Data Center implementation (I)
Hi there!
As we’re coming to an end of the weekly posts, we’d like to sum up a little bit our design, talking about the processes followed from the first design made to the final solution proposed. We haven’t done a proper Cisco comparison until now, so, as we now have our design, it will be easier to look for a similar solution with Cisco.
The design of a proper data center is not easy and you have to consider many things so the explanation of the implementation will be divided into two posts. In this one we’ll talk about networking (LAN and SAN) and the servers used. The other post will be focused on security, management and QoS.
Networking
If you have been reading us, you might remember the first 3 hierarchal tier design we proposed. Well, we were thinking and looking for other Brocade solutions and we found this one:
The previous topology is a leaf-spine design, where all switches are connected, with a Layer 3 link with all superior tiers. Brocade also proposes a border leaf structure where all security devices such firewalls must be connected. On the leaf switches we’ll connect servers used as for storage as to provide services. The layer 1 protocol used on the SAN network will be Fiber Channel over Ethernet (FCoE) and Ethernet for LAN.
We’ve chosen this topology because in case the bank grows in a future, another super-spine structure can be added easily. Besides, we consider that it’s better to start with a small topology offering basic services and also using the cloud rather than spending a huge amount of the budge on the whole infrastructure.
Servers
For the hole LAN network will be using Brocade VDX family devices, following the recommendations of Brocade. So the equipment will be:
- Brocade VDX 6740 for Leaf tier
- Brocade VDX 8770 for the Spine tier
- Brocade VDX 6940-36Q for the Border leaf
- Brocade MLX-4 for the outgoing router
For the servers’ equipment we’ve been asked to use Oracle’s devices. Analysing the advantages of wheather to use blades or servers on racks, we’ve decided to use rack devices instead of blades because we think that for small companies this is a more flexible option. So which devices we’ll use? Oracle Server X6-2.
Because VMware is a very used technology for different manufacturers and because it’s most known than other options, we’ll use VMware to virtualize our servers. Thanks to virtualization we’ll be able to have multiple applications running on the same physical device. And thanks to their multiple tools they have, we could also use some to manage and transfer data in case a machine goes down.
After reading all our previous post, you can’t miss the final proposed solution, so keep reading our next post: Data Center implementation (II).
Add new comment