the flood list), type-4 Ethernet segment (DF election), type-5 IP prefix.

Defined in RFC 7432, RFC 9136. See also: EVPN

F

Front-end network

The network joining GPU servers to users, APIs, orchestration and the rest of the estate. Looks like a normal data-centre network, but inference and data loading can push it hard.

See also: Back-end network, DPU

Full bisection bandwidth

Any half of the hosts can send to the other half at full line rate, as far as capacity goes: a 1:1 fabric. With per-flow ECMP, large flows can still collide on one link, so it is a capacity property, not a guarantee of no blocking.

See also: Oversubscription, ECMP

G

Glean

When a router has a route to a subnet but no ARP entry for the host, it resolves ARP first and typically drops the packet that triggered it. The first packet to a silent host is lost.

GPU

Graphics Processing Unit: thousands of simple cores running the same operation over large data in parallel; the engine of AI training.

See also: TPU, HBM

GPUDirect RDMA

The NIC reads and writes GPU memory directly over PCIe, without staging data in CPU memory.

See also: RDMA

GPUDirect Storage

Moves data between storage and GPU memory directly, through the NIC or NVMe, without a copy through CPU memory.

Defined in NVIDIA GPUDirect Storage. See also: GPUDirect RDMA

H

HBM

High Bandwidth Memory: stacked memory beside a GPU or TPU, delivering terabytes per second.

See also: GPU, TPU

Head-of-line blocking

Traffic stuck behind a blocked packet or a paused queue, although its own path is free. PFC causes it because it pauses a whole priority, not one flow.

See also: PFC

Headroom

Buffer a switch keeps per lossless port and priority to absorb the data still in flight after it sends a PFC pause. Depends on link speed, cable length and MTU.

See also: PFC

Hold timer

How long a BGP speaker waits without hearing from a peer before declaring the session down and removing its routes. 180 seconds by default on IOS-XE and IOS-XR, with keepalives every 60.

Defined in RFC 4271. See also: BFD

Hot-potato routing

Hand traffic to another network at the nearest exit: with otherwise equal routes, BGP prefers the lowest IGP cost to the next hop.

See also: Optimal Route Reflection

I

iBGP full mesh

Every iBGP speaker in an AS peers with every other, because a router does not pass iBGP-learned routes to other iBGP peers. n routers need n(n-1)/2 sessions.

Defined in RFC 4271. See also: Route reflector

IGMP

Internet Group Management Protocol: how hosts tell the router on their LAN which multicast groups (and, in IGMPv3, which sources) they want. A querier asks periodically; silence means nobody is left.

Defined in RFC 3376 (IGMPv3), RFC 2236 (IGMPv2). See also: IGMP snooping, PIM, SSM

IGMP snooping

A switch listens to IGMP and forwards each multicast group only to the ports that asked for it, plus multicast router ports. Memberships expire unless something keeps querying.

Defined in RFC 4541. See also: IGMP

Immersion cooling

Whole servers sit in a non-conductive liquid that absorbs the heat. Works, but needs purpose-built servers and handling; direct-to-chip is more common for GPU racks.

See also: Direct-to-chip cooling

Incast

Many senders transmitting to one receiver at the same moment, overflowing the receiver's switch port. Typical of collective operations and storage reads.

See also: ECN, PFC, Oversubscription

InfiniBand

A lossless network designed for RDMA: credit-based flow control at the link layer, and a central Subnet Manager that assigns addresses and computes routes.

Defined in InfiniBand Trade Association. See also: Subnet Manager, LID, Credit-based flow control

Ingress replication

Flooded traffic (broadcast, unknown unicast, multicast) is copied by the sending VTEP, one unicast VXLAN copy per remote VTEP in the segment, instead of using IP multicast in the underlay.

Defined in RFC 8365. See also: VTEP, EVPN route types

IS-IS

Intermediate System to Intermediate System: a link-state IGP. Every router floods its links and computes shortest paths to every other; common in large enterprise and service-provider cores.

Defined in ISO/IEC 10589, RFC 1195. See also: SR-MPLS

K

KV cache

The keys and values a model has computed for every token of a conversation so far: its working memory during inference. It grows with context length and can be moved between GPUs or offloaded to CPU memory and storage.

See also: Prefill and decode

L

LACP

Link Aggregation Control Protocol: negotiates which links form one bond (802.3ad). A link is removed after three missed messages: 3 s at the fast rate, 90 s at the slow rate. Defined in IEEE 802.1AX.

See also: MLAG, Ethernet segment

LID

Local Identifier: the 16-bit address the Subnet Manager gives each InfiniBand port; switches forward on it.

See also: Subnet Manager

LLR

Link Level Retry: the link retransmits a corrupted frame itself, hop by hop, so the transport rarely sees the loss. Optional in Ultra Ethernet.

See also: Ultra Ethernet

Local bias

Split horizon for EVPN over VXLAN: the leaf where flooded traffic enters delivers it to its own segments; peers drop copies that came from a leaf sharing the same segment.

Defined in RFC 8365 section 8.3.1, RFC 9746. See also: Designated forwarder, Ethernet segment

M

MAC Mobility

An EVPN extended community with a sequence number. When a host moves, its new leaf advertises a higher number, and every switch prefers the newest location.

The same mechanism detects duplicates: by default RFC 7432 flags a MAC that moves 5 times in 180 seconds.

Defined in RFC 7432 section 15. See also: EVPN

MLAG

Multi-chassis link aggregation: two switches present themselves as one LACP partner, synchronised over a peer link with a vendor-specific protocol. EVPN multihoming is the standards-based alternative.

See also: LACP, Ethernet segment

MSDP

Multicast Source Discovery Protocol: RPs tell each other about active sources over TCP. Used between PIM domains and for anycast RP.

Defined in RFC 3618. See also: Anycast RP, RP

MTU

Maximum Transmission Unit: the largest packet a link carries. VXLAN adds 50 bytes, so a 1500-byte host packet needs 1550 on every underlay link. Path MTU discovery (PMTUD) relies on ICMP "fragmentation needed" messages.

See also: VXLAN

multipath-relax

A BGP option that lets paths with different AS paths of the same length share traffic (ECMP). Without it, BGP only balances over paths with identical AS paths.

See also: ECMP, ADD-PATH

N

NCCL

The NVIDIA Collective Communications Library: implements collectives over NVLink, PCIe, InfiniBand and RoCE, choosing algorithms such as ring and tree.

Defined in NVIDIA NCCL. See also: Collective

NSCC

Network-Signal-based Congestion Control: Ultra Ethernet's mandatory sender-side congestion control, driven by ECN and delay. RCCC is an optional receiver-credit-based scheme.

Defined in Ultra Ethernet Consortium. See also: Ultra Ethernet

NVLink

NVIDIA's GPU-to-GPU interconnect inside a server or rack (the scale-up network), much faster than the network NICs. NVSwitch connects many GPUs at full NVLink speed.

See also: Scale-up, GPUDirect RDMA

O

OCS

Optical circuit switch: mirrors redirect light between fibres to rewire connections without converting to electrical signals. Google uses them to reconfigure TPU pods.

See also: TPU

Optimal Route Reflection

A route reflector computes each client's best path from that client's position in the IGP topology, not its own, so clients keep their nearest exit. Needs the link-state database on the RR.

Defined in RFC 9107. See also: Route reflector, Hot-potato routing

ORIGINATOR_ID

Set by the first route reflector to the router ID of the route's source in the AS. A router that receives a route carrying its own router ID as ORIGINATOR_ID discards it.

Defined in RFC 4456. See also: Cluster ID

Oversubscription

The ratio between bandwidth that can enter a switch tier and bandwidth that can leave it. 1:1 gives full bisection bandwidth, though per-flow hashing can still put two flows on one link, so it is not strictly non-blocking. GPU back-end fabrics are usually built 1:1.

See also: Clos, Incast

P

Packet spraying

Sending the packets of one flow over many paths instead of one, for even load. It needs a receiver that accepts packets out of order.

See also: ECMP, Ultra Ethernet

Packet trimming

A congested switch cuts a packet down to its header and forwards that instead of dropping it silently, so the receiver learns at once what to request again.

See also: Ultra Ethernet

Parallel file system

A file system that spreads files over many storage servers so thousands of clients can read and write at once at very high total bandwidth. Lustre and GPFS are examples.

See also: GPUDirect Storage

Path hunting

After a failure, BGP routers try and withdraw successive alternative paths before converging, multiplying updates. One shared spine ASN avoids it.

Defined in RFC 7938 section 5.1. See also: ASN

PFC

Priority Flow Control (IEEE 802.1Qbb): a switch tells its neighbour to pause one traffic priority when its buffer fills, so that priority is not dropped. It can spread congestion backwards (pause storms, head-of-line blocking).

See also: ECN, Headroom, Head-of-line blocking

PFC storm

A broken endpoint, such as a hung host or a NIC whose receive ring stopped draining, sends pause frames without end and freezes a priority on its link and beyond. Different from congestion spreading, which is PFC working as designed. The PFC watchdog exists to break it.

See also: PFC

PIM

Protocol Independent Multicast: routers build multicast distribution trees by sending joins towards the source or the RP, using the unicast routing table for direction (RPF).

Defined in RFC 7761. See also: RPF, RP, SSM, BIDIR

Prefill and decode

The two phases of serving a model: prefill reads the whole prompt (compute-heavy), decode produces the answer one token at a time (memory-bound). Disaggregated inference runs them on separate GPUs.

Defined in NVIDIA Dynamo. See also: KV cache

PUE

Power Usage Effectiveness: total facility power divided by IT power. 1.0 would mean no overhead at all; large modern sites run at about 1.1 to 1.2.

Defined in ISO/IEC 30134-2. See also: CDU

PXN

PCI x NVLink: NCCL moves data over NVLink to the GPU whose NIC is on the destination's rail, then sends it from there, keeping inter-node traffic rail-local. Introduced in NCCL 2.12.

Defined in NVIDIA NCCL 2.12 blog. See also: NCCL, Rail-optimized

Q

Queue pair

The RDMA connection endpoint: a send queue and a receive queue on the NIC. Applications post work to them; the NIC does the transfer.

See also: RDMA

R

Rail-optimized

A GPU cluster design where GPU n of every server connects to leaf (rail) n, so collective traffic between same-numbered GPUs stays one hop away.

See also: Clos, Scalable unit

RDMA

Remote Direct Memory Access: a NIC reads and writes another machine's memory directly, without the operating system on the data path. The basis of GPU-to-GPU networking.

See also: RoCEv2, InfiniBand, Queue pair

RoCEv2

RDMA over Converged Ethernet version 2: the InfiniBand transport carried in UDP (port 4791) over routed Ethernet. It performs best on a network that avoids drops.

Defined in IBTA RoCEv2 annex. See also: RDMA, PFC, ECN, DCQCN

Route distinguisher

An 8-byte prefix that makes routes from different VRFs or switches unique in BGP. It identifies a route; it does not decide who imports it (the route target does).

Defined in RFC 4364. See also: Route target

Route reflector

An iBGP router that passes routes between its clients so they do not need a full mesh. It advertises only its best path per prefix unless ADD-PATH is used.

Defined in RFC 4456. See also: ADD-PATH

Route target

A BGP extended community that decides which switches import a route into which VRF or VNI. Both sides must agree on the value.

Automatic route targets are derived from the local ASN and the VNI. With one ASN per leaf they differ on every leaf: FRR and Cumulus import them as a wildcard (*:VNI), other implementations need an exact match.

Defined in RFC 4364, RFC 8365. See also: Route distinguisher, EVPN

Router MAC

An EVPN extended community carrying the MAC address a leaf uses for routed VXLAN traffic. Remote leaves use it as the inner destination MAC with symmetric IRB.

Defined in RFC 9135. See also: Symmetric IRB

RP

Rendezvous point: in ASM, the router where receivers and sources meet. Receivers join a shared tree towards it; sources register with it.

Defined in RFC 7761. See also: ASM, Anycast RP, SPT

RPF

Reverse path forwarding: a router accepts a multicast packet only on the interface it would use to reach the packet's source, which prevents loops. Joins are sent the same way, towards the source.

Defined in RFC 7761. See also: PIM

S

Scalable unit

A repeatable building block of a GPU cluster: a fixed number of servers with their leaves, deployed and tested as a unit (NVIDIA reference architectures use 32 servers).

See also: Rail-optimized

Scale-up

Scale-up: GPUs joined by NVLink inside a server or rack, behaving like one large GPU. Scale-out: many servers joined by the back-end network.

See also: NVLink, Clos

SHARP

Scalable Hierarchical Aggregation and Reduction Protocol: InfiniBand switches perform part of a reduction themselves, halving the data GPUs must send.

See also: All-reduce, InfiniBand

SPT

Shortest-path tree: the (S,G) tree from a source to its receivers along the unicast shortest path. In ASM, last-hop routers normally switch to it from the shared tree once they learn the source.

Defined in RFC 7761. See also: ASM, RP

SR-MPLS

Segment Routing over MPLS: the IGP advertises a label (a prefix SID) for each router's loopback, so traffic can be label-switched to any router without LDP. Lets the core forward without BGP routes.

Defined in RFC 8660. See also: IS-IS

SSM

Source-Specific Multicast: receivers ask for (source, group), so the tree goes straight to the source with no RP. 232.0.0.0/8 is reserved for it; hosts use IGMPv3.

Defined in RFC 4607. See also: ASM, IGMP

STP edge port

A switch port marked as connecting to a host forwards immediately instead of going through spanning-tree learning. BPDU guard shuts it down if a switch appears on it.

Straggler

The slowest participant in a synchronous step. Every other GPU waits for it, so with N ranks the step time follows roughly the (1 - 1/N) quantile of per-rank latency.

See also: Collective

Subnet Manager

The InfiniBand control plane: discovers the fabric, assigns LIDs and programs every switch's forwarding tables (OpenSM, or NVIDIA UFM).

See also: InfiniBand, LID

Symmetric IRB

Integrated Routing and Bridging where both the ingress and egress leaf route: the packet crosses the fabric in the tenant L3 VNI, from one router MAC to the other.

Defined in RFC 9135. See also: VNI, Anycast gateway, Router MAC

T

TCAM

Ternary content-addressable memory: the fast lookup memory in a switch ASIC used for ACLs, QoS and features such as ARP suppression. It is finite and often has to be carved between features.

TPU

Google's Tensor Processing Unit: built around matrix-multiply units (MXUs), systolic arrays of 128x128, or 256x256 from v6e. TPU pods connect chips directly in a torus.

See also: GPU, OCS

U

UALink

Ultra Accelerator Link: an open scale-up interconnect for joining accelerators in a pod, an alternative to NVLink. AMD's Helios first carries it over Ethernet.

Defined in UALink Consortium. See also: Scale-up, NVLink

Ultra Ethernet

An Ethernet transport and set of link features for AI and HPC from the Ultra Ethernet Consortium: multipath packet spraying, out-of-order delivery, and new congestion control (NSCC, optional RCCC).

Defined in Ultra Ethernet Consortium. See also: NSCC, Packet spraying

V

VNI

VXLAN Network Identifier: the 24-bit segment number in the VXLAN header. An L2 VNI carries one VLAN; an L3 VNI carries routed traffic for one tenant VRF.

With symmetric IRB, every leaf of a tenant shares one L3 VNI. Switches attach it to a local "transit" VLAN that has no ports; only the VNI travels on the wire.

Defined in RFC 7348. See also: VXLAN, Symmetric IRB, VRF

VRF

Virtual Routing and Forwarding: a separate routing table on a switch, one per tenant, so tenants can reuse addresses without seeing each other.

See also: VNI, Route target

VTEP

VXLAN Tunnel End Point: the switch (or host) that puts frames into VXLAN and takes them out. Usually identified by a loopback address on each leaf.

Defined in RFC 7348. See also: VXLAN, EVPN

VXLAN

Virtual eXtensible LAN: carries an Ethernet frame inside a UDP packet (port 4789) between two switches, adding 50 bytes, so a Layer 2 segment can cross a routed network.

The header carries a 24-bit VNI, so up to about 16 million segments instead of 4094 VLANs. The outer IP header goes from one VTEP to another; the network in between only routes IP.

Defined in RFC 7348. See also: VTEP, VNI, EVPN, MTU

W

WRED

(Weighted) Random Early Detection: a queue starts marking (with ECN) or dropping packets with rising probability between a minimum and a maximum threshold.

See also: ECN