Read excerpts from a convo between Wayne Geils and Chris Rechsteiner on how to avoid cloud bill surprises, what smart companies do differently than everyone else, and more.
On cloud bill cost surprises:
Wayne Geils: There is a point where Dunning-Kruger becomes the curse of every IT manager. If you’re not familiar, it’s that law that says basically the less exposure you have to a technology, the more you think you’ve mastered it. And then only once you truly have developed mastery, you’re kind of like at the point of well, like, I know it, but it’s complicated. And that’s why a lot of these organizations, when they first get exposed to cloud and their teams get a year under their belt of running *a* workload, they’re like, “I can solve everything with the cloud!” And then a horrible bill shows up and bad problems happen.
Chris Rechtsteiner: If a cloud assessment, the upfront work, is done and it’s done intelligently, then those cost surprises don’t exist. True, an assessment might cost more. It might take longer. But at least your risks are known. And knowing prevents three quarters of the battle.
On vMotion really being as easy as it sounds:
Chris Rechtsteiner: In the transition of VMware to off prem, the term that constantly comes up is vMotion. And the mindset is all around the idea that it’s very easy. You just vMotion something from location A to location B. From on prem to the cloud, or wherever to wherever. Is it really that easy?
Wayne Geils: Yes.
On dev/infrastructure team biases:
Wayne Geils: VMware been around so long that people just take it for granted. In the maelstrom of the last five years of technology changes and cloud adoption, a lot of what VMware has done has probably been lost in the noise. A lot of people assume that if it’s not cloud, it can’t be efficient, it can’t work as well.
Chris Rechtsteiner: You’re in or you’re out, very binary.
Wayne Geils: Right. And I’ve seen firsthand in a lot of the organizations that we help with, the team running VMware is normally associated with the old legacy architecture. So they aren’t inherently listened to. They aren’t inherently given a seat at the table as development starts driving new directions. They become a cloud-first methodology instead of being a right-technology-first methodology.
Chris Rechtsteiner: You have a group of people who are established personally, professionally, and socially in an organization, and all of a sudden that changes. And it’s like, you’ve been a valued member of this organization for 10 years, and oh by the way, we’re going to a different technology, you don’t know shit. That’s not a technological issue. That’s a cultural issue. Because now you have these people who have tremendous institutional knowledge who are essentially just devalued.
Wayne Geils: Remember, like developers are the artistic. They’re creators. They embrace risk. And then we have the infrastructure team that is risk-averse. Their entire model is to not take any risks. That is a real problem in this cloud adoption world. Because now you have this team of hard-charger risk takers who have this bias, right? And then the senior management, the C-suite is like, “Well, the cloud team’s the new team, they’re doing the right solution because HBR said I should be cloud so I’m going to be cloud.”
On the potential of micro-tenancy:
Chris Rechtsteiner: What else do you see that’s going to happen coming into 2020 around the VMware on hyperscale platforms? What developments are out there that are going to further change the way people think about this as a solution and really influence their decisions as to whether they pursue it?
Wayne Geils: Across our entire industry right now, change is rapid. Everybody is releasing as many features as quickly as they can. So I see a lot of change coming. But the biggest one that I’m holding out for, that I think the most impactful, will be micro-tenancy. Today, there is some real limitations in NSX that allow a cluster to only have a single virtual gateway, a single routing instance to that cluster. And that makes micro-tenancy impossible. Once that’s solved, which VMware I think is aggressively actively working for today, I believe that micro-tenancy will open the door for a lot of adoption. Today, the barrier to get started is a bit stiff. You have to set up an entire SDDC, which is three to five nodes of dedicated — to make it work is 150-200 VMs, really a workload to move. And while there are many, many organizations that have that size workload, to do that as a new technology or a new model is a lot to ask.
Chris Rechtsteiner: That’s a steep adoption curve. Without a doubt. Especially from experimentation perspective.
Wayne Geils: Yeah. Which is why I think we’re seeing a lot of the early adoptions be disaster recovery plays or those kind of things, because that’s a safe container to do this technology test. So micro-tenancy will allow them then to work with service providers or other organizations that will stand up a much larger-scale cluster, give the access out to enterprises that want to test with this, but in much smaller 10-, 20-VM footprints. So they become comfortable with the technology before they do a larger, more aggressive adoption.
On the biggest challenge of using VMware on hyperscale clouds:
Chris Rechtsteiner: What do you see as the biggest challenge for organizations that are looking to use VMware on the hyperscale platforms?
Wayne Geils: I believe they’re gonna be surprised how well this works, how simply this deploys, how quickly they get resources running in a way that they didn’t think they could. But their IT teams are going to get distracted or befuddled in this entire smorgasbord of new hyperscale cloud services that have accounts a click away. And this silo thing we talked about earlier, that is getting to organizations that went too quick into cloud and built these big, multi-skilled teams is gonna be a real risk, where instead of keeping their team core focused, they’re going to be really easily distracted. And to avoid that, I think organizations that are going into the cloud need to build strong, focused relationships with managed service providers that are in that cloud.
Because that Dunning-Kruger thing, where your team will easily think they’re a master of RDS or they’ve completely understood Amazon Redshift and big data, isn’t true when you have a managed service provider. You need somebody, especially for production workloads. You need somebody that has been exposed to it, that has done it before.
On what smart organizations do differently than everyone else:
Wayne Geils: Smart organizations, organizations that are forward looking, are not worrying about the technology that delivers the solution to their business problem. That’s a secondary. They worry about how to fix the business problem. What is the most efficient way? How do they harness their applications? Then they look at what technologies can fit that bill, whether they be provided by services in the cloud, or managed service providers, or in-house-grown applications.