Question About Synology's SMB3.0 Muti-channel, check #17 for solution

[FAhentai]

Hi guys.
I have a Synology DS918+ that uses USB3.0 2.5G nic with modified driver. Recently I have realized that Synology starts to support smb3.0 muti-channel officially, thus I had a try. In conclusion, I would say that it worked, yet not as good as I expected.
Here is the details:
My Ds918+ has total 4 Disks, 2X8T and 2X16T, 2x8t is in raid 1 mode and there is no raid for 2X16T. All of those disks are Seagate Ironwolf (not Ironwolf Pro) and both of the 16T one has about 40% space left.
My current network system is NBN BOX(modem)-----PPPOE to Primary router-----Asus AX89X------QNAP Switch, QSW-2104-2T(connect to router with 10G RJ45)------Primary computer with 2.5G ethernet connection, meanwhile DS918+ is connected to ax89x's LAN 1 and LAN 2, with the IP address next to each other. The test result seems like, well, it worked but not 100% worked. When I was copy files from Nas, the speed stays at 220MB/S，it is very stable. However, when I try to write files to NAS, the speed is only 120-130. The peak speed is only 150MB, while when I was using 2.5G, this speed is between 160 to 220.
Later I thought it could be the switch's problem, so I connect my computer directly to AX89X's 10G RJ45, I found the result is almost the same.
However, today I had another try. I happen to have 2X2.5G USE3.0 NIC, thus, I pluged both of them onto my lap top and had a try, with 2X2.5G NIC connect to the switch. I found out that no matter if it is reading or writting to NAS, the speed stays at 220M. However, with one of the NIC removed. The reading speed droped to 160-180, and writting spead droped to 140MB.

Now I am totally confused, thus, here are my questions
1. When you are doing SMB3.0 Muti-channel asymmetrically, Eg your Nas uses 2X1G and your computer uses 1X2.5G, is there always some sort of performance lose comparing with symmetrically (Eg 2X1G from your Nas and 2X either muti-gig or 1G connection to your computer)? I also watched some review, it seems like QNAP and Ugreen Nas had the same problem.
2. Why my primary computer uses 2.5G ethernet just like my laptop, there is such performance difference?

 

[lantis3]

How does SMB 3.0 multiple channels work depends on a lot of factors.

https://barreto.home.blog/2012/06/2...a-feature-of-windows-server-2012-and-smb-3-0/

https://kb.synology.com/en-nz/DSM/tutorial/smb3_multichannel_link_aggregation

HDD speed = 220 Megabytes (MB) = 1,845,493,760 Bits < 2 x 1Gbits < 2.5 Gbits,
your HDD speed can't saturate 2x1Gbps ethernet or one 2.5Gbps ethernet

140MB = 1120 Mbits, 1120Mbits x 2 = 2240 Mbits< 2.5Gbits

1. USB NICs usually don't perform well (especially cheap ones)
2. Which NAS disk set did you use for copy from/write to?
3. Read/write performance depends a lot on where files are located on the disk and how fast the read/write head can access them, it's mechanical. The max number claimed by vendor only happens in optimal condition like read/write big sequential files.

 

[FAhentai]

How does SMB 3.0 multiple channels work depends on a lot of factors.

https://barreto.home.blog/2012/06/2...a-feature-of-windows-server-2012-and-smb-3-0/

https://kb.synology.com/en-nz/DSM/tutorial/smb3_multichannel_link_aggregation

HDD speed = 220 Megabytes (MB) = 1,845,493,760 Bits < 2 x 1Gbits < 2.5 Gbits,
your HDD speed can't saturate 2x1Gbps ethernet or one 2.5Gbps ethernet

140MB = 1120 Mbits, 1120Mbits x 2 = 2240 Mbits< 2.5Gbits

1. USB NICs usually don't perform well (especially cheap ones)
2. Which NAS disk set did you use for copy from/write to?
3. Read/write performance depends a lot on where files are located on the disk and how fast the read/write head can access them, it's mechanical. The max number claimed by vendor only happens in optimal condition like read/write big sequential files.

You can not just simply do the math like this mate. Synology is a little bit more complex than that. I give you an example, when I was using 2.5G USB NIC on Synology, as long as the files is less than 6G, when I try to write file from my primary computer (All the disk inside my primary computers are SSD, either SATA3 or M.2 PCI-E GEN4, total about 15.5T, I must be mad） to the NAS, the speed will always stay at 280MB/S+, why? Cuz it goes to Synology's ram first (My DS918+ has 8G RAM), the speed 220-160 only happens when I try to transfer something larger than that. And 99.999% of files that I transfer to my nas is my movies, so yeah, they are all large files that require no 4K reading and writting. While for the reading, it is different, cuz it always goes to HDD first so that it can only go somewhere near 220MB-230MB, regardless I am using 2X1G or just 1X2.5G cuz it is restricted by HDD speed.
This time when I did the test, I just paste and copy some movie file that is less than 6G to my computer and then doing the opsite way and had a comparation with using 2.5G. What I found out is that when I was doing the reading from NAS, it seems normal, cuz 220M/S is the best a HDD can do. However, writting to NAS is not the case, it seems like just 10%-20% better than 1G ethernet. However, later when I tried to use my laptop (all disk inside are SSD), which is like what I said above, 2XUSB2.5G NIC that connected to the switch at one time and doing the same test, the problem is gone, no matter if it is reading or writting, they both work at 220M/S, while the only difference is when I just use one of the NIC to my laptop, it seems like the reading speed dropped while the writting speed increased. If it is just a few MB I would say that is normal, however, the diference is already like 30MB+: when I was using the lap top with one 2.5G, its reading speed dropped about 50MB while the writting speed inscreased 30MB.


Also, there are something on top of this. Of course, my current NAS, DS918+ is no where near capable with my expectation of future. However, for Synology, almost any model that has 5+ bays, it will usually have 4X1G ports. Meanwhile, if you check Aliexpress, you will found that those 2X10G spf+ and 4X2.5G switch is cheap, so cheap that you can not even believe, like this one https://www.aliexpress.com/item/100...UqZRJI1N&utparam-url=scene:search|query_from:
not even 50 Aud get delievered. Comparing with going to 10G (in fact some of those models can not even go 10G, eg DS1517+) , this could be more cost effective. However, the writting spead problem really turns people down

 

[lantis3]

Did you open task manager and watch ethernet performance for NICs and individual CPU core usage?

Do your SSD/NVMe drives on PC/laptop have dram or they are dram-less?

View: https://youtu.be/0rFL8PGkmMU

 

[FAhentai]

Did you open task manager and watch ethernet performance for NICs and individual CPU core usage?

Do your SSD/NVMe drives on PC/laptop have dram or they are dram-less?

View: https://youtu.be/0rFL8PGkmMU

1.Of course I did check when I was using 2XUSB 3.0 2.5G NIC on my laptop, and it is clear that 2X NIC get 900Mb/s+ data stream on each NIC and also in that case everything is normal, no matter if it is writting or reading, the overall speed is always = 2X1G speed. The problem only happens when I have 2X1G from nas and 1X2.5G to computer. and the problem only happens to writting (copy from computer to NAS)
2. Mate, it is irrelevatnt. No matter if it is SATA or Nvme, Dram or Dram-less, on computer side the SSD will never be the bottleneck that slow the speed down unless I try to copy so many files from NAS that is more than enough to use all its SLC cache, which is something I have never done before.

 

[bill001g]

So I didn't read the details of this thread since it appeared to be mostly NAS stuff which is not my strongest area.

In the last post I noticed the discussion of 2x1g vs 1x 2.5.

The problem you have is that port bonding does not work the way most people think. Making this more confusing is there are proprietary methods that some device,like nas, support. Linux also has support for some proprietary methods.

So the standard method is called 802.3ad. This is the only method that tend to be supported if you are using a switch or router inbetween. I think windows also only supports 802.3ad unless you load other software.

802.3ad load balances by session. What this means is a single file transfer will always use the same ethernet port. So 1 would be used to 100% and the other left idle. Even with multiple sessions the algorithm is simplistic. It does not look at the actual utilization and it can put all the sessions on the same port and leave the other unused.

It was really designed for a central server with many end clients accessing. With a large number of sessions the load tends to balance more just because of randomness.

Now the proprietary ones can do things like balance by packet but you then get a new problem. Say you have a transfer made up of 1 large 1500 byte packet and then 15 100 byte packets. It will send the first large packet on cable 1. It will then send the 15 other packets on cable 2. The first 14 of the smaller packets will arrive before the first large packet does. The receiving device will think the large packet was lost and request a retransmission. In some cases it will think there is too much loss and just restart the session. You can get very band performance using some of these proprietary methods on some data types.

Very advanced forms of port bond chop the data up at the bit level. They make sure the packets are in the correct order before they give them to the OS. This is very uncommon method used on special load balancing appliance boxes.

Pretty much as soon as faster ethernet ports were available corporate users completely abandoned any kind of port bonding. It is now mostly used as a extremely fast backup path if a switch or port in a server were to fail.

 

[FAhentai]

So I didn't read the details of this thread since it appeared to be mostly NAS stuff which is not my strongest area.

In the last post I noticed the discussion of 2x1g vs 1x 2.5.

The problem you have is that port bonding does not work the way most people think. Making this more confusing is there are proprietary methods that some device,like nas, support. Linux also has support for some proprietary methods.

So the standard method is called 802.3ad. This is the only method that tend to be supported if you are using a switch or router inbetween. I think windows also only supports 802.3ad unless you load other software.

802.3ad load balances by session. What this means is a single file transfer will always use the same ethernet port. So 1 would be used to 100% and the other left idle. Even with multiple sessions the algorithm is simplistic. It does not look at the actual utilization and it can put all the sessions on the same port and leave the other unused.

It was really designed for a central server with many end clients accessing. With a large number of sessions the load tends to balance more just because of randomness.

Now the proprietary ones can do things like balance by packet but you then get a new problem. Say you have a transfer made up of 1 large 1500 byte packet and then 15 100 byte packets. It will send the first large packet on cable 1. It will then send the 15 other packets on cable 2. The first 14 of the smaller packets will arrive before the first large packet does. The receiving device will think the large packet was lost and request a retransmission. In some cases it will think there is too much loss and just restart the session. You can get very band performance using some of these proprietary methods on some data types.

Very advanced forms of port bond chop the data up at the bit level. They make sure the packets are in the correct order before they give them to the OS. This is very uncommon method used on special load balancing appliance boxes.

Pretty much as soon as faster ethernet ports were available corporate users completely abandoned any kind of port bonding. It is now mostly used as a extremely fast backup path if a switch or port in a server were to fail.

Sorry mate but you are totally wrong!
SMB3.0 gets nothing to do with 802.3ad,aka link aggregation. I have tried this before since both my router and NAS support that. Link aggregation can only do some thing like auto balance, nevertheless, it can not help point to point, single task, which means that if you try to load or write files, this can not help, eg: suppose you have 2X1G bound on your Synology nas with 802.3ad link aggregation turned on, while on anther hand you have 2.5G to your computer, you loading and writting speed will be still restricted to 1G ethernet. The only difference is, when you have 2 computers connect to the nas and try to load and write to the nas simultaneously, both of your computer will get 1G speed. Yet the only thing I don't know about this stuff is, how about the situation that is link aggregation to link aggregation? Eg, if I have a computer that has 2 ethernet ports work at 1G, same as your NAS, then you do link aggregation for both computer end and Nas end, what will happen?
-----------------------------------------------------------------------------------------------------------------------------
However, SMB3.0 muti-channel is different from that. 1st, you can only use SMB protocol, which means that FTP or other protocol can not get any benfits from that. 2. If you try to use that, you have to TURN OFF THE BOND, which means that you have to stop link aggregation before you can use that. In fact, if you tried it with Synology, you will realized that as long as you enabled SMB muti-channel, when you try to create a bond of your internet, regardless what kind of bond it is, there will be a warning jump out says that "This will stop SMB Muti-channel" 3. Smb 3.0 muti channel can effectively increase the speed of point to point single task,the results I showed above is the evidence.
The problem is, on theory, when you are doing SMB3.0 muti channel, it should work both asymmetrically and symmetrically. Eg, no matter if it is Nas 2X1G to computer 1X2.5G, or Nas 2X1G to computer 2x1G/2.5G or even 10G, the final speed should be 2X1G which works out 220MB/S+. While based on my test, it seems like, to unleash full speed of SMB muti-channel, a symmetrtric enviroment is still needed. Eg, if you have a nas that supports SMB3.0, that has 2X1G, it will be better your computer has 2X1G (in fact speed does not matter, even if your nas go through a switch does not matter, the only thing matters is how many ethernet ports on your computer, Eg, if you have 2X1G, it is fine, if you have 2X2.5G, it is still fine), otherwise, it still works, but not as good as we expected :loading from Nas is still okay, 2X1G speed, while writting things onto NAS, the speed will be only 10%- 20% quicker than 1G.
After I post this thread, I did check out some of the NAS review and test, it seems like in the case of asymmteric connection, both UGreen and QNAP nas perform the same: Reading is okay, always doubled, while writting is still not very much improved.
-----------------------------------------------------------------------------------------------------------------------------------
The primary reason I want to work this out is because I want old Nas to work a little bit longer. In fact, Nas is not a computer that you do gaming on it, it shuold be treated as your a router or, like the computerized cash register that is used in supermarket: You will use it till its life ends, rather than obsolete it because it is out of time.
However, the reality is, for HDD, it will not last much longer in home domestic market, even for normal commercial usage. Why? Cuz the the price of making a HDD is not cheap, while for SSD, the price can gets cheaper and cheaper. Last year I have heard a rumour, says that YTMC is planning to release some of its slow NAND chip that is good enough to make SATA SSD, the price? 18T 1899 RMB, works out 262 usd, tho it did not happen, yet now there is official announcement says they have a tech breakthrogh on QLC Nand.Thus, it will not be too long for us to have cheaper than HDD SATA SSD. When that time come, what are you going to do? Buy a new nas? My answer is no. In fact, the potential is not fully dig out.
Here are some spicific case, Eg Synology DS 918+/DS920+, I will just put a USB 3.0 5G NIC from QNAP so that I can get about 420MB/S speed (yes, it is 5G NIC however the speed is restricted by USB3.0 so that is why it can only go that fast ).However, there is stil a noticeble cost.
Meanwhile,for smb 3.0 muti-channel can offer a 0 cost and even better solution for many Nas. Eg: I don't know if you guys has noticed or not, Snology has many models come with 4X1G Lan DS1517+ or some other Nas from Synology that has 4X1G, SMB muti-channel offers 440M speed in that case.Besides,for QNAP, ASUSTOR, UGREEN, you will found out it they have so many models come with 2x2.5G Lan. In those cases, just get a dirty cheap 2X10G+4x2.5G switch and turn on SMB3.0 muti-channe. No more problem.But, the writting speed performance in those cases is totally out of my expectation. **** [Moderator edit to remove profanity.]

 

[lantis3]

Create a Windows VM on Synology that has two virtual NICs and each one mapped to individual NIC on NAS or the 2 x 2.5G usb ethernet adapters (either virtual or hardware passthrough?) you just got, see if the VM can take full advantage of the SMB 3.0 multiple channel. If it could, then it's probably Synology's driver problem,

OP no need to mock people, if you think everyone here is incompetent, you can go contact Synology.

 

[FAhentai]

Create a Windows VM on Synology that has two virtual NICs and each one mapped to individual NIC on NAS, see if the VM can take full advantage of the SMB 3.0 multiple channel. If it could, then it's probably Synology's driver problem,

OP no need to mock people, if you think everyone here is incompetent, you can go contact Synology.

Come on, I get no intension to mock any one. I get something better to do.
Contact Synology? I have already done it. However, I have not too much expectation for them to work out the problem, even they are able to do so they are more likely to ignore the problem, why? They are evil. This is not something that is just conspiracy, this is the fact. I can give you many examples. If you ever check its DS918+ DS1517+ DS920+ DS923+ DS1522+'s hardware, you will found out that they are deliberately either not fully release all the function of the hardware, or even worse, disable some function that is already avaliable on the hardware so that they can force you to buy either higher grade model or some Synology only upgrade parts. USB 3.0 NIC is another example, in fact case of SMB3.0 Muti-channel can also be an example: more than 3 years ago, there are already many vedio tut teachs people how to enable it based on the old DSM system whitch there is no official support of SMB3.0 muti-channel.
Logically speaking, the reason behind this could be something that is obvious. It is Synology's driver, or I should say, DSM system's problem. Otherwise there is no reason for its working fine with 2X1G (NAS) to 2X2.5G(Laptop). However, the primary reason for me to post such thing here is to ask if any one is possible to found a solution for this since I know in this place there are so many people with insanly high skill with those stuff.

 

[lantis3]

Don't know what your 2.5G ethernet chipset is.

Don't know if it's related. Full/half duplex issue? Either Linux or Windows?

Fixing performance issues with Realtek RTL8156B 2.5GbE USB dongle in Ubuntu - CNX Software

A few days ago, I reviewed a USB 3.0 to 2.5 Gbps Ethernet adapter based on Realtek RTL8156B chip in Ubuntu 20.04, and let's say the reliability and

www.cnx-software.com

 

[FAhentai]

Don't know what your 2.5G ethernet chipset is.

Don't know if it's related. Full/half duplex issue? Either Linux or Windows?

Fixing performance issues with Realtek RTL8156B 2.5GbE USB dongle in Ubuntu - CNX Software

A few days ago, I reviewed a USB 3.0 to 2.5 Gbps Ethernet adapter based on Realtek RTL8156B chip in Ubuntu 20.04, and let's say the reliability and

www.cnx-software.com

It is clear that the problem is not from the NIC. Otherwise there is no way that I when 2 of them are there the speed is fine while just one of them online the reading speed is fine but the writting speed is slow.

Btw, believe or not, I get the answer from Synology, guess what they say? They simply repeat what I told them(the facts), and said we do understand this is going to affect your expirence. Then, showed me the the FAQ, which does not answer my question but clearly states that as long as there is a proformance improvment, smb3.0 muti-channel is succussfully deployed. Just like what I expected. There is no expecation from them

 

[Ralston18]

@FAhentai

I have been loosely following this thread and have some sense of the overall discussion.

What I am not so sure about is the proverbial "big picture".

Is is possible for you to provide some sort of diagram showing devices, connections, configurations, read/write speeds etc.?

Actually two diagrams:

1) a diagram showing what you expect or require to happen with respect to file transfer speeds and so forth.

2) a diagram showing what is actually happening with respect to file transfer speeds and so forth.

And no harm if another one or two diagrams prove necessary.

I do not know what is happening nor do I have an answer (full disclosure here).

However, it just seems to me that laying it all out in comparative diagrams may prove helpful.

Just my thoughts on the matter.

 

[FAhentai]

@FAhentai

I have been loosely following this thread and have some sense of the overall discussion.

What I am not so sure about is the proverbial "big picture".

Is is possible for you to provide some sort of diagram showing devices, connections, configurations, read/write speeds etc.?

Actually two diagrams:

1) a diagram showing what you expect or require to happen with respect to file transfer speeds and so forth.

2) a diagram showing what is actually happening with respect to file transfer speeds and so forth.

And no harm if another one or two diagrams prove necessary.

I do not know what is happening nor do I have an answer (full disclosure here).

However, it just seems to me that laying it all out in comparative diagrams may prove helpful.

Just my thoughts on the matter.

https://ibb.co/HdWR19v Pic https://ibb.co/HdWR19v
I don't know why forum can not read this platform, thus I just put the link there.
In fact the problem is already been identified. Yet there is no solution.Any way, back to the diagram there, in the case of CASE A, the reading speed (From nas to PC) is right, 220MB/S, stable, however, the write speed, 120MB to 140MB. However, in the CASE B, both write and read speed are right, max at 220MB/S.
And I am very sure that it is not my devices that restricted the speed, cuz back to the old days I was using a USB3.0 to 2.5G NIC on the Synology, both reading and writting speed is more than that with the same test.

 

[Ralston18]

Thank you.

Imgur (www.imgur.com) seems to work for most posters.

In any case, I can see and open the link with no problems.

And between Case A and Case B all is the same except Computer and Laptop.

Computer apparently not having 2x2.5G

So my next thought/question: is it expected or viable for the computer to have 2x2.5G?

Computer specs vs laptop specs?

What are the differences? Any differences?

 

[FAhentai]

Thank you.

Imgur (www.imgur.com) seems to work for most posters.

In any case, I can see and open the link with no problems.

And between Case A and Case B all is the same except Computer and Laptop.

Computer apparently not having 2x2.5G

So my next thought/question: is it expected or viable for the computer to have 2x2.5G?

Computer specs vs laptop specs?

What are the differences? Any differences?

No. This test only prove one thing, as I said before, on theory, when you are doing SMB3.0 muti channel, it should work both asymmetrically and symmetrically. Eg, no matter if it is Nas 2X1G to computer 1X2.5G, or Nas 2X1G to computer 2x1G/2.5G or even 10G, as long as the the one side the final speed should be 2X1G which works out 220MB/S+. Thus, case A and Case B there should work out the same writting and reading speed.
However, based on the test result shows there is a performance difference.Case A's writting speed is lower than Case B. Generally, this kind of performance difference applies to many other brands as well, Such as Qnap and Ugreen. However, case A is far more practical comparing with case B, and also, it has a lot of potentials. I write the details on #7.

 

[Ralston18]

"when you are doing SMB3.0 muti channel, it should work both asymmetrically and symmetrically. Eg, no matter if it is Nas 2X1G to computer 1X2.5G, or Nas 2X1G to computer 2x1G/2.5G or even 10G, as long as the the one side the final speed should be 2X1G which works out 220MB/S+."

If the final speeds are not as it should be or as expected then there is something different.

Are you familar with Powershell?

E.g.:

https://learn.microsoft.com/en-us/p...e/get-smbconnection?view=windowsserver2022-ps

Powershell has quite a collection of Smb related cmdlets.

Try running (as admin)

Get-SmbServerConfiguration

Get-SmbClientConfiguration ( on laptop and computer).

Compare cmdlet results: look for what should be identical and look for what should be different.

Here are the results from my computer:

Windows PowerShell
Copyright (C) Microsoft Corporation. All rights reserved.

Install the latest PowerShell for new features and improvements! https://aka.ms/PSWindows

PS C:\WINDOWS\system32> Get-SmbClientConfiguration


CompressibilitySamplingSize : 524288000
CompressibleThreshold : 104857600
ConnectionCountPerRssNetworkInterface : 4
DirectoryCacheEntriesMax : 16
DirectoryCacheEntrySizeMax : 65536
DirectoryCacheLifetime : 10
DisableCompression : False
DormantFileLimit : 1023
EnableBandwidthThrottling : True
EnableByteRangeLockingOnReadOnlyFiles : True
EnableCompressibilitySampling : False
EnableInsecureGuestLogons : True
EnableLargeMtu : True
EnableLoadBalanceScaleOut : True
EnableMultiChannel : True
EnableSecuritySignature : True
EncryptionCiphers : AES_128_GCM, AES_128_CCM, AES_256_GCM, AES_256_CCM
ExtendedSessionTimeout : 1000
FileInfoCacheEntriesMax : 64
FileInfoCacheLifetime : 10
FileNotFoundCacheEntriesMax : 128
FileNotFoundCacheLifetime : 5
ForceSMBEncryptionOverQuic : False
KeepConn : 600
MaxCmds : 50
MaximumConnectionCountPerServer : 32
OplocksDisabled : False
RequestCompression : False
RequireSecuritySignature : False
SessionTimeout : 60
SkipCertificateCheck : False
UseOpportunisticLocking : True
WindowSizeThreshold : 8

= = = =

Your devices will likely be configured somewhat differently - my computer not being in your environment.

The objective simply being to demonstrate that there could be many reasons that the speeds are not as expected.

And you may need to use other Smb cmdlets to discover some difference or differences. I cannot say which particular cmdlets (if any) would discover the difference. Get-SmbShare would/could be one to run and compare results.

Lastly, and it took me some time to remember it - look at the SMB versions.

https://learn.microsoft.com/en-us/w...ct-enable-and-disable-smbv1-v2-v3?tabs=server

I do remember seeing some past posts regarding problems and issues regarding smb versions.

No harm in confirming that the versioning is correct.

You not have to change anything to look and compare. Get cmdlets are passive and just provide information.

 

[FAhentai]

"when you are doing SMB3.0 muti channel, it should work both asymmetrically and symmetrically. Eg, no matter if it is Nas 2X1G to computer 1X2.5G, or Nas 2X1G to computer 2x1G/2.5G or even 10G, as long as the the one side the final speed should be 2X1G which works out 220MB/S+."

If the final speeds are not as it should be or as expected then there is something different.

Are you familar with Powershell?

E.g.:

https://learn.microsoft.com/en-us/p...e/get-smbconnection?view=windowsserver2022-ps

Powershell has quite a collection of Smb related cmdlets.

Try running (as admin)

Get-SmbServerConfiguration

Get-SmbClientConfiguration ( on laptop and computer).

Compare cmdlet results: look for what should be identical and look for what should be different.

Here are the results from my computer:

Windows PowerShell
Copyright (C) Microsoft Corporation. All rights reserved.

Install the latest PowerShell for new features and improvements! https://aka.ms/PSWindows

PS C:\WINDOWS\system32> Get-SmbClientConfiguration


CompressibilitySamplingSize : 524288000
CompressibleThreshold : 104857600
ConnectionCountPerRssNetworkInterface : 4
DirectoryCacheEntriesMax : 16
DirectoryCacheEntrySizeMax : 65536
DirectoryCacheLifetime : 10
DisableCompression : False
DormantFileLimit : 1023
EnableBandwidthThrottling : True
EnableByteRangeLockingOnReadOnlyFiles : True
EnableCompressibilitySampling : False
EnableInsecureGuestLogons : True
EnableLargeMtu : True
EnableLoadBalanceScaleOut : True
EnableMultiChannel : True
EnableSecuritySignature : True
EncryptionCiphers : AES_128_GCM, AES_128_CCM, AES_256_GCM, AES_256_CCM
ExtendedSessionTimeout : 1000
FileInfoCacheEntriesMax : 64
FileInfoCacheLifetime : 10
FileNotFoundCacheEntriesMax : 128
FileNotFoundCacheLifetime : 5
ForceSMBEncryptionOverQuic : False
KeepConn : 600
MaxCmds : 50
MaximumConnectionCountPerServer : 32
OplocksDisabled : False
RequestCompression : False
RequireSecuritySignature : False
SessionTimeout : 60
SkipCertificateCheck : False
UseOpportunisticLocking : True
WindowSizeThreshold : 8

= = = =

Your devices will likely be configured somewhat differently - my computer not being in your environment.

The objective simply being to demonstrate that there could be many reasons that the speeds are not as expected.

And you may need to use other Smb cmdlets to discover some difference or differences. I cannot say which particular cmdlets (if any) would discover the difference. Get-SmbShare would/could be one to run and compare results.

Lastly, and it took me some time to remember it - look at the SMB versions.

https://learn.microsoft.com/en-us/w...ct-enable-and-disable-smbv1-v2-v3?tabs=server

I do remember seeing some past posts regarding problems and issues regarding smb versions.

No harm in confirming that the versioning is correct.

You not have to change anything to look and compare. Get cmdlets are passive and just provide information.

Thank you, but, you know what? With case A, it seems like, the problem has been worked out.....And you can not even believe how I made it.
I turned on my wifi on my computer. My router works on WIFI6 160MHZ mode, and my connection speed is 1922M/1922M, I did test this before, with 2.5G NIC on Synology, and wifi only, both my write and read speed are at 150MB/s or something, needless to say it is more than 1G. Thus, at the end it makes my case A become case B. Here is a screen shot when I was doing the test

https://imgur.com/a/NbmM6so

View: https://imgur.com/a/NbmM6so

it is clear that both my ethernet and wifi has some load.
Thus, it seems like, well, no more problem.
And let's think about even bigger picture. As I mentioned above on #7, there are many products, from QNAP, Ugreen, Etc, has 2X2.5G NIC. We also know, that 1. for WIFI7, if 240MHZ mode and 320MHZ mode is on, it is very likely that your wifi speed will be faster than 2.5G ethernet. 2. Sata 3 SSD's speed is less than 5G ethernet 3. Sata SSD is very likely to be cheaper than HDD in next 2-3 years. Hummmm, intersting.