Troubleshooting cloud applications is still a networking manager's job, says PacketPushers blogger and network administrator Ethan Banks. Banks says managers have to be able to pinpoint network performance problems, even if those problems aren't within traditional network boundaries. Banks says that if there is an issue with applications that are in the cloud and not on the physical network, it is still the job of the network manager to diagnose what is going on. And, he writes, it's more important to figure out what the problem is rather than what the problem is not. But this is easier said than done. Banks cites tools such as Thousand Eyes as the kind of technology that allows networking professionals to gain visibility and gather data in a meaningful way across physical and cloud networks. Instead of being forced to sift through tons of data, Banks relies on data visualizations created by Thousand Eyes to pinpoint trouble areas.
Read more about how Banks is using Thousand Eyes for network troubleshooting.
RSA Conference begins with a reminder of the current state of security
Enterprise Strategy Group senior analyst Jon Oltsik says he experienced a bit of a somber moment at the beginning of this year's RSA Conference. He attended a dinner hosted by Pacific Crest Securities and was struck by the opening remarks of the host, cybersecurity consultant Richard Clarke. Oltsik paraphrases Clarke's words: "I don't believe that anyone would have dreamed that the industry would be as big as it is today. So we've all had a good ride and made a little bit of money along the way … But here's the problem: Few if any of us would have guessed that the cybersecurity risks to our nation, our critical infrastructure, and our sensitive data would be worse today than it was 10 years ago." Clarke encouraged attendees to stay focused on the task at hand and to take steps to improve security.
Read more of Oltsik's initial takeaways from day 1 at the RSA Conference 2015.
Choosing between merchant and custom silicon
Current Analysis analyst Mike Fratto says that when evaluating custom versus merchant silicon, it isn't the either/or decision that some people think it is. Choosing a vendor has become more complex, fueled by the rise in popularity of white box switching and associated switching products as well as the increased focus on software. Fratto says that merchant silicon is for the IT buyer who needs standard fast forwarding and routing of typical protocols. Features such as Layer 2/3 protocol support, stacking, port density and performance will be pretty much the same for all merchant silicon. On the other hand, if an IT buyer is looking for custom silicon, Fratto says it's important to understand the advantages of custom silicon and have those performance claims backed up by data.
Read more on how Fratto advises IT buyers to choose between merchant and custom silicon.
Virtualized Geek hopes for increased SDN integration
Virtualized Geek blogger Keith Townsend says that when it comes to the recently announced Maxta/Cumulus partnership, he hoped for a deeper level of software defined networking integration into the hyperconverged infrastructure. His ideal scenario is what he calls an "infrastructure brick," where a data center is designed on scale-out nodes that provide storage, compute and network on commodity hardware. In terms of the Maxta partnership, the storage component merely becomes more network aware to build data resiliency on network topology. Why does Townsend want deeper integration of SDN? Because "…the same customer interested in building a data center based purely on hyperconverged storage and compute would benefit from an integrated network configuration," says Townsend.
Read more about Townsend's idea of an infrastructure brick