What's a sensible MTU?

What's a sensible value to set for the MTU of a router, running at around 3M bps downstream rate, under Windows XP? With a new router, I've been noticing that, sometimes, one or two websites I visit daily become unobtainable; that is, they don't resolve and the browser (IE7) simply gives up, stating that it can't get the website. Other websites are fine, though. I suspect this is an MTU problem of some sort. However, changing the router's MTU value from its default of 1458 has had no effect.

I've performed a ping fragmentation test in Windows (pinging a normally reachable popular website) and found that the Windows MTU was sometimes 1430, sometimes 1432. I gather that Windows XP sets its own MTU automatically and dynamically.

So knowing that the router's MTU should preferably equal the Windows MTU, or be slightly greater, I tweaked the router's MTU up slightly, to 1460 (1432 + 28). This seems to have had no effect, though.

I'm assuming, of course, that the router's firmware is bug-free in this respect.

Do I actually need to increase the router's MTU a lot further beyond 1460?

1500 works for me.
Sometimes it depends on your ISP (but I remember you don't wish to say who they are) so it might be worth asking them.

Your problems could also be DNS related.

Yes, I've seen it written (vaguely so) that sometimes it can be due to the ISP. However, in this case, the non resolving of particular websites started only after getting the new router.

What would have to be wrong or ill-configured in either Windows or the router to make it a DNS-related problem?

the MTU should be set to 1500 as this is the ethernet and windows default.

If you want to run a smaller MTU set it on the PC to avoid potential fragmentation issues in the router. Lack of access to secure sites (inc Ebay etc) is often blocked if packet fragmentation results from the router MTU being less than that of windows.

The occasionally-unobtainable websites I'm encountering aren't secure sites, and the router's MTU has never been less than Windows XP's MTU, so I think we can rule that one out. But I wasn't aware that MTU was also controllable in Ethernet. Clearly, optimised packeting needs to take place at the Ethernet level also, but is an Ethernet MTU actually user-configurable? If so, how? And would the Ethernet MTU value depend on whether the connection was wired or wireless?

It was my impression that nowadays, with Windows XP and 2K, the operating system automatically sets its own optimised MTU. Thus, although there may be a Windows default value of 1500 when the OS is first installed, the value that Windows uses adapts to the type of Internet connection being used. In fact, this appears to be supported by the fact that my ping test gave a Windows value (without the packet header overhead) of about 1430.

Perhaps the best approach should be for me to just gradually bump up the existing router MTU toward 1500, from 1460, and see if and at what point the non-resolving of those few particular websites disappears? It could take me a while, though, as on some days the problem is there, on other days it isn't.

Edited by deleted (Tue 17-Nov-09 13:59:14)

Seems more like a DNS problem, which could occur with any ISP, rather than an MTU issue. Why not try OpenDNS and see if it goes away?

the best approach is to just set the router's MTU to 1500. I've done this to other Netgears and the "page won't display" issue goes away.

If you use ping tests to find the MTU at which it breaks you are merely finding the lowest MTU in the system and not measuring the Windows MTU. If path MTU discovery is enabled and working then XP is supposed to reduce its MTU.

if you haven't changed the setting then Windows is at 1500 MTU. DrTCP is a tool for changing this. Wireless uses a bigger packet "over the air" to cover the overheads.

In reply to a post by meditator:

...Perhaps the best approach should be for me to just gradually bump up the existing router MTU toward 1500...

OR just try 1500 as I and others have mentioned???

Previously you mentioned you talk to your mystery ISP "I interact with their professional engineers" - it would be interesting to hear what they say...

In reply to a post by yarwell:

the MTU should be set to 1500 as this is the ethernet and windows default.

If you want to run a smaller MTU set it on the PC to avoid potential fragmentation issues in the router. Lack of access to secure sites (inc Ebay etc) is often blocked if packet fragmentation results from the router MTU being less than that of windows.

Mmmm........... I remember tweaking the MTU on my computer using the advice of a well-known American site, and when I finished, I was unable to send out any emails !

"If it ain't broke, don't fix it"

Yes, zebedee, I had much the same experience back in the days when I was using Windows 2000. Using Dr TCP caused me more problems than it solved. Nowadays, though, I gather that Windows XP automatically adjusts its own MTU to an optimum value, so the use of an app such as Dr TCP (there are several others around, I think) should not be necessary.

The issue I'm trying to resolve here, though, is not the Windows value for the MTU but the router's value. With what little I've learned about MTU, the router's value needs to be at least the same value as that of Windows's. If that condition's not satisfied, then either the packets will tend to get fragmented or packets will not be used efficiently. In one extreme, packets can be completely dropped and the website become unobtainable, which is precisely what I'm occasionally observing.

TT router MTU 1432

That's right for TalkTalk - unfortunately we don't know the OP's ISP...

b4dger,

Less of the sarcasm, please!

Yes, I've just been in touch with my ISP about this. Their view is that the default for the Windows value is 1500. They've also mentioned that Path MTU Discovery may not work correctly, in so far as deriving the smallest overall MTU for efficient working, if ICMP is disabled.

ICMP is disabled, in my case, but for reasons of security I'm loath to permanently enable it.

What are they recommending you try, given your circumstances?

EDIT: One of the reasons open forums can be useful is so that people can search for
'mtu yourisp' etc. etc.

Edited by b4dger (Wed 18-Nov-09 12:49:02)

You've a good point there, yarwell. Yes, when you think about it, doing that ping test is merely finding the lowest overall MTU in the system.

Do you happen to know the path for the Windows MTU value in the Windows Registry? I could at least go there then and see what value my XP is currently using.

They're not recommending any one particular value; all they've said is that the default value for the Windows MTU is 1500. As I've already pointed out as well, they've said that automatic adjustment may not function properly if ICMP is disabled, thereby preventing the website server from pinging the router. They may comment further, in which case I'll post anything of note here.

Postscript: Well, what the contact at my ISP said was not quite what I wrote above. So, to clarify, here's precisely what the contact said:

"On the router I'd suggest 1500, as this is the Windows default.

Path MTU discovery should mean the MTU is automatically sized to the
smallest MTU setting on the path between your PC and the remote web
server, but it's possible that this may not be working correctly -
particularly if you have disabled ICMP on your router".

Edited by deleted (Wed 18-Nov-09 15:02:42)

Your test conclusion and logic is flawed.

Pinging a remote server to try and discover the Windows MTU is not going to work. You are actually discovering the maximum payload possible without fragmentation over the entire route to that server.

You should set the MTU on the PPP interface to a value that prevents fragmentation. Ideally you should also set the same MTU on the LAN interfaces (ethernet / wireless). If using Cisco kit a simpler solution is to use the mss-adjust setting on the LAN side to force the packet size down for the whole ethernet segment (your LAN).

As you are seeing maximum possible packet sizes of 1430 or 1432, this suggests both an ethernet backhaul of some description from the exchange and that your PPP session is further tunnelled within your ISPs network - most likely an L2TP tunnel between LAC and LNS. Unless of course you running your connection through some kind of VPN which will introduce its own overheads. You most certainly don't want to be increasing the MTU on your router.

There is much discussion about optimal IP MTU for DSL circuits, a common misconception being that having it as large as possible is best.

A more theoretical way to look at this is to consider the various transport layers that your connection uses and optimise your payload for the most heavily utilised. In the case of a typical DSL connection to the internet the bottleneck is clearly the DSL section (using ATM), so you should try to optimise for that to maximise throughput.


SO where do we start...


ATM Cell size is 53 bytes. However 5 bytes of that is used for the header leaving 48 bytes for payload.

The IP data payload your PC (etc.) generates is encapsulated into the PPP session adding 2 bytes of header to every packet.

To optimally use the ATM part of the connection (from DSLAM/MSAN to DSL CPE) you really want to fill every ATM cell completely. Bear in mind that ATM will pad the final cell for a packet if the packet does not completely fill the cell, rather than take the 1st chunk of the next packet and incorporate it.

So if your IP + PPP packet size is not divisible by 48 then you are effectively introducing artificial overhead or waste in that some ATM cells are not fully loaded.

So some examples:
IP MTU = 1400 + 2 bytes PPP = 1402
1402/48 = 29.21 (to 2 d.p.)


IP MTU = 1430 + 2 bytes PPP = 1432
1432/48 = 29.83 (to 2 d.p.)


Neither are optimal for PPPoA, the closest would be 1440 (1440/48 = 30). However, you have already ascertained that your connection can't support that without fragmentation.

So 1390 is really the highest IP MTU optimal for the PPPoA transit in this case. (1390+2)/48 = 29.


Of course this is all highly theoretical and assumes that you are going to be sending/receiving maximum sized IP packets the whole time.

Whether in practise this would result in higher throughput is questionable. On the one hand the DSL link is the bottleneck in your end to end connection to wherever you are pulling data from, so optimising that makes sense. The trade off is that you're having to generate and process more packets at the IP layer. Whilst the core fabric of the internet will have no issue churning these extra packets back and forth, your equipment might - having a decent ethernet adaptor will mitigate this in most cases.

You will still have to contend with XP's rather flaky IP stack and it's discovery methods of course.

All food for thought, not that anyone will listen of course... everyone seems to be convinced that bigger is always better when it comes to MTU.

GeeTee,

Most interesting to get an alternative view. I do kinda agree with your contention that biggest is not necessarily best, because what I remember, when I last visited the issue of MTU some several years ago now, is that that's precisely what was concluded by the pundits of the time.

Getting MTUs optimised is by no means straightforward, it seems. Even the article on it, on kitz.co.uk, gives many different examples of Windows MTU, each apparently working fine for the individuals concerned. Sure, 1500 is a good starting point for the Windows value but it won't necessarily be an efficient value for Windows to be using. Indeed, as far as I can gather, Windows XP will automatically try to optimise the value.

On your treatise posted here, although you've arrived at an MTU as low as 1390, you still seem to be of the view that this is only theoretical and so I'm still thinking that perhaps, in practice, my best bet might be to just try some values. Incidentally, I didn't think that 48 was the divisor, since most Windows MTU values I've seen used are 1430, 1458, 1478 and 1500. But perhaps I'm not understanding precisely in which part of the end-to-end connection you're considering.

My memory's not so good these days but I thought that Windows stored its MTU value in the Windows Registry. I've looked for it on my machine and have not been able to find it. I tried HKey Local Machine\System\Current Control Set\Services\TCPip\Parameters\Interfaces, and then tried looking in all of the adapter IDs listed there. But there were no DWords for an MTU anywhere there. I stress that I don't want to alter the Windows MTU, I just want to find out what its current value is, so that I can see if the router's value is correspondingly correct and, if it isn't, I can then alter it in the router's GUI.

Edited by deleted (Wed 18-Nov-09 14:35:03)

In reply to a post by meditator:

Yes, zebedee, I had much the same experience back in the days when I was using Windows 2000. Using Dr TCP caused me more problems than it solved. Nowadays, though, I gather that Windows XP automatically adjusts its own MTU to an optimum value, so the use of an app such as Dr TCP (there are several others around, I think) should not be necessary.

I don't think you're right there. As far as I'm aware, only Vista and Windows 7 dynamically alter the MTU, not XP.

In reply to a post by meditator:

I thought that Windows stored its MTU value in the Windows Registry.

I'm guessing it will only be there if it's not left to default.

Default and Manually changing: http://support.microsoft.com/kb/826159

Well, I've no concrete proof that WinXP automatically adjusts. It's just that it kinda sticks in my mind from way back. You could well be right.

Thanks for that link to the Microsoft KB article on MTU. I notice, though, that that KB article applies only to the Pro version of WinXP.

Zebedee,

Googling, I've found this:

http://setiathome.berkeley.edu/forum_thread.php?id=2...

It dates back to 2005 but nevertheless it appears to say that Windows XP does work out and set its own MTU. If you change the Windows MTU value, Windows will ignore that change, anyway. (I do vaguely remember Win2K reacting like that, when I was using Win2K some years ago. I ended up getting in a bit of a mess, trying to use Dr TCP to change it, only to find that the Dr TCP value was always ignored). So unless WinXP has undergone changes in that dept since then, it does look as though I was correct.

Well, maybe they're right - who knows ?

One good piece of advice in there anyway - turn MTU path discovery ON.

Yes, that'd be fine, but there's no facility in my WinXP's Registry for doing that. I've tried looking in a number of different likely places for that setting in the Registry but, in WinXP Home, an MTU Discovery folder of any sort does not seem to exist. That could mean that the Windows MTU is indeed fixed at 1500 (and therefore there should never be a need to alter it) but, equally, it could mean that the Windows MTU does change dynamically (and so, again, there's no point in manually changing it). Possibly, things are done differently in WinXP Pro.

As I've pointed out, I'm not actually wanting to change the Windows MTU, I'd just like to know what current value it's using on my machine, so that I can then judge the corresponding value to configure in the router.

You rarely see the divisor of 48 in discussion about IP MTU because this value is entirely specific to IP data encapsulated in a PPPoA session being carried on an ATM transport layer. I'd hoped my explanation made the reason for using that divisor clear.

Most MTU discussions are based on information gleaned from articles found littering the tinterweb which primarily revolve around optimising MTU for IP transit over Ethernet.

Based on the fact you supplied that you experience fragmentation with an MTU anything over 1430, why would you be considering trying values greater than that?

It is basic maths that determines that 1390 is the highest IP MTU you can run IF you want to fully utilise the DSL part of your end to end connection at the same time as avoiding fragmentation. No theory there, just simple numbers.

The theoretical aspect refers to whether trying to optimise the ATM transport in this way will result in any real world benefit in terms of increased throughput. In theory it should. In my own experience it does. Will it for everyone - most likely not as other factors come into play.

I think you are also heading down a deadend looking for what XP has your MTU set to currently in order to determine what to set your router too. ~correction here~> XP automatically configures its RWIN value not its MTU so unless you have configured something different it will be at its default of 1500.

edit - for correction

Edited by deleted (Wed 18-Nov-09 18:48:02)

Path MTU Discovery is fundamentally broken on the internet as a whole due to the number of core routers and devices that either ignore it or actively drop the ICMP traffic related to it. It's worse than useless.

The MTU limitation you have is going to be on your PPP connection to your ISP. Worry about not fragmenting over that section of your connection and everything else will be fine.

In reply to a post by meditator:

Yes, that'd be fine, but there's no facility in my WinXP's Registry for doing that.

HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters

EnablePMTUDiscovery ....... set to 1. (default value)

Which version of Windows are you using, zebedee? In my version of WinXP, there is no such parameter as EnablePMTUDiscovery in that location.

It's present in Windows Media Centre Edition, built upon XP Pro; can't speak for XP Home.

Tho' I've never seen my Win's MTU value changed from its TCPOptimizer-set value of 1454.

Edited by XRaySpeX (Thu 19-Nov-09 00:33:23)

Let me emphasise, GeeTee, that although I appreciate one or two basic things about MTU, I don't consider myself an expert in this area. After all, why would I be posting this thread if I already knew all the basic answers?

Let me repeat what I described much earlier, though. The value of MTU at which a ping to a website of my choice returned non-fragmented packets was 1432. Any value higher than that caused fragmentation. So yes, you might well ask why I've been thinking of configuring my router to a value higher than 1432 + 28 = 1460. Well, the only reason is that everyone else has been telling me that I should set my router's MTU to be the same as Windows's MTU, that being notionally 1500 (1472 + 28). On the basis that that 1500 is the maximum MTU value and not necessarily the actual MTU in use, it's seemed logical to try router MTU values of between 1460 and 1500 - since the golden rule is to get the router's MTU to be either exactly the same as Windows's MTU, or to have it slightly larger; packets then don't get fragmented as they pass through my router. Apparently, you should, in any event, never go above 1500. But maybe this is screwy logic?

The author of the article on MTU at kitz.co.uk publishes a long list of empirical Windows MTUs, obtained from various (perhaps unreliable) sources, and none of them are anything as low as 1390.

By the way, do you mean actually 1390 + 28, as the MTU for Windows?

I've reflected on what you say about RWIN and I think you're possibly correct about Windows's MTU value not automatically adapting, but the RWIN value adapting instead. So yes, maybe I got that misconstrued.

You may also be correct about any number of routers between my PC and the website blocking Path MTU Discovery, but certainly what my ISP has eluded to could alone be the cause of the overall MTU in my setup not being optimum at present, namely that if the website cannot determine from my end the most suitable packet size to use, the website may well decide to simply allocate the Windows default size of 1500 and, with my router's MTU currently being at present only 1460, a mismatch will occur. It may well be that the only way to allow the website to use Path MTU Discovery is if ICMP at my end is enabled (allowing pinging of my router by websites). However, enabling ICMP (at least, on a permanent basis) raises a potential security issue. Thus, maybe I'm left with a strict choice between either leaving ICMP disabled and accepting that the website will just throw 1500-size packets at my PC and a good deal of them getting fragmented, or that I enable ICMP and thereby throw away the stealthed nature of my connection. Either way, though, I'd still need to know what value for MTU I should set in my router. I'm still not convinced that 1500 is the correct value to use. And 1390 + 28 also seems unlikely; if the Windows MTU were that value, then with 1460 being the value currently in my router, why would I be having problems still?

Surely, the way I'm running my ping test at present is valid? It's involving all elements of a connection between my PC and a remote website, so it's very much a realistic scenario. If it's giving me the answer 1432 (+28) as being the optimum unfragmented end-to-end size, then what can be gained by setting the router's MTU to something that's considerably lower than that? Or is that that particular point of yours is more about more-completely filling the packets than avoiding fragmentation? (In one respect, this is a 'chicken and egg' situation, isn't it?).

Edited by deleted (Thu 19-Nov-09 00:52:55)

what was that tweak then.

if you do the following.

set the mtu on router to 1500 and disable any mtu clamping type feature.
set the mtu on the pc to 1500 or below and keep mtu discovery enabled.

you should never be in a situation where you dont have the ability to send data (assuming no other fault), maybe reduced performance if the 1500 isnt optimal but not no connectivity.

windows xp still defaults to 1500, meaning if you tweak to 1500 you actually changing nothing.

the default in xp sp3 is to have a 1500 mtu, icmp mtu discovery enabled and blackhole mtu discovery enabled. The default mtu changes if you are not using an ethernet connection from the pc, such as dialup modem or wireless.

if you set to 1500 and turn off mtu discovery, windows then ignores the setting and forces a 576 mtu value.

Edited by Chrysalis (Thu 19-Nov-09 08:49:10)

In reply to a post by meditator:

Which version of Windows are you using, zebedee? In my version of WinXP, there is no such parameter as EnablePMTUDiscovery in that location.

I'm using XP Professional SP3.

I have to tell you, zebedee, that EnablePMTUDiscovery, or viewing an MTU value in the Registry, is not available in Windows XP Home, which happens to be what I'm using (that's with SP3).

Chrysalis,

My guess is that your advice here is based on the use of a utility such as Dr TCP. Furthermore, I reckon that, along with several other contributors, you're assuming that both versions of WinXP (Pro and Home) can be dealt with in the same way. I'm afraid that isn't the case, as no MTU or MTU Discovery settings are made available in the Windows Registry in WinXP Home (which is what I have). There's no way that I can alter the Windows MTU, even if I wanted to and, as you've realised yourself, Windows would ignore any new setting for MTU anyway.

It appears that, when you run under WinXP Home, you've fewer options at your disposal when dealing with this issue. I'm beginning to conclude that probably the only sensible thing I can do is to simply assume that the Windows MTU is indeed 1500 and so configure my router's MTU to this same value. I think yarwell reckoned it worked okay with Netgear routers.

Edited by deleted (Thu 19-Nov-09 11:56:31)

FWIW As I mentioned in the first reply in this thread 1500 works for me!

Why not give it a try? You can always change it back...

Edited by b4dger (Thu 19-Nov-09 13:09:03)

Yes, as my slightly earlier reply indicated, I've now changed my router MTU to 1500. Thus far, no problems.

When I now ping a remote website with an MTU value of anything greater than 1472, it fragments. (Before, the value was 1432). I guess this agrees with Windows's MTU and the router MTU both being 1500 now, ie. 1472 + 28 = 1500.

In reply to a post by meditator:

My guess is that your advice here is based on the use of a utility such as Dr TCP. Furthermore, I reckon that, along with several other contributors, you're assuming that both versions of WinXP (Pro and Home) can be dealt with in the same way.

You're assuming an awful lot ! You asked how to switch path discovery on in XP, and you were told how to do it. We have no way of telling which version of XP you are using, and no way of knowing that you are using a version of XP which is not really designed for people tinkering with the TCP/IP settings. It's rather like buying a Ford Kaa, and complaining that you can't find the switch for the Xenon headlamps

yes as yarwell said never lower your router mtu, always control it at the end point (usually the pc).

regarding your xp home comment, then that is confusing. I have never heard of such a limitation and even just now a quick google shows nothing indicating xp home has no ability to adjust mtu values or the mtu discovery setting. By the way the mtu values are not there by default even in XP pro, they have to be created.

Edited by Chrysalis (Thu 19-Nov-09 14:30:36)

I'm sure that if you use something like TCPOptimizer ( http://www.speedguide.net/downloads.php ), it will create the relevant MTU entries in the Registry even for XP Home as it probably did for my XP MCE/XP Pro. It's spec. does not exclude XP Home.

Edited by XRaySpeX (Thu 19-Nov-09 15:21:20)

Come on now, zebedee, there was no way that I or any other punter could possibly have known that these Registry settings weren't available, when we purchased WinXP Home! The same applies to lots of other features of Windows. Many of us who bought the Home version did so because we didn't think the extra cost of the Pro version was worth it. Several of my ex-colleagues who work in the IT field bought Home machines for they and their families' use at home precisely with that in mind. It's just unfortunate that, over the years, we've encountered one or two annoying limitations of the Home version.

TCP Optimiser? Now, that rings a bell. I think I've also used that, in the dark and distant past, with Win2K. My memory of it is very vague now but I've a feeling that you're right, that it does create entries in the Register. However, my doubt would be whether it actually works with WinXP Home. I think it's maybe useful for literally creating a value that you choose but is no use if what you want to find out is the current unmodified default value in the Registry - because if it ain't there, you can't get it!

Anyway, I'm happy with how things are now, so there's no need to contemplate such things. Thanks for reminding me, though.

I read somewhere, maybe in this thread, that if it ain't there, then it defaults to 1500.

TCP Optimiser is useful for calculating an optimal value, not just a value of your choice; it takes the hard work outta all those repeated Pings that you may have wot of!

In reply to a post by meditator:

Yes, as my slightly earlier reply indicated, I've now changed my router MTU to 1500. Thus far, no problems.

Sorry I didn't/can't see that post? I can see you mentioned "I'm beginning to conclude..." but can't find any reference where you actually changed things.

People here are trying to help - but it's not always easy

Many Windows inbuilt default values have no actual matching registry entry unless the default is changed, at which time the appropriate registry value is created, either manually or by an application such as TCPOpitimzer.

TCPOptimzer has a Registry tab which shows the actual registry values for anything that does have a default registry entry. It also has an MTU tab that allows you to to a run a fragmentation test on any site you enter. It also has a latency test tab and BDP calculator to help your optimise your connection for your commonest sites. Though the makers of TCPOptimizer also have a test page on their site www.speedguide.net that does much the same thing for you and will give you suggested values for your PC and its connection.

that EnablePMTUDiscovery, or viewing an MTU value in the Registry, is not available in Windows XP Home

that's what DrTCP does, it allows you to see/change/insert the relevant values. Works here on XP home.

there isn't a registry entry for MTU unless one has been created, the lack of one means it's using 1500 coded in somewhere. DrTCP shows it up.

Yes, my thanks once again go to those who've reminded me in this thread of both Dr TCP and TCP Optimiser. I have used them in the past, in the days when I was using Win2K, I think. In the end, I gave up using them, as they didn't help for what I was trying to do at the time.

As for now, what I've been trying to do is (a) find out from others what they think is a fairly optimised value for Windows MTU, so that I can appropriately set the MTU of my router, and (b) if possible, to look in Windows Registry and read the current Windows MTU. Several of you have been kind enough to recommend the value 1500 to me and consequently I'm now using that value. I've not, however, been able to see a default MTU value in the Registry, as it's now apparent that, in WinXP Home edition, no such data is shown there. This doesn't matter, though, as I'm content to use the value that everyone's recommended.

Just out of interest, are there any of you using WinXP who are not using the default MTU of 1500? If so, which value are you using, and does it definitely cater for all the websites you use?

Edited by deleted (Fri 20-Nov-09 15:52:39)

Just out of interest, are there any of you using WinXP who are not using the default MTU of 1500?

turned out I was using 1470 as the machine had been running over a Sharedband link aggregation system. As it was set on the PC I had no issue accessing sites as the router was at 1500.

Problems only arise when a PC fires bigger packets than the hole in the router allows to pass.

"Problems only arise when a PC fires bigger packets than the hole in the router allows to pass".

Yup, understood. Prior to starting this thread, my Windows would have been using 1500 and my router was using 1460. (Both figures include the 28). But now they're both 1500.

Just out of interest, are there any of you using WinXP who are not using the default MTU of 1500? If so, which value are you using, and does it definitely cater for all the websites you use?

The Netgear MTU setting is 1500

XP Home, MTU setting using TCP Optimiser is 1430

Tcp WindowSize=51480

Seems to be OK with the websites I've visited so far.


Edit >>> I'm using a fixed speed 2Mbps connection.

Edited by Apprentice (Fri 20-Nov-09 17:19:11)

In reply to a post by meditator:

Just out of interest, are there any of you using WinXP who are not using the default MTU of 1500? If so, which value are you using, and does it definitely cater for all the websites you use?

Yes, as I indicated earlier, I am using MTU=1454 calculated into the Registry by TCPOptimizer. I never have any problem with any communication with any Internet destination (it's not just Browsers that it applies to but all Net programs).

My router is set to MTU=1492, probably a hang-over from when Win XP was tuned for my previous USB Modem, when TCPOptimizer calculated MTU=1492 for XP. Looks like I never bothered to reset the MTU in the router.

EDIT: Oh, BTW, my TCPWindowSize as calculated by TCPOptimizer is the largish 520352.

Edited by XRaySpeX (Fri 20-Nov-09 17:33:51)

OK - so now both router and PC are running at 1500 - try doing your unfragmented ping test again and see if you can pass 1500 byte packets without fragmentation.

I suspect the answer will be no.

e.g. something like this.....

H:\>ping -f -l 1500 www.thinkbroadband.com

Pinging www.thinkbroadband.com [80.249.99.130] with 1500 bytes of data:

Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

Ping statistics for 80.249.99.130:
Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

I suspect the answer will be no.

of course it will, because to test a 1500 MTU you need to ping with a 1472 packet !

Indeed so. My bad, been a long week.

I'm just interested to see what results the OP gets now everything is aligned.

In reply to a post by meditator:

"Problems only arise when a PC fires bigger packets than the hole in the router allows to pass".

Yup, understood. Prior to starting this thread, my Windows would have been using 1500 and my router was using 1460. (Both figures include the 28). But now they're both 1500.

So, basically you are back to where you were before you started fiddling with it. You should try more rewarding hobbies, like building the Taj Mahal out of match sticks

I've already done that, GeeTee, and it does pass all the packets unfragmented.

As yarwell points out, the ping value for that is 1472.

No, you've got it wrong, I'm not back to where I started at all. I started with the router MTU being 1458. An initial tweak to 1460 made no difference and, at that point, I began this thread. The router's MTU is now 1500 (which includes the 28 overhead).

something I should add here is that some routers will clamp the mtu. eg. on the billion I think its called tcp clamping.

Whenever this kind of feature is available on a router I would advise to turn it off, just keep it at a unclamped 1500 setting and any adjustment (if required) to be done on the pc.

In reply to a post by meditator:

I've not, however, been able to see a default MTU value in the Registry, as it's now apparent that, in Win XP Home edition, no such data is shown there.

But ANY version of Windows, not just XP Home, will only have an MTU registry entry if the default value has ever been changed. Thus if ANY version of Windows has no MTU registry entry then it will have the default value of 1500.

In reply to a post by jcym:

In reply to a post by meditator:

I've not, however, been able to see a default MTU value in the Registry, as it's now apparent that, in Win XP Home edition, no such data is shown there.

But ANY version of Windows, not just XP Home, will only have an MTU registry entry if the default value has ever been changed. Thus if ANY version of Windows has no MTU registry entry then it will have the default value of 1500.

And it follows that possibly the MTU path discovery wouldn't be there unless changed either ?

Not a bad thing, when you think about it, otherwise the registry would become very bloated with entries full of default values.

correct, there is a LOT of registry entries that do not exist by default. The code in windows will use a default value when they dont exist.

of the top of my head there is 20-30 registry keys inside tcp/parameters that do not exist by default but are configurable. Of course that is just one tiny part of the registry.