Rashesh, or JT as he’s known around the office, talked on Wednesday about how your applications can find a “better home” in your data center—where they run faster, more efficiently, for less money. To break that down a bit more, let’s take a look at the newest entry into the Cisco server portfolio, the new UCS B230 blade server, or as Sean McGee has dubbed it, “The Goldilocks Blade”. This new entry in the Cisco server portfolio gives you more ways to optimize your infrastructure for specific application needs, always with an eye to your perennial operational and environmental constraints.
My teammate Scott Ciccone, who supports the server elements of the Unified Computing System, invited me up to join the fun at the Intel Developer Forum (IDF) in San Francisco earlier this week. We had the B230 out with its cover off, and the sight of 32 DIMMs packed onto a two-socket, half-width blade drew plenty of curious passersby. Scott’s quick overviews of the blade consistently turned into thoughtful 20 to 30 minute conversations, so I asked him to share a blog-size version of those discussions with you:
You can also find a couple of customer takes on the B230 here and here. But as JT observed, however exciting server or sports car specs can be, our IDF conversations tended to quickly become architectural discussions. As people began to absorb, for example, what all those DIMMs combined with our Virtual Interface Card could mean for VM density, and how you can use UCS Manager to rapidly configure the optimal type of server with the right settings for a given workload, the implications for operational efficiency and scalability--and better app performance--started to become clear.