Optimize Your Data Center: Ditch the Legacy Technologies

In our journey toward two-switch data center we covered:

It’s time for the next step: get rid of legacy technologies like six 1GE interfaces per server or two FC interface cards in every server.

Need more details? Watch the Designing Private Cloud Infrastructure webinar. How about an interactive discussion? Register for the Building Next-Generation Data Center course.

5 comments:

  1. I agree having 6 x 1GE links per physical server harkens back to 2004. I remember visiting a state of the art DC run by Telus Canada in 2004. We needed 14 servers and their sales and solution's team were showing us around and I remember them pitching how rock solid their servers were because they had 8 physical NICs to keep the traffic types physically separated. Now DCs such as Rackspace offer 10G bonded links, and soon 25G interfaces will start showing up. So we are in a time where we might have to throw out those old recipes and start experimenting to come up with new ones!
    Replies
    1. You guessed what the next video in the series will be ;)
  2. One other thing, its not sufficient to just focus on hardware. I think engineering and operations teams need to think about their applications and how they are deployed. Case in point, how did we provide high availability in the past? sometimes we would use something like Veritas cluster manager and other times we would build something similar through various building blocks like Windows Server Failover Clustering at the host level, then layering SQL Server Failover Clustering, and now layering on Microsoft's Always On Availability groups which required shared storage through a SAN. But now with VMware DRS and HA services and the fact that Always On Availability groups no longer requires shared storage we are challenged with how to utilize the best parts of these products to build a highly available service that's easy to maintain and provides the best value to the business. Interesting times!
    Replies
    1. Thanks for mentioning the Always-On Availability Groups. While I did study how database replication works in SQL Server, it must have been before SQL Server 2012, and the implementation was a bit kludgy. Now (with the Availability Group Listener) it makes even more sense.
  3. Optimize your data center:

    Ditch the proprietary technologies:

    i.e. software depencies:

    VMWare VSAN
    VMware NSX
    3Par SAN
    :

    Let´s fill that list...


Add comment
Sidebar