HP V1910-24G Switch JE006A control speed

[ankitj1611]

Hello

i am using HP V1910-24G Switch JE006A in my office.i want a setup like this

My internet speed is 5Mbps.Suppose there are two ports 21 and 23 and.i want that port 21 receives 2Mbps dedicated.Means i want to control bandwidth.is this possible??

I am in admin panel of my switch and found Device>Flow control Device>port Management


But not able to do so

Please help me.Give me a tutorial.

Thanks

 

[bill001g]

Hard to say exactly how to do this. Would be nice if HP stuck to a single OS for their switches but they have many. I suspect it can be done but I only really know the procurve based switches well.

There are huge issue with doing this since the bottleneck is not in the switch it is in the router or even in the ISP.

So to do the easy one first. You apply a packet marker to the port 21 and mark all its packets say to af31. Now on the port going to the router you can apply a CAR policy that guarantees a minimum bandwidth to all traffic with af32 markings. So between the router and the switch you can guaranty the 2m. But is is likely a complete waste of time because even on a 100m port the traffic there still is 98m more that other ports can use which is more then the 5m total. You of course could build a artificial limit to limit the port to 3m for all the other traffic. In either case you would likely have to put a policy on the router that does similar but since the packets are marked you likely could put in a similar policy.

Now this only limits the outbound traffic going to the internet. It is practically impossible to control the inbound...ie download traffic. When you hit the 5m limit it is the ISP that controls which traffic is sent and which is discarded. You have no way to control if it send 4m for one users or 3m.

What you can do that partially works is to further limit the traffic and hope the application will detect it and slow down. So say the ISP has decided to send 4m of traffic to machine1 and 1m to machine2 but you want to only allow 3m to machine1 so machine2 always can get its 2m.

What you do is limit all traffic not going to machine 2 to 3m. So when the ISP sends 4m to machine1 and 1m to machine2 you throw away 1m or the 4m leaving only 3m. This of course at this moment does not make machine 2 get 2m he still only has the 1m the ISP sent. What you are hoping is that the application will detect this extra loss and slow its requests down for data. Eventually in theory it should not request more than 3m because it does not want to detect errors. Because it is requesting less it leave the 2m for machine2. The huge flaw in this design is that it is all based on the error recovery mechanism in the tcpip stack in the OS. It work surprising well for most applications under microsoft so its worth a try but there are some poorly behaved applications that just ignore errors and continue. Also when you are talking small bandwidths it can be very inconsistent because it takes time to adjust and the tcpip stack constantly will try to for more bandwidth and then slow down again.

In any case the way you configure this is to use a ACL to match the IP address of all the machine that are NOT to have the guarantee. You would then use a rate limiter inbound on the port going to the internet router to limit the traffic.

If you have no background with QoS or things like ACL you are going to have to study for a while. These switches are much more complex than a consumer grade switch.

This can be done on some internet routers also but is is almost as complex.