SHARE
Service Provider / October 29, 2015

OpenStack Summit Tokyo – A Quick readout on interesting trends, and how Gigamon can speed these up!

This week, all the geeks and netizens are gathered in Tokyo for the Fall OpenStack summit. Apparently 5000 of us from 56 countries made the trek out here to participate and I believe this is the largest one outside of the US.

Couple of interesting themes caught my attention.

  1. Telecom operators and Service Providers using OpenStack for realizing Network Functions Virtualization (NFV). Attended a very interesting session with NEC and NTT about how OPNFV wants to integrate with OpenStack.
  2. Enterprises making the case of using OpenStack to build large scale clouds. And contrary to a very badly worded article/rumor, HP Helion is not dead! The coolest hangout place at the Summit is the HP lounge, where we feasted on some delicious mochi.
  3. Containers support within OpenStack

There was also an interesting debate session about OpenStack’s transition from its original vision to become more like VMware. The question that we were left with at the end was “Is OpenStack becoming too VMwarized? 🙂 ” but I think the audience seemed to sort of agree that even though that might be true, at least the enterprises have a choice of which parts of OpenStack to use for their implementation, giving them the flexibility they need.

Coming back to the interesting themes above, a few questions to ponder. As Service Providers and Enterprises migrate to the OpenStack based clouds, how are they going to do the following?

  1. Secure these clouds for IDS, Forensics, SIEM etc.
  2. Ensure SLAs and compliancy for the applications that have been migrated
  3. Analyze and troubleshoot subscriber traffic, especially for a mobile operator.
  4. Reuse existing traffic/packet based monitoring and tooling infrastructure (CAPEX/OPEX savings)

An obvious answer to these problems seems to be traffic based visibility. But that may not be enough considering that OpenStack adds tenant constructs, for obvious good reasons (isolation etc.) which do limit how one tenant can access their traffic without impacting another tenant. Remember, no vSwitch access for the tenants to do port mirroring etc.

My colleague, Anil Rao, wrote a thought-provoking blog this week laying out the challenges of monitoring in a multi-tenant OpenStack deployment and how Gigamon is approaching the problem through Tap-as-a-Service (Taas) integration with Neutron. During the summit, Anil and I had excellent discussions with a couple of the world’s largest carriers and they agreed on the need for this kind of visibility from an operators’ perspective.

While TaaS works its way through the OpenStack community for adoption, we at Gigamon are not waiting. We also announced a tenant friendly visibility solution this week, to initiate discussions with the providers and private cloud operators at the summit.

Framework of OpenStack and Gigamon Cloud monitoring solution.

Their immediate response was ‘that’s nice, one tenant doesn’t need to pay a tax for another tenant’s visibility’. While we designed it for this advantage, hearing it from prospective customers/tenants using monetary terms, sounded pretty neat, which basically meant that the solution resonated with them.

With this solution, service providers and enterprises can safely migrate and even accelerate adoption of OpenStack clouds, since the tenant workload traffic can indeed be secured and analyzed.

Check out the solution brief for this new option. We are currently in beta trials, drop us a note or reach out to your Gigamon sales contacts if you are interested in trying this out.


}
Back to top