Linux Tcp Receive Window Auto Tuning
May 05, 2016 The auto-tuning feature allows newer versions of Windows to be able to adjust the RWIN value automatically on the fly so to speak. While it does offer improvements in certain realms, it can also cause network performance issues. I wanted to detail a few scenarios of TCP Auto-tuning slow network performance and when I have observed this issue. Disabling linux tcp receiver window auto tuning. I'm working on linux (2.6.43.8-1.fc15.i686.PAE) and face some problem to disable auto tuning of the TCP receiver window. I'm working on linux (2.6.43.8-1.fc15.i686.PAE) and face some problem to disable auto tuning of the TCP receiver window. I'm trying to find a given setting that would force my client to advertise a fixed receiver window. Linux Network (TCP) Performance Tuning with Sysctl Submitted by Sarath Pillai on Sun, - 19:06 Before the Linux kernel version 1.3.57, there was no mechanism other than recompiling the kernel, if you wanted to modify some system parameters.
-->Oct 10, 2016 Tuning 10Gb NICs highway to hell If you are trying to achieve a maximum performance with 10Gb or 40Gb NICs in RHEL or similar prepare yourself to a battle. This article is for experienced users, don’t mess up with default kernel parameters if you. Oct 08, 2016 If you want a good full breakdown see: The Cable Guy TCP Receive Window Auto-Tuning Over the internet your info flows through a pipe. It does so in bunches at a time called packets. Think of it as a ball rolling through the pipe.
Applies to: Windows Server 2019, Windows Server 2016, Windows Server (Semi-Annual Channel)
Use the information in this topic to tune the performance network adapters for computers that are running Windows Server 2016 and later versions. If your network adapters provide tuning options, you can use these options to optimize network throughput and resource usage.
The correct tuning settings for your network adapters depend on the following variables:
- The network adapter and its feature set
- The type of workload that the server performs
- The server hardware and software resources
- Your performance goals for the server
The following sections describe some of your performance tuning options.
Enabling offload features
Turning on network adapter offload features is usually beneficial. However, the network adapter might not be powerful enough to handle the offload capabilities with high throughput.
Important
Do not use the offload features IPsec Task Offload or TCP Chimney Offload. These technologies are deprecated in Windows Server 2016, and might adversely affect server and networking performance. In addition, these technologies might not be supported by Microsoft in the future.
For example, consider a network adapter that has limited hardware resources.In that case, enabling segmentation offload features might reduce the maximum sustainable throughput of the adapter. However, if the reduced throughput is acceptable, you should go ahead an enable the segmentation offload features.
Note
Some network adapters require you to enable offload features independently for the send and receive paths.
Enabling receive-side scaling (RSS) for web servers
RSS can improve web scalability and performance when there are fewer network adapters than logical processors on the server. When all the web traffic is going through the RSS-capable network adapters, the server can process incoming web requests from different connections simultaneously across different CPUs.
Important
Avoid using both non-RSS network adapters and RSS-capable network adapters on the same server. Because of the load distribution logic in RSS and Hypertext Transfer Protocol (HTTP), performance might be severely degraded if a non-RSS-capable network adapter accepts web traffic on a server that has one or more RSS-capable network adapters. In this circumstance, you should use RSS-capable network adapters or disable RSS on the network adapter properties Advanced Properties tab.
To determine whether a network adapter is RSS-capable, you can view the RSS information on the network adapter properties Advanced Properties tab.
RSS Profiles and RSS Queues
The default RSS predefined profile is NUMAStatic, which differs from the default that the previous versions of Windows used. Before you start using RSS profiles, review the available profiles to understand when they are beneficial and how they apply to your network environment and hardware.
For example, if you open Task Manager and review the logical processors on your server, and they seem to be underutilized for receive traffic, you can try increasing the number of RSS queues from the default of two to the maximum that your network adapter supports. Your network adapter might have options to change the number of RSS queues as part of the driver.
Increasing network adapter resources
For network adapters that allow you to manually configure resources such as receive and send buffers, you should increase the allocated resources.
Some network adapters set their receive buffers low to conserve allocated memory from the host. The low value results in dropped packets and decreased performance. Therefore, for receive-intensive scenarios, we recommend that you increase the receive buffer value to the maximum.
Note
If a network adapter does not expose manual resource configuration, either it dynamically configures the resources, or the resources are set to a fixed value that cannot be changed.
Enabling interrupt moderation
To control interrupt moderation, some network adapters expose different interrupt moderation levels, different buffer coalescing parameters (sometimes separately for send and receive buffers), or both.
You should consider interrupt moderation for CPU-bound workloads. When using interrupt moderation, consider the trade-off between the host CPU savings and latency versus the increased host CPU savings because of more interrupts and less latency. If the network adapter does not perform interrupt moderation, but it does expose buffer coalescing, you can improve performance by increasing the number of coalesced buffers to allow more buffers per send or receive.
Performance tuning for low-latency packet processing
Many network adapters provide options to optimize operating system-induced latency. Latency is the elapsed time between the network driver processing an incoming packet and the network driver sending the packet back. This time is usually measured in microseconds. For comparison, the transmission time for packet transmissions over long distances is usually measured in milliseconds (an order of magnitude larger). This tuning will not reduce the time a packet spends in transit.
Following are some performance tuning suggestions for microsecond-sensitive networks.
Set the computer BIOS to High Performance, with C-states disabled. However, note that this is system and BIOS dependent, and some systems will provide higher performance if the operating system controls power management. You can check and adjust your power management settings from Settings or by using the powercfg command. For more information, see Powercfg Command-Line Options.
Set the operating system power management profile to High Performance System.
Note
This setting does not work properly if the system BIOS has been set to disable operating system control of power management.
Enable static offloads. For example, enable the UDP Checksums, TCP Checksums, and Send Large Offload (LSO) settings.
If the traffic is multi-streamed, such as when receiving high-volume multicast traffic, enable RSS.
Disable the Interrupt Moderation setting for network card drivers that require the lowest possible latency. Remember, this configuration can use more CPU time and it represents a tradeoff.
Handle network adapter interrupts and DPCs on a core processor that shares CPU cache with the core that is being used by the program (user thread) that is handling the packet. CPU affinity tuning can be used to direct a process to certain logical processors in conjunction with RSS configuration to accomplish this. Using the same core for the interrupt, DPC, and user mode thread exhibits worse performance as load increases because the ISR, DPC, and thread contend for the use of the core.
System management interrupts
Many hardware systems use System Management Interrupts (SMI) for a variety of maintenance functions, such as reporting error correction code (ECC) memory errors, maintaining legacy USB compatibility, controlling the fan, and managing BIOS-controlled power settings.
The SMI is the highest-priority interrupt on the system, and places the CPU in a management mode. This mode preempts all other activity while SMI runs an interrupt service routine, typically contained in BIOS.
Unfortunately, this behavior can result in latency spikes of 100 microseconds or more.
If you need to achieve the lowest latency, you should request a BIOS version from your hardware provider that reduces SMIs to the lowest degree possible. These BIOS versions are frequently referred to as 'low latency BIOS' or 'SMI free BIOS.' In some cases, it is not possible for a hardware platform to eliminate SMI activity altogether because it is used to control essential functions (for example, cooling fans).
Note
The operating system cannot control SMIs because the logical processor is running in a special maintenance mode, which prevents operating system intervention.
Performance tuning TCP
You can use the following items to tune TCP performance.
Precision auto body. At Your Service. At Precision Tune Auto Care, we work hard to keep your car on the road.Because we go the extra mile, you get a lot more of them from your car. Count on Precision Tune for complete car care.from top to bottom. Precision Tune Auto Care of Margate, Florida provides fast and affordable auto repair and maintenance. Let our certified technicians keep your vehicle safe and reliable. Precision Tune Auto Care of Tallahassee, Florida provides fast and affordable auto repair and maintenance. Let our certified technicians keep your vehicle safe and reliable. Precision Tune Auto Care of Delray Beach, Florida provides fast and affordable auto repair and maintenance. Let our certified technicians keep your vehicle safe and reliable.
TCP receive window autotuning
In Windows Vista, Windows Server 2008, and later versions of Windows, the Windows network stack uses a feature that is named TCP receive window autotuning level to negotiate the TCP receive window size. This feature can negotiate a defined receive window size for every TCP communication during the TCP Handshake.
In earlier versions of Windows, the Windows network stack used a fixed-size receive window (65,535 bytes) that limited the overall potential throughput for connections. The total achievable throughput of TCP connections could limit network usage scenarios. TCP receive window autotuning enables these scenarios to fully use the network.
For a TCP receive window that has a particular size, you can use the following equation to calculate the total throughput of a single connection.
Total achievable throughput in bytes = TCP receive window size in bytes * (1 / connection latency in seconds)
For example, for a connection that has a latency of 10 ms, the total achievable throughput is only 51 Mbps. This value is reasonable for a large corporate network infrastructure. However, by using autotuning to adjust the receive window, the connection can achieve the full line rate of a 1-Gbps connection.
Some applications define the size of the TCP receive window. If the application does not define the receive window size, the link speed determines the size as follows:
- Less than 1 megabit per second (Mbps): 8 kilobytes (KB)
- 1 Mbps to 100 Mbps: 17 KB
- 100 Mbps to 10 gigabits per second (Gbps): 64 KB
- 10 Gbps or faster: 128 KB
For example, on a computer that has a 1-Gbps network adapter installed, the window size should be 64 KB.
This feature also makes full use of other features to improve network performance. These features include the rest of the TCP options that are defined in RFC 1323. By using these features, Windows-based computers can negotiate TCP receive window sizes that are smaller but are scaled at a defined value, depending on the configuration. This behavior the sizes easier to handle for networking devices.
Note
You may experience an issue in which the network device is not compliant with the TCP window scale option, as defined in RFC 1323 and, therefore, doesn't support the scale factor. In such cases, refer to this KB 934430, Network connectivity fails when you try to use Windows Vista behind a firewall device or contact the Support team for your network device vendor.
Review and configure TCP receive window autotuning level
You can use either netsh commands or Windows PowerShell cmdlets to review or modify the TCP receive window autotuning level.
Note
Unlike in versions of Windows that pre-date Windows 10 or Windows Server 2019, you can no longer use the registry to configure the TCP receive window size. For more information about the deprecated settings, see Deprecated TCP parameters.
Note
For detailed information about the available autotuning levels, see Autotuning levels.
To use netsh to review or modify the autotuning level
To review the current settings, open a Command Prompt window and run the following command:
The output of this command should resemble the following:
To modify the setting, run the following command at the command prompt:
Note
In the preceding command, <Value> represents the new value for the auto tuning level.
For more information about this command, see Netsh commands for Interface Transmission Control Protocol.
To use Powershell to review or modify the autotuning level
To review the current settings, open a PowerShell window and run the following cmdlet.
Boot Camp gives you full access to all the power your MacBook or iMac have to offer, including all RAM, multiple processors and cores, 3D graphics, and all the ports.Using your powerful Mac for both Mac OS and Windows used to require virtual devices, and while VMs have some advantages (allowing you to run both operating systems at the same time), for many users just being able to boot into either is all they want. This means that none of your Mac content is touched at all, and Boot Camp has no effect on the operation of Mac OS.The process of running Boot Camp to install a Windows partition includes running a preparation utility on Mac OS (after backing everything up of course), creating the partition and installing Windows, and then selecting which OS to boot into every time the machine boots up. Asistente boot camp mac. To select an OS during startup you simply hold down the Option key and select the startup disk you want.
The output of this cmdlet should resemble the following.
To modify the setting, run the following cmdlet at the PowerShell command prompt.
Note
In the preceding command, <Value> represents the new value for the auto tuning level.
For more information about these cmdlets, see the following articles:
Autotuning levels
You can set receive window autotuning to any of five levels. The default level is Normal. The following table describes the levels.
Level | Hexadecimal value | Comments |
---|---|---|
Normal (default) | 0x8 (scale factor of 8) | Set the TCP receive window to grow to accommodate almost all scenarios. |
Disabled | No scale factor available | Set the TCP receive window at its default value. |
Restricted | 0x4 (scale factor of 4) | Set the TCP receive window to grow beyond its default value, but limit such growth in some scenarios. |
Highly Restricted | 0x2 (scale factor of 2) | Set the TCP receive window to grow beyond its default value, but do so very conservatively. |
Experimental | 0xE (scale factor of 14) | Set the TCP receive window to grow to accommodate extreme scenarios. |
If you use an application to capture network packets, the application should report data that resembles the following for different window autotuning level settings.
Autotuning level: Normal (default state)
Autotuning level: Disabled
Autotuning level: Restricted
Autotuning level: Highly restricted
Autotuning level: Experimental
Deprecated TCP parameters
The following registry settings from Windows Server 2003 are no longer supported, and are ignored in later versions.
- TcpWindowSize
- NumTcbTablePartitions
- MaxHashTableSize
All of these settings were located in the following registry subkey:
HKEY_LOCAL_MACHINESystemCurrentControlSetServicesTcpipParameters
Windows Filtering Platform
Windows Vista and Windows Server 2008 introduced the Windows Filtering Platform (WFP). WFP provides APIs to non-Microsoft independent software vendors (ISVs) to create packet processing filters. Examples include firewall and antivirus software.
Note
A poorly-written WFP filter can significantly decrease a server's networking performance. For more information, see Porting Packet-Processing Drivers and Apps to WFP in the Windows Dev Center.
For links to all topics in this guide, see Network Subsystem Performance Tuning.
Internet protocol suite |
---|
Application layer |
Transport layer |
Internet layer |
|
Link layer |
|
TCP tuning techniques adjust the network congestion avoidance parameters of Transmission Control Protocol (TCP) connections over high-bandwidth, high-latency networks. Well-tuned networks can perform up to 10 times faster in some cases.[1] However, blindly following instructions without understanding their real consequences can hurt performance as well.
Network and system characteristics[edit]
Bandwidth-delay product (BDP)[edit]
Bandwidth-delay product (BDP) is a term primarily used in conjunction with TCP to refer to the number of bytes necessary to fill a TCP 'path', i.e. it is equal to the maximum number of simultaneous bits in transit between the transmitter and the receiver.
High performance networks have very large BDPs. To give a practical example, two nodes communicating over a geostationary satellite link with a round-trip delay time (or round-trip time, RTT) of 0.5 seconds and a bandwidth of 10 Gbit/s can have up to 0.5×1010bits, i.e., 5 Gbit = 625 MB of unacknowledged data in flight. Despite having much lower latencies than satellite links, even terrestrial fiber links can have very high BDPs because their link capacity is so large. Operating systems and protocols designed as recently as a few years ago when networks were slower were tuned for BDPs of orders of magnitude smaller, with implications for limited achievable performance.
Buffers[edit]
The original TCP configurations supported TCP receive window sizebuffers of up to 65,535 (64 KiB - 1) bytes, which was adequate for slow links or links with small RTTs. Larger buffers are required by the high performance options described below.
Buffering is used throughout high performance network systems to handle delays in the system. In general, buffer size will need to be scaled proportionally to the amount of data 'in flight' at any time. For very high performance applications that are not sensitive to network delays, it is possible to interpose large end to end buffering delays by putting in intermediate data storage points in an end to end system, and then to use automated and scheduled non-real-time data transfers to get the data to their final endpoints.
TCP speed limits[edit]
Maximum achievable throughput for a single TCP connection is determined by different factors. One trivial limitation is the maximum bandwidth of the slowest link in the path. But there are also other, less obvious limits for TCP throughput. Bit errors can create a limitation for the connection as well as RTT.
Window size[edit]
In computer networking, RWIN (TCP Receive Window) is the amount of data that a computer can accept without acknowledging the sender. If the sender has not received acknowledgement for the first packet it sent, it will stop and wait and if this wait exceeds a certain limit, it may even retransmit. This is how TCP achieves reliable data transmission.
Even if there is no packet loss in the network, windowing can limit throughput. Because TCP transmits data up to the window size before waiting for the acknowledgements, the full bandwidth of the network may not always get used. The limitation caused by window size can be calculated as follows:
Linux Tcp Window Size
where RWIN is the TCP Receive Window and RTT is the round-trip time for the path.
At any given time, the window advertised by the receive side of TCP corresponds to the amount of free receive memory it has allocated for this connection. Otherwise it would risk dropping received packets due to lack of space.
The sending side should also allocate the same amount of memory as the receive side for good performance. That is because, even after data has been sent on the network, the sending side must hold it in memory until it has been acknowledged as successfully received, just in case it would have to be retransmitted. If the receiver is far away, acknowledgments will take a long time to arrive. If the send memory is small, it can saturate and block emission. A simple computation gives the same optimal send memory size as for the receive memory size given above.
Packet loss[edit]
When packet loss occurs in the network, an additional limit is imposed on the connection.[2] In the case of light to moderate packet loss when the TCP rate is limited by the congestion avoidance algorithm, the limit can be calculated according to the formula (Mathis, et al.):
where MSS is the maximum segment size and Ploss is the probability of packet loss. If packet loss is so rare that the TCP window becomes regularly fully extended, this formula doesn't apply.
TCP options for high performance[edit]
A number of extensions have been made to TCP over the years to increase its performance over fast high-RTT links ('long fat networks' or LFNs).
TCP timestamps (RFC 1323) play a double role: they avoid ambiguities due to the 32-bit sequence number field wrapping around, and they allow more precise RTT estimation in the presence of multiple losses per RTT. With those improvements, it becomes reasonable to increase the TCP window beyond 64 kB, which can be done using the window scaling option (RFC 1323).
The TCP selective acknowledgment option (SACK, RFC 2018) allows a TCP receiver to precisely inform the TCP sender about which segments have been lost. This increases performance on high-RTT links, when multiple losses per window are possible.
Path MTU Discovery avoids the need for in-network fragmentation, increasing the performance in the presence of packet loss.
See also[edit]
References[edit]
- ^'High Performance SSH/SCP - HPN-SSH'. Psc.edu. Retrieved January 23, 2020.
- ^'The Macroscopic Behavior of the TCP Congestion Avoidance Algorithm'. Psc.edu. Archived from the original on May 11, 2012. Retrieved January 3, 2017.
External links[edit]
- RFC 1323 - TCP Extensions for High Performance
- RFC 2018 - TCP Selective Acknowledgment Options
- RFC 2582 - The NewReno Modification to TCP's Fast Recovery Algorithm
- RFC 2488 - Enhancing TCP Over Satellite Channels using Standard Mechanisms
- RFC 2883 - An Extension to the Selective Acknowledgment (SACK) Option for TCP
- RFC 3517 - A Conservative Selective Acknowledgment-based Loss Recovery Algorithm for TCP
- RFC 4138 - Forward RTO-Recovery (F-RTO): An Algorithm for Detecting Spurious Retransmission Timeouts with TCP and the Stream Control Transmission Protocol (SCTP)
- TCP Tuning Guide, ESnet
- DrTCP - a utility for Microsoft Windows (prior to Vista) which can quickly alter TCP performance parameters in the registry.
- Information on 'Tweaking' your TCP stack, Broadband Reports
- TCP/IP Analyzer, speedguide.net
- NTTTCP Network Performance Test Tool, Microsoft Windows Server Performance Team Blog
- Best Practices for TCP Optimization - ExtraHop