Re: [ntp:questions] SNTP with 1ms of precision?
David Woolley wrote: > Rob wrote: > >> >> What nearly all of them do is send a query at the specified (or fixed) >> interval, and put the time from the reply packet in the clock, stepping >> it to that value. >> >> Maybe it cannot be rightfully called SNTP, but it is what we have. > > The SNTP supplied with the reference implementation (4.2.5p32) seems > close to that. It spreads the steps out, but using tickadj, to slew at > +/-500ppm, typically, but seems to have no concept of frequency. The devices that I am talking about probably do not use this reference implementation, and if they do they probably have heavily modified it as they underlying "OS" does not have tickadj. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Rob wrote: What nearly all of them do is send a query at the specified (or fixed) interval, and put the time from the reply packet in the clock, stepping it to that value. Maybe it cannot be rightfully called SNTP, but it is what we have. The SNTP supplied with the reference implementation (4.2.5p32) seems close to that. It spreads the steps out, but using tickadj, to slew at +/-500ppm, typically, but seems to have no concept of frequency. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Danny Mayer wrote: RFC 2030, which RFC5905 obsoleted, said the same thing. It doesn't The substring "disci" appears nowhere in RFC 2030; together with failing to find anything of the sort in a quick skim, and a specific skim of the wording around the word "clock!, I conclude that RFC 2030 says nothing about what clock discipline algorithm should be used. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
David, The basic definition of SNTP has not changed over the yeas, although rfc5905 does clarify the intended scope and role of primary servers, secondary servers and clients. It was the expected, but not required, model that the Unix adjtime() system call be used if the offset was less than an unspecified value and the settimeofday() if greater. There was no intent, either in the earlier SNTP specifications or rfc5905 to specify the SNTP clock discipline algorithm itself. Dave David Woolley wrote: Danny Mayer wrote: On 6/16/2010 5:22 PM, Maarten Wiltink wrote: "Marcelo Pimenta" wrote in message news:aanlktilq6m8apeoasibr-o8mhwifqkfv9xyf6mudr...@mail.gmail.com... [...] The NTP algorithm is much more complicated than the SNTP algorithm. The short, short version: there is no SNTP algorithm. SNTP is NTP _without_ the algorithms. Using NTP means continuously adjusting the speed of your clock so it tracks real time as best you can make it, while SNTP is simply asking what time [they think] it is. This is a totally inaccurate statement. See RFC 5905 Section 14. SNTP is That RFC was published after this thread was started! You can't go changing the definitions just for you convenience. Even if it had been published, say six months earlier, the reality is that de facto and historic definitions would still dominate the market. merely a subset of the full NTP protocol. An SNTP server is one with it's own refclock and not dependent on any other upstream servers while and SNTP client is one with a single upstream server and no dependent clients. An SNTP client or an SNTP server should be disciplining the clock in the same way as an NTP server. An SNTP server should continuously adjust the speed of your clock otherwise it's not SNTP compliant. In reality, most SNTP clients step the clock. A few may use a simple frequency control scheme. Once you go much beyond that, it becomes simpler to use a full NTP client, but maybe configure only one server. In fact, RFC 1305 doesn't require any specific clock discipline for NTPv3 clients; that is in an appendix, rather than in the main specification. The important clarification about SNTP, ignoring any recent attempt to redefine it, is that it doesn't specify an algorithm, rather than that it requires the use of only a trivial algorithm. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Danny Mayer wrote: > There has been no attempt to redefine SNTP. RFC 2030 is 15 years old so > it's hardly recent. Tools like ntpdate are not SNTP clients by the > definitions of any RFC you care to name, it just sets the clock. SNTP is > a simplified protocol to discipline the clock and not just set it. There are many, many simple devices (printers, routers etc) and even operating systems out there that provide a time sync option called SNTP on their config screens. You set a server address and sometimes an interval. What nearly all of them do is send a query at the specified (or fixed) interval, and put the time from the reply packet in the clock, stepping it to that value. Maybe it cannot be rightfully called SNTP, but it is what we have. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 7/8/2010 2:50 AM, David Woolley wrote: > Danny Mayer wrote: >> On 6/16/2010 5:22 PM, Maarten Wiltink wrote: >>> "Marcelo Pimenta" wrote in message >>> news:aanlktilq6m8apeoasibr-o8mhwifqkfv9xyf6mudr...@mail.gmail.com... >>> >>> [...] The NTP algorithm is much more complicated than the SNTP algorithm. >>> The short, short version: there is no SNTP algorithm. SNTP is NTP >>> _without_ the algorithms. Using NTP means continuously adjusting the >>> speed of your clock so it tracks real time as best you can make it, >>> while SNTP is simply asking what time [they think] it is. >> >> This is a totally inaccurate statement. See RFC 5905 Section 14. SNTP is > > That RFC was published after this thread was started! You can't go > changing the definitions just for you convenience. Even if it had been > published, say six months earlier, the reality is that de facto and > historic definitions would still dominate the market. > RFC 2030, which RFC5905 obsoleted, said the same thing. It doesn't change anything that I've said. I'm now using RFC 5905 because that's what we should be referring to as of now but it merely documents the status of the protocol and doesn't change it in any substantial way. >> merely a subset of the full NTP protocol. An SNTP server is one with >> it's own refclock and not dependent on any other upstream servers while >> and SNTP client is one with a single upstream server and no dependent >> clients. An SNTP client or an SNTP server should be disciplining the >> clock in the same way as an NTP server. An SNTP server should >> continuously adjust the speed of your clock otherwise it's not SNTP >> compliant. > > In reality, most SNTP clients step the clock. A few may use a simple > frequency control scheme. Once you go much beyond that, it becomes > simpler to use a full NTP client, but maybe configure only one server. > That's not an SNTP client despite people's attempts to call it that, see RFC2030. > In fact, RFC 1305 doesn't require any specific clock discipline for > NTPv3 clients; that is in an appendix, rather than in the main > specification. > > The important clarification about SNTP, ignoring any recent attempt to > redefine it, is that it doesn't specify an algorithm, rather than that > it requires the use of only a trivial algorithm. There has been no attempt to redefine SNTP. RFC 2030 is 15 years old so it's hardly recent. Tools like ntpdate are not SNTP clients by the definitions of any RFC you care to name, it just sets the clock. SNTP is a simplified protocol to discipline the clock and not just set it. Danny ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Danny Mayer wrote: On 6/16/2010 5:22 PM, Maarten Wiltink wrote: "Marcelo Pimenta" wrote in message news:aanlktilq6m8apeoasibr-o8mhwifqkfv9xyf6mudr...@mail.gmail.com... [...] The NTP algorithm is much more complicated than the SNTP algorithm. The short, short version: there is no SNTP algorithm. SNTP is NTP _without_ the algorithms. Using NTP means continuously adjusting the speed of your clock so it tracks real time as best you can make it, while SNTP is simply asking what time [they think] it is. This is a totally inaccurate statement. See RFC 5905 Section 14. SNTP is That RFC was published after this thread was started! You can't go changing the definitions just for you convenience. Even if it had been published, say six months earlier, the reality is that de facto and historic definitions would still dominate the market. merely a subset of the full NTP protocol. An SNTP server is one with it's own refclock and not dependent on any other upstream servers while and SNTP client is one with a single upstream server and no dependent clients. An SNTP client or an SNTP server should be disciplining the clock in the same way as an NTP server. An SNTP server should continuously adjust the speed of your clock otherwise it's not SNTP compliant. In reality, most SNTP clients step the clock. A few may use a simple frequency control scheme. Once you go much beyond that, it becomes simpler to use a full NTP client, but maybe configure only one server. In fact, RFC 1305 doesn't require any specific clock discipline for NTPv3 clients; that is in an appendix, rather than in the main specification. The important clarification about SNTP, ignoring any recent attempt to redefine it, is that it doesn't specify an algorithm, rather than that it requires the use of only a trivial algorithm. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 6/17/2010 9:46 AM, Marcelo Pimenta wrote: > Yep but I have an equipment for specific use with RFC2030 implemented > because of a standard that I have follow that say it is mandatory SNTP - > RFC2030. My equipment ask for time every minute and frequence tolerance of > the crystal is less then 1ms/min. They are in a intranet with 2 switches > isolated so... my question was about the SNTP because some people said to me > that SNTP cannot garantee 1ms accuracy. Is it possible, in that described > cenario to no have 1ms of accuracy? I expect that a implementation > of RFC2030 made by a person at equipment X and the implementation of the > same RFC at the same equipment X BUT by another person could give me > different results?? Someone has been feeding you a bunch of nonsense. If you have a Meinberg GPS directly attached to your server then an SNTP server is all you need. See RFC 5905. For the precision of the unit you should consult the specs of the unit. There are delays getting the data from the GPS to your server and the usual delays for NTP and the O/S in disciplining the clock but that's the only areas you need to worry about (apart from CPU load from other processes). SNTP has nothing to do with guaranteeing accuracy. The error budget comes from the various components involved and the Internet is not involved here. If you use an SNTP server provided by anyone it will obviously not give you the same exact results as a different implementation just because the code is different and it is not easy to account for variations in the code. Even a different version of the same code will give you different results but with a refclock hanging off the box you will get very good results. Danny ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 6/16/2010 5:22 PM, Maarten Wiltink wrote: > "Marcelo Pimenta" wrote in message > news:aanlktilq6m8apeoasibr-o8mhwifqkfv9xyf6mudr...@mail.gmail.com... > > [...] >> The NTP algorithm is much more complicated than the SNTP algorithm. > > The short, short version: there is no SNTP algorithm. SNTP is NTP > _without_ the algorithms. Using NTP means continuously adjusting the > speed of your clock so it tracks real time as best you can make it, > while SNTP is simply asking what time [they think] it is. This is a totally inaccurate statement. See RFC 5905 Section 14. SNTP is merely a subset of the full NTP protocol. An SNTP server is one with it's own refclock and not dependent on any other upstream servers while and SNTP client is one with a single upstream server and no dependent clients. An SNTP client or an SNTP server should be disciplining the clock in the same way as an NTP server. An SNTP server should continuously adjust the speed of your clock otherwise it's not SNTP compliant. Danny ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: Yes!! My question is about ms(MILIsecond) a collegue of this forum said that NTP give nsec accurancy... Is impossible to have usec imagine nsec! ns (as in less than us) is absolutely possible for NTP, but it cannot be done over a network, you need a good deal of hardware support: Poul-Henning Kemp has made several ntp servers where the local clock is slaved to UTC with better than 100 ns precision: To do so he needed to replace the motherboard crystal with an Rb oscillator, use external hw to measure the interrupt latency and use a timing-optimized GPS clock source. Terje -- - "almost all programming can be viewed as an exercise in caching" ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-18, Richard B. Gilbert wrote: > David J Taylor wrote: >> "Richard B. Gilbert" wrote in message >> news:fcgdnxzbxfvy7ifrnz2dnuvz_sadn...@giganews.com... >> [] >>> If you have a GPS timing receiver (different from navigation >>> receiver), with a PPS output, one edge of that pulse should be >>> accurate to within 50 nanoseconds. A serial output will tell you >>> which second is marked by the pulse. >> >> .. although the very popular GPS 18x LVC specifies: >> >> "Accuracy: Measurement Pulse Output Time: +/- 1 microsecond at [the] >> rising edge of the pulse." >> >> so not quite as good as 50 nanoseconds. >> >> Cheers, >> David > > The difference is not significant unless you can somehow take advantage > of the greater accuracy! The difference in price between the Motorola > device whose specifications I was citing and the GPS 18x LVC is quite > significant; ISTR that I paid about $300 for mine. If I had to replace > it, I'd go with the GPS 18X LVC. Agreed. The interrupt serviced routines are simply not accurate enough to make use of the accuracy better than about 1us. In tests I ran I got roughly 1us random variations in the itnerrupt servicing. > > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"unruh" wrote in message news:slrni1mjri.fnp.un...@wormhole.physics.ubc.ca... > On 2010-06-17, Marcelo Pimenta wrote: >> [...] What's your SO? > Define SO. Operating System in Italian. Groetjes, Maarten Wiltink ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"Richard B. Gilbert" wrote in message news:roadnwsbv-fooybrnz2dnuvz_oodn...@giganews.com... [] The difference is not significant unless you can somehow take advantage of the greater accuracy! Of course. The difference in price between the Motorola device whose specifications I was citing and the GPS 18x LVC is quite significant; ISTR that I paid about $300 for mine. If I had to replace it, I'd go with the GPS 18X LVC. I just didn't want anyone to think that the GPS 18x LVC was 50ns (as you didn't say what device you were quoting). Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
unruh wrote: > NO idea what this means. SNTP uses the ntp ehternet protocol Ie, the I presume that "ethernet" in there is a slip of the keyboard. Being pedantic, NTP protocol messages are carried in UDP datagrams which are carried in IPv4 datagrams (with IPv6 to be named later :) which are then carried in "whatever" - 99 times out of 10 that may be an Ethernet frame, but it does not have to be, thanks to the magic that is IP and layering :) rick jones -- oxymoron n, Hummer H2 with California Save Our Coasts and Oceans plates these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
David J Taylor wrote: "Richard B. Gilbert" wrote in message news:fcgdnxzbxfvy7ifrnz2dnuvz_sadn...@giganews.com... [] If you have a GPS timing receiver (different from navigation receiver), with a PPS output, one edge of that pulse should be accurate to within 50 nanoseconds. A serial output will tell you which second is marked by the pulse. .. although the very popular GPS 18x LVC specifies: "Accuracy: Measurement Pulse Output Time: +/- 1 microsecond at [the] rising edge of the pulse." so not quite as good as 50 nanoseconds. Cheers, David The difference is not significant unless you can somehow take advantage of the greater accuracy! The difference in price between the Motorola device whose specifications I was citing and the GPS 18x LVC is quite significant; ISTR that I paid about $300 for mine. If I had to replace it, I'd go with the GPS 18X LVC. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-17, Marcelo Pimenta wrote: > 2010/6/16 David Woolley > >> Marcelo Pimenta wrote: >> >> >>> Rob, my understading about the use of SNTP and NTP is: while SNTP provides >>> time synchronization within *one *network, NTP allows a global time >>> >> >> You are confusing it with timed. SNTP also expects to use global time, and >> NTP can be used with an arbitrary timebase, providing all stratum 0s have >> the same arbitrary time (i.e., typically there is only one). > > > I thought that SNTP was used in a kind of "Intranet" with no routers and no > internet, so the low latency there is no need of so much calculations to > adjust the clock. No. SNTP is a system that uses the ntp protocol to exchange time ( timestamp the output packet, the other side timestamps the receive time and emit time, and the client then uses that to estimate the "true time difference" via dt=(t4+t1)/2-(t2+t3)/2). SNTP thenb uses this to discipline the local clock by some method unspecified. > > >> >> >> synchronization on the internet. SNTP provides the current time, the >>> current >>> number of leap seconds and the warning flags marking the intriduction of a >>> leap second correction. >>> >>> The NTP algorithm is much more complicated than the SNTP algorithm. NTP >>> normally uses multiple time servers to verify the time and then controls >>> the >>> >> >> There is no SNTP algorithm, beyond basic validity checking. SNTP covers >> any use of NTP wire formats that falls short of being a compliant NTP >> implementation. SNTP implementations could still use PLL clock discipline >> code, or could use linear regression solutions for offset and rate. > > > Yeh!! You are right!! RFC is the same. But, the server can respond SNTP e > NTP request? Or the request is the same? If not the same, who decide it? The > client? If client ask SNTP to the server, it will respond SNTP? NO idea what this means. SNTP uses the ntp ehternet protocol Ie, the packets interchanged are ntp packets. There is no "asks SNTP" There is the ntp packets and that is what sntp uses. > > >> >> slew rate of the device. The algorithm determines if the values are >>> accurate >>> using several methods including fudge factors and identifying time servers >>> that don't agree with the other time servers. It then speeds up or slows >>> down the device drift rate so device time is always correct and there >>> won't >>> be any subsequent time jumps after the initial correction. >>> >>> SNTP usually uses just one Ethernet Time Server to calculate the time and >>> then it "jumps" the system time to the calculated time. It can, however, >>> >> >> One issue that has maybe not been stressed is that this strategy makes you >> extremely vulnerable to scheduling delays, e.g. you would likely never get a >> 1ms error bound on Windows if you used this strategy. >> > > Yep, Windows have a lot of "things" to do. I am using a private SO for > specific use. > > > >> Early W32Time implementations, and current ones in out of the box >> configurations are SNTP implementations, but often use internet sources. >> They do violate the specification by allowing multiple SNTP strata. > > > So, Windows XP use SNTP when I put an IP server to sincronize through > Internet? No idea what Windows uses. It is not necessary that it uses sntp. It could use anything, including guessing the time. > >> >> >> have back-up Ethernet Time Servers in case one is not avaiable(not two at >>> the same time). >>> >>> In my case, I have* only one* network. My Time Server is not a machine, is >>> a >>> meinberg GPS. In my point of view, if my source time were machines, maybe >>> NTP could be better to find a middle line between all these machines used >>> as >>> time servers. But if I am using a very good and reliable GPS(Meinberg) >>> with >>> a lot of satellites giving it the correct time, and it's pluged directly >>> in >>> my switch, I think that in this case, NTP will not make any difference. >>> >> >> I think you are being asked why you don't just use a standard ntpd for the >> platform, and also how you know that your total system meets the scheduling >> constraints for 1ms accuracy with SNTP. > > > I can't use ntpd, I have an equipment for specific use with RFC2030 > implemented because of a standard that say it is mandatory SNTP - RFC2030. > My equipment ask for time every minute and frequence tolerance of the > crystal is less then 1ms/min. They are in a intranet with 2 switches > isolated so... my question was about the SNTP because some people said to me > that SNTP cannot garantee 1ms accuracy. Is it possible, in that described > cenario to no have 1ms of accuracy? Your specifications are silly. Yes, it is possible on a local internet to have accuracy ( at least most of the time) be less than 1ms. I have done it. If some kid cuts the intern et wre of switches off hte server, there is no guarentee of any accuracy at all. > > >> >> ___ >>
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-17, Marcelo Pimenta wrote: > 2010/6/17 unruh > >> On 2010-06-16, Marcelo Pimenta wrote: >> > 2010/6/15 unruh >> > >> >> On 2010-06-14, Marcelo Pimenta wrote: >> >> > Hi everybody!! >> >> > >> >> > My question is about Time Accuracy of NTP/SNTP protocol. I want to >> know >> >> if >> >> > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a >> >> > network with 50 hosts(is the same with 100 hosts?) using only swithes, >> no >> >> > routing. All these hosts are syncronized with a Meinberg GPS. The >> traffic >> >> is >> >> > low(~25%). Is it possible to have the same time accuracy of IRIG-B for >> >> > example? >> >> >> >> I think you mean accuracy of 1ms. Yes. it can. On a local network, >> >> 100usec is easily possible ( and depending on the routers, it can be >> >> better than that). >> >> >> > >> > Yes, you are right, accuracy of 1ms. On a local network 100usec?? Even if >> we >> > use only switches(no routers), how is that possible if I have 4 types of >> > Latency increasing about 80us? Algorithms to compensate the delay? Even >> in >> > SNTP? >> >> ntp relies on the outgoing and ingoing times are the same. They could be >> one >> year, but if they are both exactly the same ( both exactly one year) >> ntp can deliver a time accuracy of 1nsec. >> Yes, I typically get time accuracies of tens of usec on a local network. >> Even over adsl frin home to work I get 1 msec easily ( and yes I can >> detect the accuracy by >> having a gps PPS receiver at both endw to test the accuracy to 1us) >> > > NTP can deliver a time accuracy of 1nsec?? I my point of view, NTP includes > methods to estimate the round-trip path delay between the server and client > but the performance is limited by SO stack latency, it is an Application > Layer protocol. In principle yes, it can. Of course in practice no. And if the delay is one year, the symmetry is not likely to be 1ns. It was simply and (extreme) example to make a point., The performance is not limited by latency as long as it is symmetric. If it is not symmetric then there is a problem > > What's the resolution of you clock? What's your SO? Define SO. > > >> >> level) >> > >> > Anyway, the question is about 1 ms. To get 100usec I'll use PTP. >> > >> > >> > >> >> > >> >> > The internal clock in my devices(hosts) starts with the crystal >> >> frequence. A >> >> > typical offset or frequence tolerance of the crystal is less then >> >> > 1ms/minute. The resolution is about 0.9ms and the devices acts as >> unicast >> >> > clients, asking GPS every minute. >> >> >> >> Why? The whole purpose of ntp is to discipline your clock so that it >> >> keeps time much better than that. >> >> >> > >> > The idea to have this is not overload the network asking time every >> second >> > to keep my accurancy in 1ms in 100% of time. And the second reason is >> about >> > how many ms your clock will be wrong in 59 sec without a frame to >> discipline >> > my clock again? PCs clock for example is not good, that's why I need a >> very >> > good crystal. >> > >> > >> >> Not sure what you mean by "the resolution is about .9ms" What is your >> >> device? >> >> >> > >> > I mean that my device can show me variations of 0.9ms, is the smallest >> > possible increase of time the clock model allows. I work with protection >> > IED(intelligent electronic device), is a term used in the electric power >> > industry to describe microprocessor-based controllers of power system >> > equipment. >> > >> > If something happens in 458ms, I will get exacly 458ms. Some equipment >> have >> > so I can trust in variation that I can capture. >> > >> >> >> >> > Is it enough? The Time Accurancy could be better if I could ask GPS >> twice >> >> a >> >> > minute? A broadcast implementation with GPS sending time every minute >> >> could >> >> > help? >> >> >> >> It depends on what you want the time for. >> >> >> > >> > I need accurancy at least of 1ms in 100% of time. >> > >> >> >> >> > >> >> > I know that these kind of issue have many variables but the main >> question >> >> > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? >> >> >> >> Yes, depending on the competence of the writer of the SNTP software. >> >> >> > >> > So, the problem is not in the standard but into implementation of SNTP to >> > not loose time to correct the clock and something like that? >> > >> > >> >> > >> >> > Thank you very much >> >> >> >> ___ >> >> questions mailing list >> >> questions@lists.ntp.org >> >> http://lists.ntp.org/listinfo/questions >> >> >> >> ___ >> questions mailing list >> questions@lists.ntp.org >> http://lists.ntp.org/listinfo/questions >> ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"David Woolley" wrote in message news:hve58p$8l...@news.eternal-september.org... [] Windows tends to use a broken SNTP. Recent versions can be configured to act as (almost?) compliant NTP clients and servers. [] .. but thanks to the great work of Dave Hart, Meinberg, and others recently Windows 2000, XP, Vista and Windows-7 can now use the reference version of NTP rather than Microsoft's own version, even to the extent of acting as stratum-1 servers with GPS reference sources. I appreciate you will know this, David, but other readers may not. I'm enthusiastic about this, so I provided a couple of write-ups on setting up NTP clients and notes on stratum-1 servers here: http://www.satsignal.eu/ntp/setup.html http://www.satsignal.eu/ntp/NTP-on-Windows-serial-port.html Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"Richard B. Gilbert" wrote in message news:fcgdnxzbxfvy7ifrnz2dnuvz_sadn...@giganews.com... [] If you have a GPS timing receiver (different from navigation receiver), with a PPS output, one edge of that pulse should be accurate to within 50 nanoseconds. A serial output will tell you which second is marked by the pulse. .. although the very popular GPS 18x LVC specifies: "Accuracy: Measurement Pulse Output Time: +/- 1 microsecond at [the] rising edge of the pulse." so not quite as good as 50 nanoseconds. Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"Marcelo Pimenta" wrote in message news:aanlktilffevh94cief43-ewqbc2dtqev4dbbpfs1d...@mail.gmail.com... [] Yes, I agree. But we were talking about NTP using Ethernet, so... that's impossible Even with Ethernet, temperature variations at the client will have a noticeable effect (at least they seem to here!). Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: 2010/6/17 Richard B. Gilbert Marcelo Pimenta wrote: Yes!! My question is about ms(MILIsecond) a collegue of this forum said that NTP give nsec accurancy... Is impossible to have usec imagine nsec! Thanks If you have a GPS timing receiver (different from navigation receiver), with a PPS output, one edge of that pulse should be accurate to within 50 nanoseconds. A serial output will tell you which second is marked by the pulse. Getting time into a computer without losing some of that accuracy can be quite difficult. Yes, I agree. But we were talking about NTP using Ethernet, so... that's impossible Sorry about that but the thread had lost its original context long before I replied! ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Steve Kostecke wrote: On 2010-06-17, Marcelo Pimenta wrote: Here's a snapshot of my "main" system showing an offset of 340us (0.340ms) from it's sys_peer over a switched GigE LAN ... Which doesn't mean the error is .34 ms; it is quite likely less than .1ms, assuming no systematic component. Offset is an optimistic estimate of what you would get with a time stepping SNTP. The NTP PLL smoothes out the offsets and provided the temperature is well controlled, the software clock maintains rather more accurate time. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
David Woolley wrote: > SNTP as used on Windows to access the Microsoft timeserver certainly > can't. Anything that runs on Windows that sets the clock on ever > poll almost certainly can't for any reasonable application load. On > Linux you have to beware of interrupt latency, NIC interrupt > aggregations. I suspect there are interrupt aggregations on more than just Linux. I would expect every *nix and even Windows to have interrupt aggregation/coalescing. I would expect it to delay notification to the host that the packet has arrived by anywhere between 5 and 20 microseconds, but that is an off-the-cuff estimate. rick jones -- Process shall set you free from the need for rational thought. these opinions are mine, all mine; HP might not want them anyway... :) feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH... ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-17, Marcelo Pimenta wrote: > Yes, I agree. But we were talking about NTP using Ethernet, so... that's > impossible Here's a snapshot of my "main" system showing an offset of 340us (0.340ms) from it's sys_peer over a switched GigE LAN ... remote refid st t when poll reach delay offset jitter = 6g-ntp .ACST. 16 a- 640 0.000 0.000 0.001 *fe80::200:24ff: .CHU1. 1 u 56 1024 377 1.130 0.340 0.017 +fe80::2a0:ccff: 241.24 4 u 865 1024 377 0.759 0.405 1.899 +fe80::240:5ff:f 18.26. 2 u 856 1024 377 0.649 0.441 0.374 A summary of today's peerstats file (w/ peer.awk): ident cntmeanrmsmax delaydistdisp = fe80::200:24ff: 198 0.563 0.436 1.087 1.037 17.633 3.433 fe80::240:5ff:f 155 0.489 0.525 1.655 0.947 20.904 4.886 fe80::2a0:ccff: 93 -0.117 0.841 2.986 1.885 23.916 8.841 (ntp-dev 4.2.7p34 on a quad core Debian system) -- Steve Kostecke NTP Public Services Project - http://support.ntp.org/ ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: I thought that SNTP was used in a kind of "Intranet" with no routers and no internet, so the low latency there is no need of so much calculations to adjust the clock. As W32Time is not fully NTP compliant, out of the box, it is probably currently true that most SNTP use is over the LAN, although the domain controller is still likely to use it over the internet. At one time, it is possible that home use shifted it towards being mainly internet. Even nowadays, ADSL routers are likely to use it over the internet; I'm not sure how the number of those relates to the number of business machines in Windows domains. Yeh!! You are right!! RFC is the same. But, the server can respond SNTP e NTP request? Or the request is the same? If not the same, who decide it? The client? If client ask SNTP to the server, it will respond SNTP? On the wire, SNTP is NTP with certain fields ignored, and certain constraints (that are often violated). I'm fairly sure that a valid SNTP server can double as a stratum 1 NTP server. An NTP server can always double as an SNTP server, and does not have to do anything to distinguish that its client is SNTP. So, Windows XP use SNTP when I put an IP server to sincronize through Internet? Windows tends to use a broken SNTP. Recent versions can be configured to act as (almost?) compliant NTP clients and servers. I can't use ntpd, I have an equipment for specific use with RFC2030 implemented because of a standard that say it is mandatory SNTP - RFC2030. Bad specification. In any case, you can still run the NTP PLL, even if you can never configure more than one server. There may be some subtlety I have missed, but I think an NTP client will be compliant with SNTP if used at stratum 2. My equipment ask for time every minute and frequence tolerance of the crystal is less then 1ms/min. They are in a intranet with 2 switches isolated so... my question was about the SNTP because some people said to me that SNTP cannot garantee 1ms accuracy. Is it possible, in that described cenario to no have 1ms of accuracy? SNTP as used on Windows to access the Microsoft timeserver certainly can't. Anything that runs on Windows that sets the clock on ever poll almost certainly can't for any reasonable application load. On Linux you have to beware of interrupt latency, NIC interrupt aggregations. Ordinary Windows applications won't be able to read the time to 1ms accuracy, unless the machine is very lightly loaded and you turn up multi-media timers to their fastest setting (and use a recent version). There is no architectural reason why an implementation of SNTP should not achieve a 1ms 100 percentile error bound, but I doubt there are many real life cases. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Rob wrote: With NTP a single query does not affect the time so much, because successive queries are fed through a filter to find the average, and the time is not stepped but slowly changed to the new value. So when there is an occasional delay in your network it will not affect the time accuracy very much. If they are more than 8 samples apart, they shouldn't affect it at all (and even that is on the pessimistic side). ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/17 Richard B. Gilbert > Marcelo Pimenta wrote: > >> Yes!! My question is about ms(MILIsecond) a collegue of this forum said >> that >> NTP give nsec accurancy... Is impossible to have usec imagine nsec! >> >> Thanks >> >> > If you have a GPS timing receiver (different from navigation receiver), > with a PPS output, one edge of that pulse should be accurate to within 50 > nanoseconds. A serial output will tell you which second is marked by the > pulse. > > Getting time into a computer without losing some of that accuracy can be > quite difficult. Yes, I agree. But we were talking about NTP using Ethernet, so... that's impossible > > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
In article , Marcelo Pimenta writes: >I thought that SNTP was used in a kind of "Intranet" with no routers and no >internet, so the low latency there is no need of so much calculations to >adjust the clock. I haven't been able to figure out what you are really trying to ask. The packets on the wire are the same for NTP and SNTP. The results you will get will depend upon the hardware and software you are using. If you have a good/reference server on your local LAN, it should be reasonable to get the clock on a client to within 1 ms most of the time. I would expect occasional quirks. How many would depend mostly on the software you are using. (If somebody is going to get hurt or killed if your clock is off by more than 1 ms, you should do something else.) You should probably setup a test bed and see how well it actually works. -- These are my opinions, not necessarily my employer's. I hate spam. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: Yes!! My question is about ms(MILIsecond) a collegue of this forum said that NTP give nsec accurancy... Is impossible to have usec imagine nsec! Thanks If you have a GPS timing receiver (different from navigation receiver), with a PPS output, one edge of that pulse should be accurate to within 50 nanoseconds. A serial output will tell you which second is marked by the pulse. Getting time into a computer without losing some of that accuracy can be quite difficult. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Yes!! My question is about ms(MILIsecond) a collegue of this forum said that NTP give nsec accurancy... Is impossible to have usec imagine nsec! Thanks 2010/6/17 David J Taylor > David, >> >> CONGRATULATIONS!!! Very good tests and presentations!! Very very good!!! >> You >> gave me real results and all your reports are consistent and show that >> it's >> impossible to have usec and nsec precision using NTP though Ethernet. You >> have SO stack latency and it's a Layer 7 protocol!!! >> >> One more time... congratulations!!! It's helped a lot!!! I hope you page >> be >> on line forever =) I have to see all things that you wrote over there >> yet >> but... I think is everything that I was looking for. >> >> Thank you very much >> > > I'm glad you liked the page - it was millisecond precision (or the lack of > it) that I was hoping to demonstrate - as you asked in the title of your > topic "SNTP with 1ms of precision?" I read ms as millisecond. Microsecond I > think you will be struggling. > > BTW: there is a simplified configuration diagram of my network here: > > http://www.satsignal.eu/mrtg/daily_ntp.html#Configuration > > Cheers, > David > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
David, CONGRATULATIONS!!! Very good tests and presentations!! Very very good!!! You gave me real results and all your reports are consistent and show that it's impossible to have usec and nsec precision using NTP though Ethernet. You have SO stack latency and it's a Layer 7 protocol!!! One more time... congratulations!!! It's helped a lot!!! I hope you page be on line forever =) I have to see all things that you wrote over there yet but... I think is everything that I was looking for. Thank you very much I'm glad you liked the page - it was millisecond precision (or the lack of it) that I was hoping to demonstrate - as you asked in the title of your topic "SNTP with 1ms of precision?" I read ms as millisecond. Microsecond I think you will be struggling. BTW: there is a simplified configuration diagram of my network here: http://www.satsignal.eu/mrtg/daily_ntp.html#Configuration Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/17 David J Taylor > NTP can deliver a time accuracy of 1nsec?? I my point of view, NTP >> includes >> methods to estimate the round-trip path delay between the server and >> client >> but the performance is limited by SO stack latency, it is an Application >> Layer protocol. >> > > Marcelo, here's what /my/ NTP achieves: > > http://www.satsignal.eu/mrtg/performance_ntp.php > > The top three PCs (Pixie, Feenix and Stamsund) are each fed with a serial > GPS/PPS signal from a Garmin GPS-18 LVC or GPS-18X LVC. Three different > operating systems. > > The bottom PCs are all Windows boxes of various versions, synced over a LAN > including a cheapo Netgear GS608 1Gb/s switch, a WRT54GL Samknows > performance measurement box, and a Netgear WNR2000v2 with wired and wireless > connections. > > I don't know whether any of those LAN-connected boxes would meet your "1ms" > requirement (although if there were all running FreeBSD they would likely be > significantly better than when running Windows). > > Cheers, > David > David, CONGRATULATIONS!!! Very good tests and presentations!! Very very good!!! You gave me real results and all your reports are consistent and show that it's impossible to have usec and nsec precision using NTP though Ethernet. You have SO stack latency and it's a Layer 7 protocol!!! One more time... congratulations!!! It's helped a lot!!! I hope you page be on line forever =) I have to see all things that you wrote over there yet but... I think is everything that I was looking for. Thank you very much > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/17 Rob > Marcelo Pimenta wrote: > >> > In my case, I have* only one* network. My Time Server is not a > machine, > >> is a > >> > meinberg GPS. In my point of view, if my source time were machines, > maybe > >> A Meingerg GPS IS a machine. > >> > > > > I agree, I said machines equipment like PC computers with Windows or > Linux > > SO. Meinberg GPS is a machine, but a machine with specific hardware and > SO > > to receive and give time. > > The Meinberg GPS runs Linux. > But I think not a normal distribution but a customized one. > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: >> > In my case, I have* only one* network. My Time Server is not a machine, >> is a >> > meinberg GPS. In my point of view, if my source time were machines, maybe >> A Meingerg GPS IS a machine. >> > > I agree, I said machines equipment like PC computers with Windows or Linux > SO. Meinberg GPS is a machine, but a machine with specific hardware and SO > to receive and give time. The Meinberg GPS runs Linux. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
NTP can deliver a time accuracy of 1nsec?? I my point of view, NTP includes methods to estimate the round-trip path delay between the server and client but the performance is limited by SO stack latency, it is an Application Layer protocol. Marcelo, here's what /my/ NTP achieves: http://www.satsignal.eu/mrtg/performance_ntp.php The top three PCs (Pixie, Feenix and Stamsund) are each fed with a serial GPS/PPS signal from a Garmin GPS-18 LVC or GPS-18X LVC. Three different operating systems. The bottom PCs are all Windows boxes of various versions, synced over a LAN including a cheapo Netgear GS608 1Gb/s switch, a WRT54GL Samknows performance measurement box, and a Netgear WNR2000v2 with wired and wireless connections. I don't know whether any of those LAN-connected boxes would meet your "1ms" requirement (although if there were all running FreeBSD they would likely be significantly better than when running Windows). Cheers, David ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: > Yep but I have an equipment for specific use with RFC2030 implemented > because of a standard that I have follow that say it is mandatory SNTP - > RFC2030. My equipment ask for time every minute and frequence tolerance of > the crystal is less then 1ms/min. They are in a intranet with 2 switches > isolated so... my question was about the SNTP because some people said to me > that SNTP cannot garantee 1ms accuracy. Is it possible, in that described > cenario to no have 1ms of accuracy? I expect that a implementation > of RFC2030 made by a person at equipment X and the implementation of the > same RFC at the same equipment X BUT by another person could give me > different results?? SNTP is much worse than NTP because with SNTP a single query that somehow gets delayed (e.g. momentary queue on a switchport) will immediately step the time and it will be wrong for the entire period until you make the next query. With NTP a single query does not affect the time so much, because successive queries are fed through a filter to find the average, and the time is not stepped but slowly changed to the new value. So when there is an occasional delay in your network it will not affect the time accuracy very much. But having said that, there is nothing in NTP or SNTP that will guarantee you an accuracy of 1ms. It is a best effort protocol. It will usually do better but there is no guarantee. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: > I thought that SNTP was used in a kind of "Intranet" with no routers and no > internet, so the low latency there is no need of so much calculations to > adjust the clock. No, SNTP is used by software manufacturers who are too lazy to implement full NTP. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/17 unruh > On 2010-06-16, Marcelo Pimenta wrote: > > 2010/6/16 Rob > > > (...) > > > In my case, I have* only one* network. My Time Server is not a machine, > is a > > meinberg GPS. In my point of view, if my source time were machines, maybe > A Meingerg GPS IS a machine. > I agree, I said machines equipment like PC computers with Windows or Linux SO. Meinberg GPS is a machine, but a machine with specific hardware and SO to receive and give time. > > NTP could be better to find a middle line between all these machines used > as > > time servers. But if I am using a very good and reliable GPS(Meinberg) > with > > a lot of satellites giving it the correct time, and it's pluged directly > in > > my switch, I think that in this case, NTP will not make any difference. > > Sure it will. ntp is designed to deliver as accurate a time as poosible > It does not care if the soruce is anohter computer, a GPS a radion > transmitter, a CDMS clock broadcast, or someone's wristwatch. > > One of its key purposes is to make sure that it gets rid of time > transmission delays as much as possible, and to make sure that the > clocks it uses are sensible. > Note that your Mienberg gps could also give bad time. Someone could > decided that they should enclose it in a coper wire mesh, making the gps > signal inaccessible. Or the computer could fail. or > > Meinberg could give me bad time but I have feedbacks about it and a contact that can close when an antenna fail or other part of GPS fail and cannot garantee the time. > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/17 unruh > On 2010-06-16, Marcelo Pimenta wrote: > > 2010/6/15 unruh > > > >> On 2010-06-14, Marcelo Pimenta wrote: > >> > Hi everybody!! > >> > > >> > My question is about Time Accuracy of NTP/SNTP protocol. I want to > know > >> if > >> > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a > >> > network with 50 hosts(is the same with 100 hosts?) using only swithes, > no > >> > routing. All these hosts are syncronized with a Meinberg GPS. The > traffic > >> is > >> > low(~25%). Is it possible to have the same time accuracy of IRIG-B for > >> > example? > >> > >> I think you mean accuracy of 1ms. Yes. it can. On a local network, > >> 100usec is easily possible ( and depending on the routers, it can be > >> better than that). > >> > > > > Yes, you are right, accuracy of 1ms. On a local network 100usec?? Even if > we > > use only switches(no routers), how is that possible if I have 4 types of > > Latency increasing about 80us? Algorithms to compensate the delay? Even > in > > SNTP? > > ntp relies on the outgoing and ingoing times are the same. They could be > one > year, but if they are both exactly the same ( both exactly one year) > ntp can deliver a time accuracy of 1nsec. > Yes, I typically get time accuracies of tens of usec on a local network. > Even over adsl frin home to work I get 1 msec easily ( and yes I can > detect the accuracy by > having a gps PPS receiver at both endw to test the accuracy to 1us) > NTP can deliver a time accuracy of 1nsec?? I my point of view, NTP includes methods to estimate the round-trip path delay between the server and client but the performance is limited by SO stack latency, it is an Application Layer protocol. What's the resolution of you clock? What's your SO? > > level) > > > > Anyway, the question is about 1 ms. To get 100usec I'll use PTP. > > > > > > > >> > > >> > The internal clock in my devices(hosts) starts with the crystal > >> frequence. A > >> > typical offset or frequence tolerance of the crystal is less then > >> > 1ms/minute. The resolution is about 0.9ms and the devices acts as > unicast > >> > clients, asking GPS every minute. > >> > >> Why? The whole purpose of ntp is to discipline your clock so that it > >> keeps time much better than that. > >> > > > > The idea to have this is not overload the network asking time every > second > > to keep my accurancy in 1ms in 100% of time. And the second reason is > about > > how many ms your clock will be wrong in 59 sec without a frame to > discipline > > my clock again? PCs clock for example is not good, that's why I need a > very > > good crystal. > > > > > >> Not sure what you mean by "the resolution is about .9ms" What is your > >> device? > >> > > > > I mean that my device can show me variations of 0.9ms, is the smallest > > possible increase of time the clock model allows. I work with protection > > IED(intelligent electronic device), is a term used in the electric power > > industry to describe microprocessor-based controllers of power system > > equipment. > > > > If something happens in 458ms, I will get exacly 458ms. Some equipment > have > > so I can trust in variation that I can capture. > > > >> > >> > Is it enough? The Time Accurancy could be better if I could ask GPS > twice > >> a > >> > minute? A broadcast implementation with GPS sending time every minute > >> could > >> > help? > >> > >> It depends on what you want the time for. > >> > > > > I need accurancy at least of 1ms in 100% of time. > > > >> > >> > > >> > I know that these kind of issue have many variables but the main > question > >> > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? > >> > >> Yes, depending on the competence of the writer of the SNTP software. > >> > > > > So, the problem is not in the standard but into implementation of SNTP to > > not loose time to correct the clock and something like that? > > > > > >> > > >> > Thank you very much > >> > >> ___ > >> questions mailing list > >> questions@lists.ntp.org > >> http://lists.ntp.org/listinfo/questions > >> > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/16 Maarten Wiltink > "Marcelo Pimenta" wrote in message > news:aanlktilaoduniqjgpigohpzvjcv_zmsw_tr7naj6b...@mail.gmail.com... > > > [...] accuracy of 1ms. On a local network 100usec?? Even if we use > > only switches(no routers), how is that possible if I have 4 types of > > Latency increasing about 80us? > > The latency is corrected for. A query-response cycle is a back-and-forth > exchange. The assumption is that the latency for the response (going > forth) is equal to the latency for the response (going back), and both > are half of the total latency, which you do know. So you estimate that > the response was sent half the total latency time before the moment you > received it. > > This often works well, and sometimes not so well. Domestic ADSL with a > saturated upstream and downstream capacity left is one common case where > it works not so well. > Yeh, I agree with you. > > > [...] > > I need accurancy at least of 1ms in 100% of time. > > That's a *very* tall order. > Yep but I have an equipment for specific use with RFC2030 implemented because of a standard that I have follow that say it is mandatory SNTP - RFC2030. My equipment ask for time every minute and frequence tolerance of the crystal is less then 1ms/min. They are in a intranet with 2 switches isolated so... my question was about the SNTP because some people said to me that SNTP cannot garantee 1ms accuracy. Is it possible, in that described cenario to no have 1ms of accuracy? I expect that a implementation of RFC2030 made by a person at equipment X and the implementation of the same RFC at the same equipment X BUT by another person could give me different results?? > > Groetjes, > Maarten Wiltink > > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/16 David Woolley > Marcelo Pimenta wrote: > > >> Rob, my understading about the use of SNTP and NTP is: while SNTP provides >> time synchronization within *one *network, NTP allows a global time >> > > You are confusing it with timed. SNTP also expects to use global time, and > NTP can be used with an arbitrary timebase, providing all stratum 0s have > the same arbitrary time (i.e., typically there is only one). I thought that SNTP was used in a kind of "Intranet" with no routers and no internet, so the low latency there is no need of so much calculations to adjust the clock. > > > synchronization on the internet. SNTP provides the current time, the >> current >> number of leap seconds and the warning flags marking the intriduction of a >> leap second correction. >> >> The NTP algorithm is much more complicated than the SNTP algorithm. NTP >> normally uses multiple time servers to verify the time and then controls >> the >> > > There is no SNTP algorithm, beyond basic validity checking. SNTP covers > any use of NTP wire formats that falls short of being a compliant NTP > implementation. SNTP implementations could still use PLL clock discipline > code, or could use linear regression solutions for offset and rate. Yeh!! You are right!! RFC is the same. But, the server can respond SNTP e NTP request? Or the request is the same? If not the same, who decide it? The client? If client ask SNTP to the server, it will respond SNTP? > > slew rate of the device. The algorithm determines if the values are >> accurate >> using several methods including fudge factors and identifying time servers >> that don't agree with the other time servers. It then speeds up or slows >> down the device drift rate so device time is always correct and there >> won't >> be any subsequent time jumps after the initial correction. >> >> SNTP usually uses just one Ethernet Time Server to calculate the time and >> then it "jumps" the system time to the calculated time. It can, however, >> > > One issue that has maybe not been stressed is that this strategy makes you > extremely vulnerable to scheduling delays, e.g. you would likely never get a > 1ms error bound on Windows if you used this strategy. > Yep, Windows have a lot of "things" to do. I am using a private SO for specific use. > Early W32Time implementations, and current ones in out of the box > configurations are SNTP implementations, but often use internet sources. > They do violate the specification by allowing multiple SNTP strata. So, Windows XP use SNTP when I put an IP server to sincronize through Internet? > > > have back-up Ethernet Time Servers in case one is not avaiable(not two at >> the same time). >> >> In my case, I have* only one* network. My Time Server is not a machine, is >> a >> meinberg GPS. In my point of view, if my source time were machines, maybe >> NTP could be better to find a middle line between all these machines used >> as >> time servers. But if I am using a very good and reliable GPS(Meinberg) >> with >> a lot of satellites giving it the correct time, and it's pluged directly >> in >> my switch, I think that in this case, NTP will not make any difference. >> > > I think you are being asked why you don't just use a standard ntpd for the > platform, and also how you know that your total system meets the scheduling > constraints for 1ms accuracy with SNTP. I can't use ntpd, I have an equipment for specific use with RFC2030 implemented because of a standard that say it is mandatory SNTP - RFC2030. My equipment ask for time every minute and frequence tolerance of the crystal is less then 1ms/min. They are in a intranet with 2 switches isolated so... my question was about the SNTP because some people said to me that SNTP cannot garantee 1ms accuracy. Is it possible, in that described cenario to no have 1ms of accuracy? > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-16, Marcelo Pimenta wrote: > 2010/6/16 Rob > >> Marcelo Pimenta wrote: >> > The idea to have this is not overload the network asking time every >> second >> > to keep my accurancy in 1ms in 100% of time. And the second reason is >> about >> > how many ms your clock will be wrong in 59 sec without a frame to >> discipline >> > my clock again? PCs clock for example is not good, that's why I need a >> very >> > good crystal. >> >> Why are you causing your own problems by sticking to SNTP instead of >> using NTP which was designed to overcome the problems you fear? >> > > > Rob, my understading about the use of SNTP and NTP is: while SNTP provides > time synchronization within *one *network, NTP allows a global time > synchronization on the internet. SNTP provides the current time, the current > number of leap seconds and the warning flags marking the intriduction of a > leap second correction. > > The NTP algorithm is much more complicated than the SNTP algorithm. NTP > normally uses multiple time servers to verify the time and then controls the > slew rate of the device. The algorithm determines if the values are accurate > using several methods including fudge factors and identifying time servers > that don't agree with the other time servers. It then speeds up or slows > down the device drift rate so device time is always correct and there won't > be any subsequent time jumps after the initial correction. No. sntp is used either for a stratum 0 time source, or a terminal client which will never be a peer. The algorithm is suppposed to be the same for both. There is nothing about "one network" or anything else. YOu seem to be describing the difference between the rdate protocol and the ntp protocol. > > SNTP usually uses just one Ethernet Time Server to calculate the time and > then it "jumps" the system time to the calculated time. It can, however, > have back-up Ethernet Time Servers in case one is not avaiable(not two at > the same time). No. That is NOT sntp. It is rdate. Why anyone would use rdate nowadays I have no idea. > > In my case, I have* only one* network. My Time Server is not a machine, is a > meinberg GPS. In my point of view, if my source time were machines, maybe A Meingerg GPS IS a machine. > NTP could be better to find a middle line between all these machines used as > time servers. But if I am using a very good and reliable GPS(Meinberg) with > a lot of satellites giving it the correct time, and it's pluged directly in > my switch, I think that in this case, NTP will not make any difference. Sure it will. ntp is designed to deliver as accurate a time as poosible It does not care if the soruce is anohter computer, a GPS a radion transmitter, a CDMS clock broadcast, or someone's wristwatch. One of its key purposes is to make sure that it gets rid of time transmission delays as much as possible, and to make sure that the clocks it uses are sensible. Note that your Mienberg gps could also give bad time. Someone could decided that they should enclose it in a coper wire mesh, making the gps signal inaccessible. Or the computer could fail. or > > >> >> ___ >> questions mailing list >> questions@lists.ntp.org >> http://lists.ntp.org/listinfo/questions >> ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-16, Marcelo Pimenta wrote: > 2010/6/15 unruh > >> On 2010-06-14, Marcelo Pimenta wrote: >> > Hi everybody!! >> > >> > My question is about Time Accuracy of NTP/SNTP protocol. I want to know >> if >> > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a >> > network with 50 hosts(is the same with 100 hosts?) using only swithes, no >> > routing. All these hosts are syncronized with a Meinberg GPS. The traffic >> is >> > low(~25%). Is it possible to have the same time accuracy of IRIG-B for >> > example? >> >> I think you mean accuracy of 1ms. Yes. it can. On a local network, >> 100usec is easily possible ( and depending on the routers, it can be >> better than that). >> > > Yes, you are right, accuracy of 1ms. On a local network 100usec?? Even if we > use only switches(no routers), how is that possible if I have 4 types of > Latency increasing about 80us? Algorithms to compensate the delay? Even in > SNTP? ntp relies on the outgoing and ingoing times are the same. They could be one year, but if they are both exactly the same ( both exactly one year) ntp can deliver a time accuracy of 1nsec. Yes, I typically get time accuracies of tens of usec on a local network. Even over adsl frin home to work I get 1 msec easily ( and yes I can detect the accuracy by having a gps PPS receiver at both endw to test the accuracy to 1us) level) > > Anyway, the question is about 1 ms. To get 100usec I'll use PTP. > > > >> > >> > The internal clock in my devices(hosts) starts with the crystal >> frequence. A >> > typical offset or frequence tolerance of the crystal is less then >> > 1ms/minute. The resolution is about 0.9ms and the devices acts as unicast >> > clients, asking GPS every minute. >> >> Why? The whole purpose of ntp is to discipline your clock so that it >> keeps time much better than that. >> > > The idea to have this is not overload the network asking time every second > to keep my accurancy in 1ms in 100% of time. And the second reason is about > how many ms your clock will be wrong in 59 sec without a frame to discipline > my clock again? PCs clock for example is not good, that's why I need a very > good crystal. > > >> Not sure what you mean by "the resolution is about .9ms" What is your >> device? >> > > I mean that my device can show me variations of 0.9ms, is the smallest > possible increase of time the clock model allows. I work with protection > IED(intelligent electronic device), is a term used in the electric power > industry to describe microprocessor-based controllers of power system > equipment. > > If something happens in 458ms, I will get exacly 458ms. Some equipment have > so I can trust in variation that I can capture. > >> >> > Is it enough? The Time Accurancy could be better if I could ask GPS twice >> a >> > minute? A broadcast implementation with GPS sending time every minute >> could >> > help? >> >> It depends on what you want the time for. >> > > I need accurancy at least of 1ms in 100% of time. > >> >> > >> > I know that these kind of issue have many variables but the main question >> > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? >> >> Yes, depending on the competence of the writer of the SNTP software. >> > > So, the problem is not in the standard but into implementation of SNTP to > not loose time to correct the clock and something like that? > > >> > >> > Thank you very much >> >> ___ >> questions mailing list >> questions@lists.ntp.org >> http://lists.ntp.org/listinfo/questions >> ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"Marcelo Pimenta" wrote in message news:aanlktilaoduniqjgpigohpzvjcv_zmsw_tr7naj6b...@mail.gmail.com... > [...] accuracy of 1ms. On a local network 100usec?? Even if we use > only switches(no routers), how is that possible if I have 4 types of > Latency increasing about 80us? The latency is corrected for. A query-response cycle is a back-and-forth exchange. The assumption is that the latency for the response (going forth) is equal to the latency for the response (going back), and both are half of the total latency, which you do know. So you estimate that the response was sent half the total latency time before the moment you received it. This often works well, and sometimes not so well. Domestic ADSL with a saturated upstream and downstream capacity left is one common case where it works not so well. [...] > I need accurancy at least of 1ms in 100% of time. That's a *very* tall order. Groetjes, Maarten Wiltink ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
"Marcelo Pimenta" wrote in message news:aanlktilq6m8apeoasibr-o8mhwifqkfv9xyf6mudr...@mail.gmail.com... [...] > The NTP algorithm is much more complicated than the SNTP algorithm. The short, short version: there is no SNTP algorithm. SNTP is NTP _without_ the algorithms. Using NTP means continuously adjusting the speed of your clock so it tracks real time as best you can make it, while SNTP is simply asking what time [they think] it is. Groetjes, Maarten Wiltink ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: > In my case, I have* only one* network. My Time Server is not a machine, is a > meinberg GPS. In my point of view, if my source time were machines, maybe > NTP could be better to find a middle line between all these machines used as > time servers. But if I am using a very good and reliable GPS(Meinberg) with > a lot of satellites giving it the correct time, and it's pluged directly in > my switch, I think that in this case, NTP will not make any difference. If that is your opinion, you should not complain about network latency and clock rate errors. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: Rob, my understading about the use of SNTP and NTP is: while SNTP provides time synchronization within *one *network, NTP allows a global time You are confusing it with timed. SNTP also expects to use global time, and NTP can be used with an arbitrary timebase, providing all stratum 0s have the same arbitrary time (i.e., typically there is only one). synchronization on the internet. SNTP provides the current time, the current number of leap seconds and the warning flags marking the intriduction of a leap second correction. The NTP algorithm is much more complicated than the SNTP algorithm. NTP normally uses multiple time servers to verify the time and then controls the There is no SNTP algorithm, beyond basic validity checking. SNTP covers any use of NTP wire formats that falls short of being a compliant NTP implementation. SNTP implementations could still use PLL clock discipline code, or could use linear regression solutions for offset and rate. slew rate of the device. The algorithm determines if the values are accurate using several methods including fudge factors and identifying time servers that don't agree with the other time servers. It then speeds up or slows down the device drift rate so device time is always correct and there won't be any subsequent time jumps after the initial correction. SNTP usually uses just one Ethernet Time Server to calculate the time and then it "jumps" the system time to the calculated time. It can, however, One issue that has maybe not been stressed is that this strategy makes you extremely vulnerable to scheduling delays, e.g. you would likely never get a 1ms error bound on Windows if you used this strategy. Early W32Time implementations, and current ones in out of the box configurations are SNTP implementations, but often use internet sources. They do violate the specification by allowing multiple SNTP strata. have back-up Ethernet Time Servers in case one is not avaiable(not two at the same time). In my case, I have* only one* network. My Time Server is not a machine, is a meinberg GPS. In my point of view, if my source time were machines, maybe NTP could be better to find a middle line between all these machines used as time servers. But if I am using a very good and reliable GPS(Meinberg) with a lot of satellites giving it the correct time, and it's pluged directly in my switch, I think that in this case, NTP will not make any difference. I think you are being asked why you don't just use a standard ntpd for the platform, and also how you know that your total system meets the scheduling constraints for 1ms accuracy with SNTP. ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/16 Rob > Marcelo Pimenta wrote: > > The idea to have this is not overload the network asking time every > second > > to keep my accurancy in 1ms in 100% of time. And the second reason is > about > > how many ms your clock will be wrong in 59 sec without a frame to > discipline > > my clock again? PCs clock for example is not good, that's why I need a > very > > good crystal. > > Why are you causing your own problems by sticking to SNTP instead of > using NTP which was designed to overcome the problems you fear? > Rob, my understading about the use of SNTP and NTP is: while SNTP provides time synchronization within *one *network, NTP allows a global time synchronization on the internet. SNTP provides the current time, the current number of leap seconds and the warning flags marking the intriduction of a leap second correction. The NTP algorithm is much more complicated than the SNTP algorithm. NTP normally uses multiple time servers to verify the time and then controls the slew rate of the device. The algorithm determines if the values are accurate using several methods including fudge factors and identifying time servers that don't agree with the other time servers. It then speeds up or slows down the device drift rate so device time is always correct and there won't be any subsequent time jumps after the initial correction. SNTP usually uses just one Ethernet Time Server to calculate the time and then it "jumps" the system time to the calculated time. It can, however, have back-up Ethernet Time Servers in case one is not avaiable(not two at the same time). In my case, I have* only one* network. My Time Server is not a machine, is a meinberg GPS. In my point of view, if my source time were machines, maybe NTP could be better to find a middle line between all these machines used as time servers. But if I am using a very good and reliable GPS(Meinberg) with a lot of satellites giving it the correct time, and it's pluged directly in my switch, I think that in this case, NTP will not make any difference. > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
Marcelo Pimenta wrote: > The idea to have this is not overload the network asking time every second > to keep my accurancy in 1ms in 100% of time. And the second reason is about > how many ms your clock will be wrong in 59 sec without a frame to discipline > my clock again? PCs clock for example is not good, that's why I need a very > good crystal. Why are you causing your own problems by sticking to SNTP instead of using NTP which was designed to overcome the problems you fear? ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
2010/6/15 unruh > On 2010-06-14, Marcelo Pimenta wrote: > > Hi everybody!! > > > > My question is about Time Accuracy of NTP/SNTP protocol. I want to know > if > > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a > > network with 50 hosts(is the same with 100 hosts?) using only swithes, no > > routing. All these hosts are syncronized with a Meinberg GPS. The traffic > is > > low(~25%). Is it possible to have the same time accuracy of IRIG-B for > > example? > > I think you mean accuracy of 1ms. Yes. it can. On a local network, > 100usec is easily possible ( and depending on the routers, it can be > better than that). > Yes, you are right, accuracy of 1ms. On a local network 100usec?? Even if we use only switches(no routers), how is that possible if I have 4 types of Latency increasing about 80us? Algorithms to compensate the delay? Even in SNTP? Anyway, the question is about 1 ms. To get 100usec I'll use PTP. > > > > The internal clock in my devices(hosts) starts with the crystal > frequence. A > > typical offset or frequence tolerance of the crystal is less then > > 1ms/minute. The resolution is about 0.9ms and the devices acts as unicast > > clients, asking GPS every minute. > > Why? The whole purpose of ntp is to discipline your clock so that it > keeps time much better than that. > The idea to have this is not overload the network asking time every second to keep my accurancy in 1ms in 100% of time. And the second reason is about how many ms your clock will be wrong in 59 sec without a frame to discipline my clock again? PCs clock for example is not good, that's why I need a very good crystal. > Not sure what you mean by "the resolution is about .9ms" What is your > device? > I mean that my device can show me variations of 0.9ms, is the smallest possible increase of time the clock model allows. I work with protection IED(intelligent electronic device), is a term used in the electric power industry to describe microprocessor-based controllers of power system equipment. If something happens in 458ms, I will get exacly 458ms. Some equipment have so I can trust in variation that I can capture. > > > Is it enough? The Time Accurancy could be better if I could ask GPS twice > a > > minute? A broadcast implementation with GPS sending time every minute > could > > help? > > It depends on what you want the time for. > I need accurancy at least of 1ms in 100% of time. > > > > > I know that these kind of issue have many variables but the main question > > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? > > Yes, depending on the competence of the writer of the SNTP software. > So, the problem is not in the standard but into implementation of SNTP to not loose time to correct the clock and something like that? > > > > Thank you very much > > ___ > questions mailing list > questions@lists.ntp.org > http://lists.ntp.org/listinfo/questions > ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On 2010-06-14, Marcelo Pimenta wrote: > Hi everybody!! > > My question is about Time Accuracy of NTP/SNTP protocol. I want to know if > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a > network with 50 hosts(is the same with 100 hosts?) using only swithes, no > routing. All these hosts are syncronized with a Meinberg GPS. The traffic is > low(~25%). Is it possible to have the same time accuracy of IRIG-B for > example? I think you mean accuracy of 1ms. Yes. it can. On a local network, 100usec is easily possible ( and depending on the routers, it can be better than that). > > The internal clock in my devices(hosts) starts with the crystal frequence. A > typical offset or frequence tolerance of the crystal is less then > 1ms/minute. The resolution is about 0.9ms and the devices acts as unicast > clients, asking GPS every minute. Why? The whole purpose of ntp is to discipline your clock so that it keeps time much better than that. Not sure what you mean by "the resolution is about .9ms" What is your device? > > Is it enough? The Time Accurancy could be better if I could ask GPS twice a > minute? A broadcast implementation with GPS sending time every minute could > help? It depends on what you want the time for. > > I know that these kind of issue have many variables but the main question > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? Yes, depending on the competence of the writer of the SNTP software. > > Thank you very much ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
Re: [ntp:questions] SNTP with 1ms of precision?
On Mon, Jun 14, 2010 at 3:03 PM, Marcelo Pimenta wrote: > > Hi everybody!! > > My question is about Time Accuracy of NTP/SNTP protocol. I want to know if > is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a > network with 50 hosts(is the same with 100 hosts?) using only swithes, no > routing. All these hosts are syncronized with a Meinberg GPS. The traffic is > low(~25%). Is it possible to have the same time accuracy of IRIG-B for > example? > > The internal clock in my devices(hosts) starts with the crystal frequence. A > typical offset or frequence tolerance of the crystal is less then > 1ms/minute. The resolution is about 0.9ms and the devices acts as unicast > clients, asking GPS every minute. > > Is it enough? The Time Accurancy could be better if I could ask GPS twice a > minute? A broadcast implementation with GPS sending time every minute could > help? > > I know that these kind of issue have many variables but the main question > is: Is possible to have precision of 1ms(could be <1ms) with SNTP? > > Thank you very much Yes. No. Maybe. A lot of factors come into play, such as the characteristics of the switches, the characteristics of the network cards, the characteristics of the load on the network and servers, etc. There are no guarantees with the NTP protocol. I will report that I *typically* see offsets of <0.5 ms on a network of ~20 newer Dell and HP servers connected to a pair of Cisco 3750s via Gigabit Ethernet. The Stratum-1 source is CDMA (essentially GPS). But during nightly backups or during periods of intense load, I see excursions out to 5 ms or more from the perspective of a monitoring system on the same network. This behavior more than meets the needs of my applications, but might not meet the needs of yours. -- RPM ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
[ntp:questions] SNTP with 1ms of precision?
Hi everybody!! My question is about Time Accuracy of NTP/SNTP protocol. I want to know if is possible to have precision of 1ms(could be < 1 ms?) using SNTP in a network with 50 hosts(is the same with 100 hosts?) using only swithes, no routing. All these hosts are syncronized with a Meinberg GPS. The traffic is low(~25%). Is it possible to have the same time accuracy of IRIG-B for example? The internal clock in my devices(hosts) starts with the crystal frequence. A typical offset or frequence tolerance of the crystal is less then 1ms/minute. The resolution is about 0.9ms and the devices acts as unicast clients, asking GPS every minute. Is it enough? The Time Accurancy could be better if I could ask GPS twice a minute? A broadcast implementation with GPS sending time every minute could help? I know that these kind of issue have many variables but the main question is: Is possible to have precision of 1ms(could be <1ms) with SNTP? Thank you very much ___ questions mailing list questions@lists.ntp.org http://lists.ntp.org/listinfo/questions
