Method and apparatus for establishing IEEE 1588 clock synchronization across a network element comprising first and second cooperating smart interface converters wrapping the network element
Apparatus for making legacy network elements transparent to IEEE 1588 Precision Time Protocol operation. Network elements are wrapped by device(s) capable of providing either transparent clock or boundary clock operation. In one embodiment, smart interface converters are used to provide transparent clock or boundary clock operation. The smart interface converters work cooperatively.
TECHNICAL FIELD
Embodiments in accordance with the present invention relate to time synchronization in computer networks, and more particularly, to IEEE 1588 time synchronization across network switches, hubs, and routers.
BACKGROUND
Achieving time synchronization in computer networking is still a challenging task. Different elements of the network, such as PCs, workstations, servers, and devices such as routers and switches, and even VoIP phones today use NTP (Network Time Protocol) to get and set their epoch (time) from network-based NTP servers. NTP provides time accuracy on the order of tens of milliseconds. In most cases, this is sufficient.
In areas such as test and process automation, requirements are more strict, requiring time synchronization between networked devices on the order of sub-microsecond accuracy.
This may be accomplished by using GPS-based solutions. However, even if the cost of GPS-based equipment at each node is acceptable, delivery of GPS signals to each location is often not practical, due to the cost of cabling, or due to the requirement of an antenna with a clear view of the necessary number of GPS satellites. When used, though, GPS can provide time synchronization on the order of 50 nanoseconds.
In test and measurement applications, such as measuring one-way packet delays in networks, accuracy on the order of sub-milliseconds is required. Deploying GPS at various points in a network to perform the required network measurements is economically and practically feasible. When it could be envisioned that a region of a computer network might have a few GPS receivers, it would not be practical to equip all network elements with direct access to GPS accuracy.
In recent years the process automation industry has started deploying the IEEE 1588 standard (IEEE 1588-2002, incorporated herein by reference and referred to as 1588) to perform time synchronization between various devices. Unfortunately, in its current form 1588 is limited to very small networks, typically a LAN segment. IEEE 1588 makes the assumption that delays introduced by network elements are symmetrical and uniform. In order for IEEE 1588 to be deployed across wide area networks (WANs), significant investment is required in IEEE 1588 compliant network elements such as switches, hubs, routers, and the like.
There are two methods known in IEEE 1588 to deal with varying and asymmetrical delays at the network switching and routing elements. These delays occur naturally due to the nature of packet-switched traffic. As packets arrive at ports of a network element such as a router or switch, they must be identified and queued to the proper destination. This process introduces variable delays, particularly if multiple packets are queued for the same destination. One method of reducing these delays is to use a boundary clock, and another is to use a transparent clock.
In examining the boundary clock solution, first consider an IEEE 1588 master clock in device A directly connected to an IEEE 1588 slave clock in device B. IEEE 1588 precision time protocol (PTP) messages pass between A and B, establishing and maintaining B's clock in synchrony with A. Introducing a network element such as a router or switch between A and B introduces additional, varying delays in the propagation of IEEE 1588 messages. With a boundary clock, the network element (usually a switch or a router) runs the IEEE 1588 PTP protocol, and is synchronized to an attached master clock, as an example, device A. The boundary clock in turn acts as a master clock to all attached slaves, as an example, device B. Boundary clocks do not pass IEEE 1588 PTP messages.
The boundary clock solution requires the network element at the MAC/PHY level to redirect IEEE 1588 signals to the built-in boundary clock. If multiple network elements are present in the path between devices A and B, all these network elements must provide boundary clock capabilities.
A transparent clock passes IEEE 1588 messages, but either intercepts and modifies some of these messages to account for delays introduced by the network element, or sends follow-up messages containing IEEE 1588 packet delay information. One implementation of a transparent switch calculates how much time a SYNC packet spends in the network element, and then modifies the timestamp of the associated FOLLOW UP packet to account for this delay; similarly, the switch calculates how much time the DELAY REQ packet spends in the network, and modifies the timestamp of the associated DELAY RES packet. The transparent clock functionality must be built into the network element.
Such solutions become more complex when multiple network elements may intervene between an IEEE 1588 master clock and IEEE 1588 slave clocks.
SUMMARY OF THE INVENTION
Delays to IEEE 1588 time protocol packets caused by network elements are compensated for by an external device that wraps the legacy network elements with cooperating smart interface converters.
In a first embodiment, a transparent clock is provided. The smart interface converters timestamp certain IEEE 1588 packets, and either modify the timestamps in IEEE 1588 packets, or send calculated delays as part of IEEE 1588 packets. The smart interface converter connected to the IEEE 1588 master clock may support more than one smart interface converter slave. An out of band link may be used to synchronize smart interface converter clocks.
In a second embodiment, a boundary clock is formed by connecting smart interface converters to legacy network elements. One smart interface converter paired with an IEEE 1588 master clock acts as a IEEE 1588 slave. One or more additional smart interface converters act as IEEE 1588 master clocks supporting IEEE 1588 slave clocks. In some implementations, the 1588 slave smart interface converter sends synchronization pulses to the IEEE 1588 master smart interface converters using an out of band link, and periodically sends EPOCH packets to maintain synchronized clocks.
DETAILED DESCRIPTION OF THE EMBODIMENTS
Network elements such as switches, routers, and the like introduce variable and unpredictable delays in packet traffic, including IEEE 1588 precision time protocol (PTP) traffic. The present invention provides network independent devices which wrap legacy network elements and provide IEEE 1588 network element delay transparency. This transparency may be provided in a distributed fashion by wrapping the network elements with cooperating smart interface converters (hereafter SIC).
Standard interface converters are pluggable building blocks used in many networking devices such as switches, hubs, routers, and the like, and are used to convert high speed signals from a first medium to a second medium. As an example, one popular type of standard interface converter known as a GBIC converts signals from optical to electrical form; optical signals carried on fiber optic cables being used to communicate over the network, and electrical signals being used within the device housing the GBIC. Other forms convert signals from twisted-pair copper conductors used in high-speed networks to electrical signals suitable for the network device.
A smart interface converter (SIC) is a standard interface converter with additional logic allowing it to process packets on the fly. This additional logic may include the ability to query the status of the SIC, perform internal tests, and/or perform data capture and analysis. The SIC device also adds the ability to inject data packets into the high speed data stream, and to modify the contents of data packets on the fly. Further details of SIC devices are described in detail in "A Method of Creating Low-Bandwidth Channel within a Packet Stream," application Ser. No. 10/688,340, filed Oct. 17, 2003, the entire disclosure of which is hereby incorporated by reference.
A first embodiment of the present invention is shown in FIG. 1. This embodiment provides a distributed transparent clock which is wrapped around single or multiple network elements such as routers, switches, and the like. The network elements may be legacy equipment that are not IEEE 1588 aware. Distributed operation is performed by at least two cooperating SICs. Each cooperating SIC timestamps IEEE 1588 SYNC and DELAY REQ packets. The cooperating SICs modify FOLLOW UP and DELAY RESP packets on the fly by adjusting their timestamps.
IEEE 1588 master clock 100 communicates 110 to smart interface converter (SIC) 120. SIC 120 communicates with network elements 130, which may be one or more switches, routers, or the like. SIC 140 also communicates with network elements 130, and communicates 150 with IEEE 1588 slave clock 160. SIC 120, a designated master clock for out of band local clock synchronization with other SIC devices, also sends timing signals via out of band channel 170 to SIC 140. Communications paths 110 and 150 are preferably Ethernet links. Out of band channel 170 may be a coaxial cable, or other signaling medium such as a fiber optic link, free radiating optical link, or RF link. SIC 120 and SIC 140 pass packet traffic bidirectionally. While FIG. 1 shows a single master SIC 120 and a single slave SIC 140, multiple slave SICs may be present.
Referring to the timing diagram in FIG. 1, Master clock 100 sends SYNC packet at time T0. This packet passes through SIC120 where it is timestamped at time T1. The SYNC packet passes through SIC 120 and network element 130 to SIC 140. SIC 140 takes another timestamp at time T2. According to the IEEE 1588 PTP protocol Master clock 100 sends a FOLLOW UP packet containing original timestamp T0. Because packets such as SYNC could be delayed by an unpredictable time interval at network element 130, according to the present invention, the FOLLOW UP packet is modified on the fly when it passes through SIC 120, including the T1 timestamp in the FOLLOW UP packet. When the FOLLOW UP packet passes through SIC 140, the original T0 timestamp is modified by adjusting for the delay, i.e. T0=T0−(T2−T1) and inserted back into the FOLLOW UP packet on the fly, and passed to IEEE 1588 slave 160. Similar processing takes place with the DELAY REQ and DELAY RES packets.
In an alternative embodiment, rather than modifying FOLLOW UP and DELAY RES packets, packets containing the necessary timestamps and sequence ID information between the SICs. As an example, rather than have SIC 120 modify the FOLLOW UP packet with timestamp T1, SIC 120 could send a packet containing the proper sequence ID and the T1timestamp to SIC 140. Then Sic 140 could calculate network element delays for the SYNC/DELAY REQ packets, and pass this data to IEEE 1588 slave 160.
In order to provide accurate timestamping, SICs 120 and 140 synchronize their local clocks with each other via out of band channel 170. In one embodiment, a coaxial cable is used. An optical cable may be used, or free-radiating signals such as RF or optical signals may be used. If SIC 120 and 140 are co-located, for example in the same enclosure, channel 170 may be a simple electrical connection or a trace on a printed circuit board interconnecting the devices.
One of the SIC devices, SIC 120 in the example shown, is designated as a master for clock synchronization sends a periodic strobe pulse to synchronize the frequencies of other SIC device clocks. In addition, on a periodic basis such as once per second or some defined interval, a special strobe signal is sent out to indicate a time boundary condition. Upon receiving the special strobe signal, the slave SIC devices adjust their local clocks to the appropriate time boundary accordingly. As an example, in a given implementation SIC 120 sends out a strobe pulse every 1.6 microseconds. This periodic strobe allows other SIC devices to synchronize their clock frequencies. Once per second SIC 120 sends out a special strobe which instructs the other SIC devices to set the fractional seconds portion of their local clocks to zero. This process insures that SIC device clocks are synchronized both in terms of frequency and time.
While the out of band clock synchronization signal is shown in FIG. 1 originating from SIC 120, this signal could be supplied by an external source, as an example, an external source using a high grade oscillator.
In an alternate embodiment, SIC 120 and SIC 140 may not only be co-located, but may share logic. In such an embodiment, for example where SIC 120 and 140 share the same clock generation circuitry, the use of an out of band channel to synchronize SIC clocks is not necessary. This would be the case, for example, if multiple SIC devices were implemented on one FPGA. Such an embodiment is shown in FIG. 2. IEEE master clock 200 communicates 210 with device400. One channel of device 400, functionally similar to SIC 120 of FIG. 1, comprises PHY 410, shared logic 450, and PHY420. PHY 420 communicates 220 with network elements 230. A second channel of device 400, functionally similar to SIC140 of FIG. 1, comprises PHY 430, shared logic 450, and PHY 440. This channel communicates 240 with network elements230, and communicates 250 with IEEE 1588 slave clock 260.
An additional embodiment of the present invention is shown in FIG. 3, providing a distributed boundary block. In this embodiment, each SIC device implements the IEEE 1588 PTP protocol, establishing separate PTP sessions.
IEEE 1588 master clock 300 communicates 310 with SIC 320 which acts as an IEEE 1588 slave clock. SIC 330 acts as a IEEE 1588 master clock as it communicates 350 between network 330 and IEEE 1588 slave clock 360.
Using IEEE 1588 protocols, SIC 320 synchronizes its internal clock with that of IEEE master 300. SIC 340 acts as IEEE 1588 master clock, with IEEE 588 slave clock 360 synchronizing its clock to SIC 340.
SIC 320 synchronizes SIC 340's clock by sending periodic synch pulses via out of band link 370. SIC 320 also injects a special EPOCH packet into the datastream through network element 330 directed to SIC 340. SIC 340 uses this packet to synchronize its EPOCH. SIC 320 sends a special pulse via out of band link 370 to indicate to SIC 340 that the follow up EPOCH packet will have EPOCH data related to this event.
In an embodiment where SIC 320 and SIC 340 are co-located or share processing circuitry, such as shown in FIG. 2, the use of the out of band channel and transmission of EPOCH packets may not be necessary.
SRC=https://www.google.com.hk/patents/US7689854
Method and apparatus for establishing IEEE 1588 clock synchronization across a network element comprising first and second cooperating smart interface converters wrapping the network element的更多相关文章
- Method and apparatus for loading a segment register in a microprocessor capable of operating in multiple modes
A microprocessor contains an address generation unit, including a segment block, for loading descrip ...
- Method and apparatus for verification of coherence for shared cache components in a system verification environment
A method and apparatus for verification of coherence for shared cache components in a system verific ...
- Method and apparatus for speculative execution of uncontended lock instructions
A method and apparatus for executing lock instructions speculatively in an out-of-order processor ar ...
- Method and apparatus for an atomic operation in a parallel computing environment
A method and apparatus for a atomic operation is described. A method comprises receiving a first pro ...
- Method and apparatus for transitioning between instruction sets in a processor
A data processor (104) is described. The data processor (104) is capable of decoding and executing a ...
- Method and apparatus for encoding data to be self-describing by storing tag records describing said data terminated by a self-referential record
A computer-implemented method and apparatus in a computer system of processing data generated by a f ...
- Method and Apparatus for Providing Highly-Scalable Network Storage for Well-Gridded Objects
An apparatus comprising a plurality of storage nodes comprising a plurality of corresponding storage ...
- Method and apparatus for training a memory signal via an error signal of a memory
Described herein is a method and an apparatus for training a memory signal via an error signal of a ...
- Method and apparatus for providing total and partial store ordering for a memory in multi-processor system
An improved memory model and implementation is disclosed. The memory model includes a Total Store Or ...
随机推荐
- 从零开始使用git第一篇:下载安装配置
从零开始使用git 第一篇:下载安装配置 第一篇:从零开始使用git第一篇:下载安装配置 第二篇:从零开始使用git第二篇:git实践操作 第三篇:从零开始使用git第三篇:git撤销操作.分支操作和 ...
- 9.3 Binder系统_驱动情景分析_服务获取过程
4. 服务获取过程 test_client客户端: (1)在用户态先构造name=“hello”的数据(服务的名字是hello),调用ioctl发送数据给service_manager(handle= ...
- UIActionSheet用法
//上拉菜单 1 UIActionSheet* sheet = [[UIActionSheet alloc] initWithTitle:nil delegate:self cancelButtonT ...
- php实现判断树的子结构
php实现判断树的子结构 一.总结 很简单的递归判断 二.php实现判断树的子结构 题目描述: 输入两棵二叉树A,B,判断B是不是A的子结构.(ps:我们约定空树不是任意一个树的子结构) 三.代码 代 ...
- stm32的DMA传输一半中断
这里本想做一个录音程序 硬件很简单: MIC(麦克风)放大滤波电路---->stm32的ADC----->DMA通道----->一个数组缓存------->通过FATFS的 ...
- JavaScript的return程序流
原文 简书原文:https://www.jianshu.com/p/cd65a26a5b0c 大纲 1.场景分析 2.代码分析 3.总结分析 1.场景分析 以下有两段代码,这两段代码都可以使用检查输入 ...
- DC综合:划分与编码风格
划分与编码风格 合理的设计划分和好的HDL编码风格对成功的综合影响很大. 逻辑划分是成功综合(和布局布线,如果布图是层次化的)的关键. 综合划分 "分而治之" 把复杂的设计化简为更 ...
- 机器学习01-kNN邻近算法
k-近邻算法 概述:k-近邻算法採用測量不同特征值之间的距离方法进行分类 长处:精度高.对于异常值不敏感.无数据输入假定 缺点:计算复杂度高,空间复杂度高,而且它没有办法各处基础数据的一些内部信息数据 ...
- ios开发网络学习六:设置队列请求与RunLoop
#import "ViewController.h" @interface ViewController ()<NSURLConnectionDataDelegate> ...
- php thinkphp uploadify
模板文件: <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w ...