Swift
asked on
Segregating Voice and Data traffic at point of origin
Hi
At one of our office locations, we run a mix of voice, data and VOIP. We have taken the stance that VOIP is data and therefore should use the existing ip addressing. The way we separate voice and data is by VLAN. The data VLAN is an even number and the voice is an odd number.
1st floor VLAN88 10.10.88.x Data LAN
1st floor VLAN89 10.10.89.x Voice LAN
This way we can tie the data VLAN to the subnet and associate VLAN and subnet for the voice. We have done this for each floor in BV and intend to extend it to the remote offices also.
My concern is, if at some later stage, we decide to bring in MPLS inter-site connectivity and decide to implement QoS or 'Class of Service', giving priority to Voice over Data, would this IP schema be sufficently capable of handling that part or we might be forced to re-think this for something more efficient?
How do you guys usually demaracte data vis a vis VoIP traffic on internal IP subnets? Is it better to use a separate class of Private IP subnets altogether, e.q 192.168.*.* or 172 range of rfp 1918?
Pls advise!!
At one of our office locations, we run a mix of voice, data and VOIP. We have taken the stance that VOIP is data and therefore should use the existing ip addressing. The way we separate voice and data is by VLAN. The data VLAN is an even number and the voice is an odd number.
1st floor VLAN88 10.10.88.x Data LAN
1st floor VLAN89 10.10.89.x Voice LAN
This way we can tie the data VLAN to the subnet and associate VLAN and subnet for the voice. We have done this for each floor in BV and intend to extend it to the remote offices also.
My concern is, if at some later stage, we decide to bring in MPLS inter-site connectivity and decide to implement QoS or 'Class of Service', giving priority to Voice over Data, would this IP schema be sufficently capable of handling that part or we might be forced to re-think this for something more efficient?
How do you guys usually demaracte data vis a vis VoIP traffic on internal IP subnets? Is it better to use a separate class of Private IP subnets altogether, e.q 192.168.*.* or 172 range of rfp 1918?
Pls advise!!
grblades
What you are planning to do will be fine. If you use the 2nd octet to denote the site (10.10 is site 1, 10.11 is site 2 etc..) then you could use superneting to reduce the number of networks across the MPLS link. However routing protocols such as EIGRP or OSPF can support a huge number of routes so this wont be an issue even if you dont use superneting.
ASKER
Blades, thanks for the reply. My concern was largely related to tagged Voice traffic when it leaves my network, either using Diffserv on the router before the traffic exits my LAN or implementing end to end QoS for the Voice part internally over my switches.
I was wondering if it gets easier if I use some other IP private subnet range alltogether apart from the 10.*.*.* and it would make th identification easier in case of future implementation of MPLS VPN between by sites.
So, you feel, I shouldn't refine this further?
I was wondering if it gets easier if I use some other IP private subnet range alltogether apart from the 10.*.*.* and it would make th identification easier in case of future implementation of MPLS VPN between by sites.
So, you feel, I shouldn't refine this further?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.