XGS1250-12 transer rate a whole 10MB (megabytes) less than my old switch on SMB transfers
I've set the QOS to queue 7 for all ports just to be sure.
I have disabled IEEE power saving.
Also have disabled the Unknown Multicast block (tested with it on and it made no difference).
16GB file from server to laptop:
(The path is using a specific VLAN from Server all the way through to the Access Point, where that VLAN underpins a WiFi profile.)
When I transfer via SMB (Windows) from fileshare on server to laptop via wifi using Zyxel XGS1250-12 2.5G port connected to NWA210AX AP, I get a maximum of 57MB/s.
But when I swap-out the xgs1250-12 for ICX7450-32ZP, I get 67MB/s.
So that is a whole 10MB/s less via the xgs1250-12.
Is this a known limitation of this switch?
I just bought it from Amazon UK, and wondering if I should return it because this is quite a big drop in throughput.
However, before I do that, I thought I'd check here to see if this can be resolved?
P.S. I know it is not fair to compare a $$$$ Layer 3 Enterprise switch to a $$ consumer device, so please note this is NOT a comparison whatsoever.
However the test is just a straightforward throughput test across a VLAN, no routing or any other features are being employed. I would expect even a consumer device to at least achieve the same throughput on a switch that is doing absolutely nothing else, just doing one job, a straightforward transfer.0
Zyxel_Nami Posts: 96
Thank you for your comment and for sharing the results of your throughput test. We understand your concern about the performance of the Zyxel switch compared to the other switch.
As you mentioned, the traffic going from Server to Laptop via Switch, it depends on various factors that may affect to the throughput. For the throughput difference between these two Switches, the hardware specification should be considered.
You can check the datasheet of XGS1250-12, its Packet Buffer size is 1.5MB meanwhile ICX7450-32ZP's is 4MB that is on different level.
Here is XGS1250-12 Specification Pagge:
If you have any further questions, please don't hesitate to let us know.
Ok, yes I did think it is merely down to specs as the test environment was not contaminated by any other variables, and the results are consistent every run. I was hoping maybe there was a hidden feature on the switch or a different firmware that might cure the drop in throughput.
Ok, I will try my old XS1930-12HP to see what I get, will report back.
Effectively I am trying to reduce the power-consumption of the network, already have a 100G switch as ToR which consumes ~50watts with 7 of 8 100G ports populated by optical QSFPs. That is the constant, perfectly fine with that, the variables are the leaf switches downstream which provide wifi/iot etc functionality.
The ICX7450-32ZP consumes a lot, hence looking at replacing it.
With rising costs of energy, and no incentive for them to ever return to normal, energy efficiency has become the highest priority in most hardware selection these days, even datacentres are starting to dump stacks of switches purely for cost mitigation.0
Ok, done the test on the XS1930-12HP, we're getting 65MB/s on average, however it does peak around 68MB/s, whereas the ICX peaks around the 70MB/s.
So this means the XGS1250-12 will have to be returned, nearly a 20% bandwidth loss is too much a hit.
Interestingly, total power consumption of the XGS1250 + injector, matches the XS1930-12HP. I did upgrade to the latest firmware and was using a DAC cable so disabled the SFP+ fan as that would only be useful I presume if I were to plug in optical-SFPs.
I guess the XS1930-12HP is indeed a very good efficient switch, just a shame it doesn't have a CLI and console port. Not sure why it was designed that way, quite offputting for those of us who rely upon remote management over console KVM.0
I just confirmed that the ICX7450-32ZP has the same size packet buffer as the 48 ports, ie. 8MB
That actually explains a lot, I've been juggling a few different switches (Cisco/Aruba/Extreme) in some high-load test scenarios over the past few years and the old 32ZP handled everything thrown at it where others often choked. If only they made a new power-efficient version of it rather than just EOL'd.
Well I have decided to purchase another XS1930-12HP now, the one I used for testing was ‘borrowed’ from another network I manage. Looks like there will be two under my management now.0
Zyxel_Nami Posts: 96
Thank you for taking the time to test and share your experience with us regarding the XGS1250-12 and XS1930-12HP switches on the Zyxel Community. We really appreciate your efforts in helping us improve our products and services.
In order to access CLI of XS1930-12HP Switch, you could consider to have the Access L3 License. This license will enable you to access the CLI of the switch as well as to perform configuration. However, please note that the device should be updated to the latest firmware (4.80) to utilize it.
Here is the XS1930 Series Specification page for your reference.
Please contact our local distributor if you want to find out more information about this license. Check it out our distributors in UK: https://www.zyxel.com/uk/en-gb/where-to-buy/distributors
We hope that this information is helpful to you, and please don't hesitate to reach out if you have any further questions or concerns. We value your feedback and are committed to providing the best support possible to our customers.
Thank you for being a part of the Zyxel Community!
- 8.5K All Categories
- 1.6K Nebula
- 71 Nebula Ideas
- 57 Nebula Status and Incidents
- 4.5K Security
- 226 Security Ideas
- 982 Switch
- 46 Switch Ideas
- 876 WirelessLAN
- 22 WLAN Ideas
- 5.1K Consumer Product
- 157 Service & License
- 280 News and Release
- 98 Success Stories
- 59 Security Advisories
- 13 Education Center
- 580 FAQ
- 263 Nebula FAQ
- 160 Security FAQ
- 76 Switch FAQ
- 74 WirelessLAN FAQ
- 7 Consumer Product FAQ
- 34 Documents
- 34 Nebula Monthly Express
- 69 About Community
- 46 Security Highlight