Unsolved
This post is more than 5 years old
36 Posts
1
2003
Vplex metro - options for connecting vplex clusters ...
Hello Guys,
For a VPLEX Metro, it must be 5ms RTT for latency, this will be synchronous data between sites .
Also, I know have some partial info. that there is requirement for CISCO OTV for inter DC WAN connection.
Guys, Can you please help to understand, for what actually.. OTV is required for actually what ?
And what are the other alternatives OTV ..
Please help..
TadhgConcannon
1 Rookie
1 Rookie
•
22 Posts
3
April 13th, 2015 08:00
HI Anil
Hope all is well.
Cisco OTV stands for "Overlay Transport Virtualization" which is technology to interconnect two data centers over layer 2. The purpose in very high level terms is to join seperate Data Centers and make them appear as one.
OTV enables Layer 2 extension between data centers over a Layer 2 or Layer 3 data center interconnect (DCI). The aggregation tier contains the functional elements to provide DCI.
OTV is not a requirement for VPLEX.
However if you are going to to deliver the all the benefits of a VMWARE HA environment with VPLEX then you will need OTV or similar to make the Data Centers appear as one.
In traditional DR environments Each DC would have an IP range associated with it. However when an applications fails from DC 1 to DC2 to the IP address will more than likely need to change. Making the recovery process a little more complicated.
With VPLEX the concept of Active Passive Data Centres is now longer a concern as either DC can be active. VPLEX delivers a true active active environment with all the benefits i.e. load balancing, Continuous Availability etc.
Take a typical environment 2 x DC. To have a Virtual Machine live migrate from DC 1 to DC 2 with VPLEX then a stretched VLAN is required
From a VMWARE HA perspective you would need a stretched VLAN that covers both DC so that if a Virtual Machine is moved from DC 1 to DC 2 its IP address will remain the same. Without a stretched VLAN this is more complicated to achieve.
Brocade have a similar technology called VXLAN which provides similar functionality as Cisco OTV
Some useful reading can be found here
Extend Layer 2 applications across distributed DC
http://www.cisco.com/c/en/us/solutions/data-center-virtualization/overlay-transport-virtualization-otv/index.html
Cisco Overlay Transport Virtualization FAQ
http://www.cisco.com/c/en/us/products/collateral/switches/nexus-7000-series-switches/qa_c67-574969.html
Using OTV to Extend Layer 2 between Two Data Centers
http://www.cisco.com/c/en/us/products/collateral/switches/nexus-7000-series-switches/white_paper_c11-644634.html
OTV Technology Introduction and Deployment Considerations
http://www.cisco.com/c/en/us/td/docs/solutions/Enterprise/Data_Center/DCI/whitepaper/DCI3_OTV_Intro/DCI_1.html
Hopefully this will get you pointed in the right direction. Any further questions reply to this thread
Kind Regards,
Tadhg
Anil2sharma
36 Posts
1
April 13th, 2015 23:00
Thanks for the nice reply.
One more query...
Lets forget about the Host side and the stretch LAN for VMWARE HA..
For VPLEX setup in two remote DC. What technologies we can use for VPLEX to VPLEX cluster SAN data traffic.
I guess it would be same like we have requirements for SRDF implementation interns of FC connectivity over WAN or other dark fiber options like DWDN or SONNET.
Please share your thoughts....
TadhgConcannon
1 Rookie
1 Rookie
•
22 Posts
1
April 14th, 2015 03:00
Hi Anil
There are two choices of connecting VPLEX Metro
Clusters in a VPLEX Metro can be connected by either Fibre Channel (8 Gbps) or Gigabit Ethernet (10 GbE).
When you order your VPLEX solution you have a choice of Fibre or Gigabit Ethernet connectivity and the EMC SA will advise based on your current DC connectivity
But in summary its the same requirements essentially as for SRDF if you have that in place currently
For more detailed information please see
h13552-vplex-ip-networking-best-practices from support.emc.com
vplex-san-connectivity-best-practices from support.emc.com
Any additional questions please reply to this thread
Kind Regards,
Tadhg