Comments (2)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 H2NS_Rüdiger_Zöllner commented Permalink

Hello Sebastian, <br /> I totally agree with your comments regarding server virtualisation and storage backend. For these backend systems the change from physical to virtual servers is not really a problem. <br /> However, there is really much more complexity coming in with server virtualisation but it is related typically to SAN components (including network, FCoE a.s.o) as the risk of bottlenecks is increasing. <br /> Yes, this can limit I/O performance and availibility but always related to SAN components instead of the Storage system. <br /> Yes Hu is also right that storage virtualisation can limit such risks but the virtualisation needs to be in the SAN and not in the backend system to be effective. <br /> Virtualisation as SVC is designed opens the opportunity for customers to select storage systems fitting to his Tier Levels and on the other hand side optimise the I/O in the server direction. <br /> In addition the failover mechanisms of the hosts getting consolidated. Only one driver (or even module) is needed within the hosts. <div>&nbsp;</div> I think Hu wrote his blogs with much sales passion, the facts are looking different from postsales perspective. <br /> Thanks for your efforts to clarify.

2 seb_ commented Permalink

Hi Rüdiger, <br /> thanks for your comment! Of course I also tend to see the appliance approach the SVC uses as very useful, but I might be biased, because coming from IBM support I just know more SVC installations and see the advantages customer's gain from the SVC. <br /> About the "sales passion" part... hmm... I try to see it that way, but to be honest also many of his older entries about the advantages of the HDS approach often look like "reality-bending by intention". <br /> :o)