AWS re:Invent.

Part I

Arthur Clune
Impossible Dream

--

I’ve promised various people that I’ll blog about what I learn at re:Invent, but given the scope of re:Invent and a need to give some focus to my week I thought it would also be useful to list the areas I want to concentrate on.

  1. Large scale transitions

    We don’t have a massive estate, but neither is it tiny. We have about 1,200 VMs hosted on VMWare and a few hundred physical servers comprising a mix of legacy systems, HPC, storage systems (2Pb across Netapp, Dell Compellent), load balancers and other systems.

    We have a couple of nice, modern data centres so are in no rush to lift and shift, but I’d love to hear from people who have done that about the business drivers, challenges and gains
  2. Serverless

    The new hotness. Who’s using it? For what? How did dev teams take to the new paradigm? How stable is it as a platform?

    And, maybe the most important question: if serverless is not to be the new Visual Basic, how are they managing software deployment (CI/CD) and testing?
  3. Managing complexity

    New things offer the promise of reducing complexity. VMs promise not having to consider hardware, Lambda offers the hope of not considering VM and ORMs promised to mean not thinking about SQL. But in any large enterprise, each new layer accretes on top of the previous ones so complexity always grows.

    How are organisations that are big enough to have a central dev team, but not so big that they have many teams and can split legacy from the new handle this?
  4. Anyone doing anything in Higher Education

    Pretty obvious this one: cool ideas from HE. There’s some sessions on using Alexa in HE and I hope to meet some fellow HE IT people at the Executive Summit
  5. Cost control and security checks in a legacy environment

    Again pretty obvious. How do we automate security checks and keep costs in reasonable bounds? I’m aware of projects like Capital One’s Cloud Custodian but what else is out there? I feel like cost control could become a full time role, and maybe that’s ok, but what scale of spend justifies that?

It’s going to be a fun week!

--

--