Every organisation is in the race of expanding the network infrastructure with high availability of applications because of the rising demand of resources. Rising demand of high availability gives the birth to geographically dispersed data centers.
Maintain high availability, resiliency and scalability is one of the challenge during designing geographically dispersed data centers. To accomplish this network designers always look for layer 1, layer 2, layer 3 connectivity between the data centers. Layer 1 connectivity provides a dedicated fiber from one data center to another data center but the solution is only possible if both the locations are in same metro otherwise cost is one of the major constraint in case of geographically dispersed locations. Layer 2 connectivity is also one of the highly deployed solutions in case when the customers use storage connectivity over FCIP (Fiber Channel Over IP). Layer 2 connectivity can be accomplished by EoMPLS or L2VPN. The last one is Layer 3 solution which could purely be based on IP or MPLS cloud.
What does customer demand?
A customer always looks for economical, preeminent and optimizes solution. This is the responsibility of network consultant, designer or architect to provide the best solution according to the requirement. But in new era of technology customers are highly qualified and know what the solution available in market is. They always look for the layer 2 solution aka l2 which means to extend the lan (logical area network) to the remote data centre. This solution works fine in case of single location. But as the location grows; demand will change to extend the LAN to other locations consequence Virtual Private LAN Services (VPLS) need to be provisioned which has its own limitations.
Current Problems with Existing Technologies
1. Data Centres need to share the problems of originating in another data centers consequence downtime increases. This is because of same broadcast domain.
2. Not able to utilize the links properly w.r.t. to the devices. (VSS – Virtual Switching System has solved the problem)
3. VLAN and Mac-Address scalability.
4. Multicast
5. Convergence time.
To overcome all the existing problems Cisco has finally come up with new solution i.e. Overlay Transport Virtualization (OTV).
What is Overlay Transport Virtualization (OTV)?
OTV is “Mac in IP Technique” which uses encapsulation not the legacy tunnelling. Customer can forward the traffic on any transport medium which carries the ip packets. The mechanism is same like of proxy arp where in IP next hop will become the proxy for all the destination mac addresses. MAC address mappings are advertised by the control protocol which restricts the unknown flooding in the remote locations and logically bind all the locations in full mesh to utilize the network and resources its best without hampering another.
This feature is only available in Nexus 7000
SDN and NFV is the next phase of technology change which will help service provider to launch the services in single click. This is all about the programmability of the networks by using open source software defined network controller.
Monday, February 15, 2010
Overlay Transport Virtualization (OTV)
Labels:
Data Centre
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment