US20130205038A1 - Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network - Google Patents

Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network Download PDF

Info

Publication number
US20130205038A1
US20130205038A1 US13/366,640 US201213366640A US2013205038A1 US 20130205038 A1 US20130205038 A1 US 20130205038A1 US 201213366640 A US201213366640 A US 201213366640A US 2013205038 A1 US2013205038 A1 US 2013205038A1
Authority
US
United States
Prior art keywords
layer
network
dcb
protocol
congestion
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US13/366,640
Other languages
English (en)
Inventor
Casimer M. DeCusatis
Mircea Gusat
Ronald P. Luijten
Cyriel J.A. Minkenberg
Fredy D. Neeser
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
International Business Machines Corp
Original Assignee
International Business Machines Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by International Business Machines Corp filed Critical International Business Machines Corp
Priority to US13/366,640 priority Critical patent/US20130205038A1/en
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION reassignment INTERNATIONAL BUSINESS MACHINES CORPORATION ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: LUIJTEN, RONALD P., NEESER, FREDY D., GUSAT, MIRCEA, MINKENBERG, CYRIEL J. A., DECUSATIS, CASIMER M.
Priority to US13/708,933 priority patent/US9356867B2/en
Priority to CN201380008254.1A priority patent/CN104094559B/zh
Priority to PCT/EP2013/051169 priority patent/WO2013117427A1/en
Priority to EP13700912.2A priority patent/EP2829025A1/en
Publication of US20130205038A1 publication Critical patent/US20130205038A1/en
Abandoned legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/11—Identifying congestion
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/19—Flow control; Congestion control at layers above the network layer
    • H04L47/193—Flow control; Congestion control at layers above the network layer at the transport layer, e.g. TCP related
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L47/00—Traffic control in data switching networks
    • H04L47/10—Flow control; Congestion control
    • H04L47/26—Flow control; Congestion control using explicit feedback to the source, e.g. choke packets

Definitions

  • the invention relates to the field of computer networking, and, more particularly, to Ethernet networks.
  • Converged Enhanced Ethernet (CEE) datacenters allow high link speeds and short delays while introducing lossless operation (and lossless traffic classes) by the means of link layer flow control (LL-FC, a.k.a. Priority Flow Control (PFC) in CEE) beyond the traditional lossy operation (lossy traffic classes).
  • LL-FC link layer flow control
  • PFC Priority Flow Control
  • a reliability system for a Converged Enhanced Ethernet network may include a plurality of end points each comprising a layer 4 transport layer, where each end point is connected to a data center bridging (DCB) layer 2 network.
  • the system may also include an adaptor between the layer 4 transport layer comprising one or more protocols, such as TCP, UDP, RCP, DCCP, XCP, etc., and the DCB layer 2 network to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer.
  • protocols such as TCP, UDP, RCP, DCCP, XCP, etc.
  • the DCB layer 2 network may generate flow control signals according to a flow control protocol supporting multiple priorities, such as Priority Flow Control (PFC).
  • the DCB layer 2 network may generate congestion control feedback signals according to a quantized congestion notification (QCN) protocol.
  • PFC and QCN can be individually or simultaneously enabled in the DCB layer 2 network. If both PFC and QCN are enabled, either one or both may be independently used by any end point.
  • the end point may be connected to the DCB layer 2 network through an end station, wherein the end station implements a quantized congestion notification (QCN) reaction point imposing rate limits based on a QCN protocol to limit network congestion in the DCB layer 2 network in response to receiving congestion control signals.
  • QCN quantized congestion notification
  • the network traffic generated by the transport layer may be carried on layer 2 with lossy operation, either by configuring the end station to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol.
  • PFC priority flow control
  • the network traffic generated by the transport layer may be carried on layer 2 with lossless operation, by configuring the end station to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station react to layer 2 flow control messages generated by the adjacent switch.
  • the network traffic generated by the transport layer may be carried on layer 2 with lossy operation, either by configuring the end station to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol.
  • the network traffic generated by the transport layer may be carried on layer 2 with lossless operation, by configuring the end station to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station react to layer 2 flow control messages generated by an adjacent switch.
  • the adaptor may preprocess the flow and congestion control feedback signals into consolidated feedback signals, with the preprocessing including at least one of delaying, aggregating, filtering, replicating, enhancing and decimating the primary feedback signals.
  • the layer 4 transport layer may be a Transmission Control Protocol (TCP), RCP, XCP, DCCP, UDP or any socket-based transport scheme, herein named TCP.
  • TCP Transmission Control Protocol
  • the interface may provide a reduced-rate consolidated feedback signal indicating congestion severity induced by a TCP flow, and in which the interface comprises a TCP congestion module for controlling TCP flow transmissions in response to the consolidated feedback signal.
  • the consolidated feedback signal may comprise at least one of a TCP flow rate limit, a TCP flow buffer occupancy metric, and a TCP flow rate limit for processing TCP ACKs [if existent, as UDP doesn't employ ACK] or Explicit Congestion Notifications (ECN) and for controlling associated TCP transmissions.
  • the congestion module adjusts a TCP flow congestion window and transmission schedule in response to the consolidated feedback signal.
  • the method may include providing a plurality of end points each comprising a layer 4 transport layer, where each end point is connected to a data center bridging (DCB) layer 2 network.
  • the method may also include positioning an adaptor between the layer 4 transport layer and the DCB layer 2 network to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer.
  • DCB data center bridging
  • the method may further include generating flow control signals at the DCB layer 2 network according to a flow control protocol supporting multiple priorities, such as Priority Flow Control (PFC).
  • the method may additionally include generating congestion control feedback signals at the DCB layer 2 network according to a quantized congestion notification (QCN) protocol.
  • QCN quantized congestion notification
  • the method may also include connecting the end point to the DCB layer 2 network through an end station, where the end station implements a quantized congestion notification (QCN) reaction point imposing rate limits based on a QCN protocol to limit network congestion in the DCB layer 2 network in response to receiving congestion control signals.
  • QCN quantized congestion notification
  • the method may further include carrying network traffic generated by the transport layer on layer 2 with lossy operation, either by configuring the end station to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol.
  • PFC priority flow control
  • the method may additionally include carrying the network traffic generated by the transport layer on layer 2 with lossless operation, by configuring the end station to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station react to layer 2 flow control messages generated by an adjacent switch.
  • PFC priority flow control
  • switch we refer to any physical or virtual device that may be used for switching, bridging, steering, sorting, routing, forwarding, scheduling packets or Ethernet frames.
  • the method may also include processing TCP ACKs and/or ECNs, and controlling associated TCP transmissions where the consolidated feedback signal comprises at least one of a TCP flow rate limit, a TCP flow buffer occupancy metric, and a TCP flow rate limit.
  • the method may further include adjusting a TCP flow congestion window and transmission schedule in response to the consolidated feedback signal via the congestion module.
  • the computer readable program codes may be configured to cause the program to provide a plurality of end points each comprising a layer 4 transport layer, where each end point is connected to a data center bridging (DCB) layer 2 network.
  • the computer readable program codes may also position an adaptor between the layer 4 transport layer and the DCB layer 2 network to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer.
  • FIG. 1 is a block diagram illustrating a Converged Enhanced network in accordance with the invention.
  • FIG. 2 is a flowchart illustrating method aspects according to the invention.
  • FIG. 3 is a flowchart illustrating method aspects according to the method of FIG. 2 .
  • FIG. 4 is a flowchart illustrating method aspects according to the method of FIG. 2 .
  • FIG. 5 is a flowchart illustrating method aspects according to the method of FIG. 4 .
  • FIG. 6 is a flowchart illustrating method aspects according to the method of FIG. 4 .
  • FIG. 7 is a flowchart illustrating method aspects according to the method of FIG. 4 .
  • FIG. 8 is a flowchart illustrating method aspects according to the method of FIG. 5 .
  • FIG. 9 is a flowchart illustrating method aspects according to the method of FIG. 5 .
  • FIG. 10 illustrates a prior art hotspot saturation tree in a 5-stage fat tree.
  • FIG. 11 illustrates explicit congestion notification buffering size in the prior art.
  • FIG. 12 is a block diagram illustrating an alternative Converged Enhanced network embodiment in accordance with the invention.
  • the system 10 includes a plurality of end points 14 a - 14 n each comprising a layer 4 transport layer 16 a - 16 n , where each end point is connected to a data center bridging (DCB) layer 2 network 18 .
  • the system 10 also includes an adaptor 20 between the layer 4 transport layer 16 a - 16 n and the DCB layer 2 network 18 to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer.
  • the DCB layer 2 network 18 generates flow control signals according to a flow control protocol supporting multiple priorities, such as Priority Flow Control (PFC) and/or the like. In another embodiment, the DCB layer 2 network 18 generates congestion control feedback signals according to a quantized congestion notification (QCN) protocol.
  • PFC Priority Flow Control
  • QCN quantized congestion notification
  • the end point 14 a - 14 n is connected to the DCB layer 2 network 18 through an end station 22 , wherein the end station implements a quantized congestion notification (QCN) reaction point imposing rate limits based on a QCN protocol to limit network congestion in the DCB layer 2 network in response to receiving congestion control signals.
  • QCN quantized congestion notification
  • the network traffic generated by the transport layer 16 a - 16 n is carried on layer 2 18 with lossy operation, either by configuring the end station 22 to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol.
  • PFC priority flow control
  • the network traffic generated by the transport layer 16 a - 16 n is carried on layer 2 18 with lossless operation, by configuring the end station to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station 22 react to layer 2 flow control messages generated by the adjacent switch 24 .
  • the network traffic generated by the transport layer 16 a - 16 n is carried on layer 2 18 with lossy operation, either by configuring the end station 22 to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol.
  • PFC priority flow control
  • the network traffic generated by the transport layer 16 a - 16 n is carried on layer 2 18 with lossless operation, by configuring the end station 22 to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station react to layer 2 flow control messages generated by an adjacent switch 24 .
  • the adaptor 20 preprocess the flow and congestion control feedback signals into consolidated feedback signals, with the preprocessing including at least one of delaying, aggregating, filtering, replicating, enhancing and decimating the primary feedback signals.
  • the layer 4 transport layer 16 a - 16 n is a Transmission Control Protocol (TCP) layer.
  • the adaptor 20 provides a reduced-rate consolidated feedback signal indicating congestion severity induced by a TCP flow, and in which the adaptor comprises a TCP congestion module for controlling TCP flow transmissions in response to the consolidated feedback signal.
  • the consolidated feedback signal may comprise at least one of a TCP flow rate limit, a TCP flow buffer occupancy metric, and a TCP flow rate limit for processing TCP ACKs and for controlling associated TCP transmissions.
  • the congestion module adjusts a TCP flow congestion window and transmission schedule in response to the consolidated feedback signal.
  • the method begins at Block 34 and may include providing a plurality of end points each comprising a layer 4 transport layer, where each end point is connected to a data center bridging (DCB) layer 2 network at Block 36 .
  • the method may also include positioning an adaptor between the layer 4 transport layer and the DCB layer 2 network to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer at Block 38 .
  • the method ends at Block 40 .
  • the method begins at Block 44 .
  • the method may include the steps of FIG. 2 at Blocks 36 and 38 .
  • the method may further include generating flow control signals at the DCB layer 2 network according to a flow control protocol supporting multiple priorities, such as Priority Flow Control (PFC) at Block 46 .
  • PFC Priority Flow Control
  • the method ends at Block 48 .
  • the method begins at Block 52 .
  • the method may include the steps of FIG. 2 at Blocks 36 and 38 .
  • the method may additionally include generating congestion control feedback signals at the DCB layer 2 network according to a quantized congestion notification (QCN) protocol at Block 54 .
  • QCN quantized congestion notification
  • the method begins at Block 60 .
  • the method may include the steps of FIG. 4 at Blocks 36 , 38 , and 54 .
  • the method may also include connecting the end point to the DCB layer 2 network through an end station, where the end station implements a quantized congestion notification (QCN) reaction point imposing rate limits based on a QCN protocol to limit network congestion in the DCB layer 2 network in response to receiving congestion control signals at Block 62 .
  • QCN quantized congestion notification
  • the method ends at Block 64 .
  • the method begins at Block 68 .
  • the method may include the steps of FIG. 4 at Blocks 36 , 38 , and 54 .
  • the method may further include carrying network traffic generated by the transport layer on layer 2 with lossy operation, either by configuring the end station to steer the traffic to a lossy priority of a priority flow control (PFC) protocol, or by not using any PFC protocol at Block 70 .
  • PFC priority flow control
  • the method begins at Block 76 .
  • the method may include the steps of FIG. 4 at Blocks 36 , 38 , and 54 .
  • the method may additionally include carrying the network traffic generated by the transport layer on layer 2 with lossless operation, by configuring the end station to steer the traffic to a lossless priority of a priority flow control (PFC) protocol and by letting the end station react to layer 2 flow control messages generated by an adjacent switch at Block 78 .
  • PFC priority flow control
  • the method ends at Block 80 .
  • the method begins at Block 84 .
  • the method may include the steps of FIG. 5 at Blocks 36 , 38 , 54 , and 62 .
  • the method may also include processing TCP ACKs and ECNs and controlling associated TCP transmissions where the consolidated feedback signal comprises at least one of a TCP flow rate limit, a TCP flow buffer occupancy metric, and a TCP flow rate limit at Block 86 .
  • the method ends at Block 88 .
  • the method begins at Block 92 .
  • the method may include the steps of FIG. 5 at Blocks 36 , 38 , 54 , and 62 .
  • the method may further include adjusting a TCP flow congestion window and transmission schedule in response to the consolidated feedback signal via the congestion module at Block 94 .
  • the method ends at Block 96 .
  • the computer readable program codes may be configured to cause the program to provide a plurality of end points 14 a - 14 n each comprising a layer 4 transport layer 16 a - 16 n respectively, where each end point is connected to a data center bridging (DCB) layer 2 network 18 .
  • the computer readable program codes may also position an adaptor 20 between the layer 4 transport layer 16 a - 16 n and the DCB layer 2 network 18 to translate at least one of flow and congestion control feedback signals, provided by at least one of the DCB network and the transport layer, to consolidated feedback signals for controlling transmission by the transport layer.
  • the system 10 provides reliability in a Converged Enhanced Ethernet network.
  • CEE Converged Enhanced Ethernet
  • DCN Converged Enhanced Ethernet
  • CEE datacenters allow high link speeds and short delays while introducing lossless operation (and lossless traffic classes) by the means of link layer flow control (LL-FC, aka PFC in CEE) beyond the traditional lossy operation (lossy traffic classes).
  • LL-FC link layer flow control
  • PFC link layer flow control
  • the lossless operation of CEE introduces new challenges, such as deadlocks and saturation tree congestion. Namely, a single hotspot saturation tree congestion can cause a total DCN collapse within a few 10s-100s of us.
  • FIG. 10 illustrates the problem (hotspot congestion box). If a sufficient fraction of all the inputs' traffic targets one of the outputs (in the figure, the output labeled 128 ), that output link can saturate: it becomes a hotspot (HS) that causes the queues in the switch feeding that link to fill up. If the traffic pattern persists, then, no matter what techniques are used to reassign buffer space, it is all ultimately exhausted. This forces that switch's LL-FC to quickly throttle back all the inputs feeding that switch. That in turn causes the previous stage to fill its buffer space. In a domino effect, the congestion eventually backs up all the way to the network inputs. This has been called tree saturation or, in other contexts, high-order Head of Line (HOL) blocking congestion spreading.
  • HOL Head of Line
  • the traffic causing the hotspot will root one or more saturation trees partly caused by the inherent traffic distribution and partly by flow interference or high-order HOL blocking.
  • lossless LL-FC offers substantial performance benefits, albeit it has the drawback, besides its complexity, of facilitating saturation tree congestion.
  • lossless ICTNs such as CEE-based DCNs will be increasingly exposed to saturation trees and congestion collapse.
  • the first attempts in the CEE context were done in IEEE 802.1Qau, by using the QCN mechanism against simple (single bottleneck), yet persistent, hotspot congestion.
  • TCP transmission control protocol
  • IP Internet Protocol
  • Double feedback loop Unlike TCP in IP networks, FCC mechanisms in DCNs are based on a dual closed-loop control system: (i) LL-FC (PFC) and (ii) end-to-end CM (either QCN, or TCP, or both).
  • the former is the smaller and faster loop taking care of LL correctness and, sometimes, performance like e.g. advanced scheduling [ETS].
  • CM involves a larger and slower loop with much longer time constants than the LL RTT; a complete CM solution may include congestion avoidance/prevention and control (after it happens). Since CM is inherently slower than its underlying LL-FC loops, it needs an aggregated view of the ICTN status—whereas the LL-FC relies only on local status.
  • CM should compensate the inertia of its larger loop by (a) acquiring global view. Feedback (QCN CNM, TCP ECNs, Vegas' delays etc.) about traffic conditions and (b) elaborating a more complex source reaction that considers the outdated global view and ideally, tries to predict the traffic based on the trends acquired so far.
  • TCP does not assume the existence of a fast and lossless LL-FC layer; nor does TCP coexist well with other flow control schemes (QCN), as proven by TCP over ATM/ABR.
  • Shallow buffers The alternative would be to over-design the switch buffers beyond the size mandated for lossless ICTNs. This, however, is not practically possible (see FIG. 11 ), but also aggravates the post-congestion phase by slowing its recovery.
  • TCP and ABR were extensively studied and improved for BE networks, we still lack conclusive evidence of their applicability and sufficiency in ICTNs.
  • recent research invalidates TCP's use for certain types of middleware, as well as the TCP Incast.
  • TCP was designed in early 80s to curb single bottleneck congestion in lossy BE networks with e2e lags of 100s of ms and 10s of MB switch buffers.
  • a CEE-based DCN is lossless (hence multi-bottleneck saturation tree congestion), fast (lags of 0.5-50 us) and shallow (10-100s KB) buffers.
  • system 10 uses the following changes/enhancements to TCP, resulting in “DC-TCP”: 1) Employ a software and/or hardware version of TCP, such as (CU)BIC, Reno, Vegas, Compound etc. in the end nodes.
  • a software and/or hardware version of TCP such as (CU)BIC, Reno, Vegas, Compound etc.
  • Future CEE DCN will implement native L2 CM, i.e. QCN (see 802.1Qau in [42]). Retain the QCN congestion detection, while disabling the QCN rate limiter in the source.
  • Congestion signaling and TCP rate limiter Replace or complement the traditional TCP rate limiter based on duplicate ACKs with a hybrid rate limiter based on backward congestion notifications (BCNs) and QCN congestion notification messages (CNMs). Feed a digested form of CNMs associated with the TCP source into TCP for window control based on L2 feedback.
  • BCNs backward congestion notifications
  • CNMs QCN congestion notification messages
  • TCP constants e.g. RTO
  • Potentially adapt to changing network size and delay in real time (optional, via delay probing or Feedback Request protocol).
  • the TCP receiver will report congestion/loss in the lossy network via duplicate ACKs.
  • the CNMs sent towards the source must be appropriately translated at the boundary between the lossless and the lossy networks.
  • One possibility is to convert CNMs to TCP window scalings in the boundary switch, as CNMs will not be understood by the lossy network.
  • FIG. 12 illustrates one embodiment of system 10 .
  • System 10 adapts TCP to a lossless DCN, by combining a re-tuned TCP flavor (CUBIC, Compound and New Reno are favored, others may apply) with L2 QCN signaling.
  • CBIC re-tuned TCP flavor
  • System 10 copes with saturation trees issues, us latency, and shallow buffers.
  • System 10 also compensates and adapts to rapidly changing DCN loads.
  • System 10 provided full TCP socket compatibility, and therefore legacy application support.
  • a method for preventing the spread of packet congestion while simultaneously preventing packet loss in the network having at least one source channel adapter, at least one destination channel adapter, and multiple fiber channel over Ethernet (FCoE) enabled switches 24 is enabled by system 10 .
  • the system 10 detects congestion occurring within the data center network.
  • the system 10 measures the extent of the congestion and generates a feedback signal (value) at the Layer 2 level, notifying the source channel adapter and destination channel adapter that congestion is occurring.
  • the system 10 also compensates for that congestion by changing the packet injection rate (within a sliding window) by an amount proportional to the magnitude of the feedback signal and dynamically readjusts the feedback signal (value) based on the extent of congestion.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
US13/366,640 2012-02-06 2012-02-06 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network Abandoned US20130205038A1 (en)

Priority Applications (5)

Application Number Priority Date Filing Date Title
US13/366,640 US20130205038A1 (en) 2012-02-06 2012-02-06 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network
US13/708,933 US9356867B2 (en) 2012-02-06 2012-12-08 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network
CN201380008254.1A CN104094559B (zh) 2012-02-06 2013-01-23 用于收敛增强以太网络的可靠性的方法和系统
PCT/EP2013/051169 WO2013117427A1 (en) 2012-02-06 2013-01-23 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network
EP13700912.2A EP2829025A1 (en) 2012-02-06 2013-01-23 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US13/366,640 US20130205038A1 (en) 2012-02-06 2012-02-06 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/708,933 Continuation US9356867B2 (en) 2012-02-06 2012-12-08 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network

Publications (1)

Publication Number Publication Date
US20130205038A1 true US20130205038A1 (en) 2013-08-08

Family

ID=47598844

Family Applications (2)

Application Number Title Priority Date Filing Date
US13/366,640 Abandoned US20130205038A1 (en) 2012-02-06 2012-02-06 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network
US13/708,933 Active 2032-02-26 US9356867B2 (en) 2012-02-06 2012-12-08 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network

Family Applications After (1)

Application Number Title Priority Date Filing Date
US13/708,933 Active 2032-02-26 US9356867B2 (en) 2012-02-06 2012-12-08 Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network

Country Status (4)

Country Link
US (2) US20130205038A1 (zh)
EP (1) EP2829025A1 (zh)
CN (1) CN104094559B (zh)
WO (1) WO2013117427A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140301197A1 (en) * 2013-04-05 2014-10-09 International Business Machines Corporation Virtual quantized congestion notification
CN104980359A (zh) * 2014-04-04 2015-10-14 中兴通讯股份有限公司 以太网光纤通道的流量控制方法、装置及系统
US9325639B2 (en) 2013-12-17 2016-04-26 At&T Intellectual Property I, L.P. Hierarchical caching system for lossless network packet capture applications
US9614765B2 (en) 2014-08-26 2017-04-04 Lenovo Enterprise Solutions (Singapore) Pte. Ltd. Quantized congestion notification (QCN) proxy function in data center bridging capabilities exchange (DCBX) protocol
CN112968811A (zh) * 2021-02-20 2021-06-15 中国工商银行股份有限公司 一种rdma网络的pfc异常处理方法及装置
US11683250B2 (en) * 2021-10-22 2023-06-20 Palo Alto Networks, Inc. Managing proxy throughput between paired transport layer connections

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6015744B2 (ja) * 2012-03-23 2016-10-26 富士通株式会社 輻輳制御方法、輻輳制御装置、通信システム及び輻輳制御プログラム
KR101536141B1 (ko) * 2014-02-13 2015-07-13 현대자동차주식회사 이더넷과 can 통신 간의 신호 변환을 제공하는 차량용 장치 및 그 제어방법

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6426944B1 (en) * 1998-12-30 2002-07-30 At&T Corp Method and apparatus for controlling data messages across a fast packet network
US20100223397A1 (en) * 2009-02-27 2010-09-02 Uri Elzur Method and system for virtual machine networking

Family Cites Families (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7564869B2 (en) 2004-10-22 2009-07-21 Cisco Technology, Inc. Fibre channel over ethernet
US7961621B2 (en) 2005-10-11 2011-06-14 Cisco Technology, Inc. Methods and devices for backward congestion notification
EP1936880A1 (en) 2006-12-18 2008-06-25 British Telecommunications Public Limited Company Method and system for congestion marking
EP2182686B1 (en) * 2007-03-12 2017-12-27 Citrix Systems, Inc. Systems and methods for providing quality of service precedence in tcp congestion control
US7821939B2 (en) 2007-09-26 2010-10-26 International Business Machines Corporation Method, system, and computer program product for adaptive congestion control on virtual lanes for data center ethernet architecture
US8458305B2 (en) * 2009-08-06 2013-06-04 Broadcom Corporation Method and system for matching and repairing network configuration
US8504690B2 (en) 2009-08-07 2013-08-06 Broadcom Corporation Method and system for managing network power policy and configuration of data center bridging
JP5621996B2 (ja) 2010-02-12 2014-11-12 日本電気株式会社 ネットワークシステム及び輻輳制御方法
US20110261686A1 (en) 2010-04-21 2011-10-27 Kotha Saikrishna M Priority Pause (PFC) in Virtualized/Non-Virtualized Information Handling System Environment
US8767742B2 (en) 2010-04-22 2014-07-01 International Business Machines Corporation Network data congestion management system
US20110261696A1 (en) 2010-04-22 2011-10-27 International Business Machines Corporation Network data congestion management probe system
JP5580706B2 (ja) 2010-09-29 2014-08-27 Ｋｄｄｉ株式会社 再送制御プロトコルを用いるデータ転送装置、プログラム及び方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6426944B1 (en) * 1998-12-30 2002-07-30 At&T Corp Method and apparatus for controlling data messages across a fast packet network
US20100223397A1 (en) * 2009-02-27 2010-09-02 Uri Elzur Method and system for virtual machine networking

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20140301197A1 (en) * 2013-04-05 2014-10-09 International Business Machines Corporation Virtual quantized congestion notification
US20150295839A1 (en) * 2013-04-05 2015-10-15 International Business Machines Corporation Virtual quantized congestion notification
US9166925B2 (en) * 2013-04-05 2015-10-20 International Business Machines Corporation Virtual quantized congestion notification
US9654410B2 (en) * 2013-04-05 2017-05-16 International Business Machines Corporation Virtual quantized congestion notification
US10182016B2 (en) * 2013-04-05 2019-01-15 International Business Machines Corporation Virtual quantized congestion notification
US9325639B2 (en) 2013-12-17 2016-04-26 At&T Intellectual Property I, L.P. Hierarchical caching system for lossless network packet capture applications
US9577959B2 (en) 2013-12-17 2017-02-21 At&T Intellectual Property I, L.P. Hierarchical caching system for lossless network packet capture applications
CN104980359A (zh) * 2014-04-04 2015-10-14 中兴通讯股份有限公司 以太网光纤通道的流量控制方法、装置及系统
US9614765B2 (en) 2014-08-26 2017-04-04 Lenovo Enterprise Solutions (Singapore) Pte. Ltd. Quantized congestion notification (QCN) proxy function in data center bridging capabilities exchange (DCBX) protocol
CN112968811A (zh) * 2021-02-20 2021-06-15 中国工商银行股份有限公司 一种rdma网络的pfc异常处理方法及装置
US11683250B2 (en) * 2021-10-22 2023-06-20 Palo Alto Networks, Inc. Managing proxy throughput between paired transport layer connections

Also Published As

Publication number Publication date
WO2013117427A1 (en) 2013-08-15
CN104094559A (zh) 2014-10-08
US20130205039A1 (en) 2013-08-08
EP2829025A1 (en) 2015-01-28
CN104094559B (zh) 2016-12-14
US9356867B2 (en) 2016-05-31

Similar Documents

Publication Publication Date Title
US9356867B2 (en) Lossless socket-based layer 4 transport (reliability) system for a converged ethernet network
US9961585B2 (en) Network-side buffer management
US8831041B2 (en) Prioritizing highly compressed traffic to provide a predetermined quality of service
US9413814B2 (en) Systems and methods for providing quality of service via a flow controlled tunnel
US8379515B1 (en) TCP throughput control by imposing temporal delay
WO2020001192A1 (zh) 一种数据传输方法、计算设备、网络设备及数据传输系统
US20140281018A1 (en) Dynamic Optimization of TCP Connections
JP2018508151A (ja) 伝送制御プロトコルｔｃｐデータパケットを送信する方法及び装置、並びにシステム
US20190253364A1 (en) Method For Determining TCP Congestion Window, And Apparatus
Kühlewind et al. Using data center TCP (DCTCP) in the Internet
EP3323229A1 (en) Method and apparatus for managing network congestion
JP2006197110A (ja) 中継装置及び通信端末装置
KR20120122523A (ko) 혼잡 윈도우 크기 조절 방법 및 그에 따른 ｔｃｐ 시스템
EP3579517A1 (en) Multicast service processing method and access device
Nabeshima Performance evaluation of multcp in high-speed wide area networks
Arumaithurai et al. Nf-tcp: Network friendly tcp
Tomita et al. Data uploading time estimation for CUBIC TCP in long distance networks
Andrew et al. An example of instability in XCP
Mareev et al. Multipoint Data Transmission Issues in High Bandwidth-Delay Product TCP/IP Networks
TWI757887B (zh) 用以促進一資料流從一發送端透過多路徑傳輸至一接收端的方法、網路控制器以及電腦程式產品
Bisio et al. Performance enhanced proxy solutions for satellite networks: state of the art, protocol stack and possible interfaces
López-Pacheco et al. Enabling large data transfers on dynamic, very high-speed network infrastructures
Chen et al. The performance of TCP congestion control algorithm over high-speed transmission links
Lopez-Pacheco et al. Performance comparison of TCP, HSTCP and XCP in high-speed, highly variable-bandwidth environments
Hung et al. Simple slow-start and a fair congestion avoidance for TCP communications

Legal Events

Date Code Title Description
AS Assignment

Owner name: INTERNATIONAL BUSINESS MACHINES CORPORATION, NEW Y

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:DECUSATIS, CASIMER M.;GUSAT, MIRCEA;LUIJTEN, RONALD P.;AND OTHERS;SIGNING DATES FROM 20120116 TO 20120119;REEL/FRAME:027656/0771

STCB Information on status: application discontinuation

Free format text: ABANDONED -- FAILURE TO RESPOND TO AN OFFICE ACTION