dcdesignmyths1

They think it's simpler than it is

They believe everyone is already using SDN and Public Cloud

part of the misunderstanding is that it needs to be a total forklifting of current equipment. Which should be totally untrue.

Brand new will fix when there's a lot of existing to interoperate with.

There's also a belief that applications don't need to change or don't need to be a part of the change. Really all of the DC silos need to go through a transformation and hopefully a collaboration!

The greatest myth is that the only purpose of a network is to move data. The purpose is to manage the connection of data to people and things.

They believe a single vendor can solve all of their problems. In fact, Multicloud is Multivendor

@dfloyer Great point, David. We so often forget what's behind the purpose of the data center. Good to keep the human element in mind.

That the network is always getting in the way, slowing down new services deployments - it doesn't have to right?

The top myth of enterprise networks is that they exist to manage connections among servers, databases, and applications. In fact, they exist to ensure service levels to sustain knowledge worker productivity 24x7.

The second myth is that the cost of moving data is trivial. Moving data is expensive. Moving large amounts of data takes a long time. In general, it is easier and cheaper to move the compute to the data.

@jameskobielus We actually consider that to be really important. Thinking about things like customer satisfaction and application responsiveness is really important.

@dfloyer Especially when it comes to public cloud! We need to be very careful when designing for our multicloud futures.

Reducing risk & faster deployment of services

RELIABILITY - The answer to the automation question is always reliability. Speed of ops doesn't matter, if you're only speeding up problems.

The objectives of IT should not change. Automation is a how. Not a why. It lets us deliver more reliable and secure services so that work can be done.

We should be focusing on automation's benefits in ensuring closed-loop real-time predictive remediation of system, network, and applciation issues before they become showstoppers.

Automation should be used to reduce the operational burden. To allow the ops team to automate the day to day tasks

That it has to be single vendor

That some people don't need to automate. Not everything needs to be automated, but the more we move away from "snowflake" networks, the better off we'll be moving towards consistent and reliable workloads and workflows.

The most prevalent data center networking myth is that it's a lights-out automated scenario in most organizations. In fact, it involves expert human personnel in most functions most of the time.

Outside of IT others sometimes see that there is a clean handoff to cloud. In reality, even if a majority runs in clouds, you still need good tools to secure, control and observe the systems.

The most prevalent myth I see as prevalent in networking is that Edge data must be moved to a datacenter. Physics & cost, especially for extracting real-time benefits form IoT data, dictate that the Edge has to include significant compute, & be treated as equal citizen.

digital business is not a project, its a day to day reality that needs to be designed into enterprise networking critical capabilities

Digital means speed. So speed of change becomes a core attribute and that requires agility which means we need to simplify networks.

One of the keys to digital business is increasing the amount of data available to applications. This means enabling distributed compute environments to work in a coordinated fashion with low latency. A real challenge for network management!

I mean, honestly, it's all about revenues for the business. Likely the more businesses can embrace the digitial transformation, the better the customer satisfaction, maybe more profits, etc. Ent. Networking should help lead us there, not be an obstacle.

Digital business involves the transformation of all processes to online, real-time, data-driven, and virtual models. Enterprise networking is the digital backbone without which digital business would be impossible.

If we, as IT, truly embrace multicloud, we will have more reliable workloads and applications. As well as less shadow IT.

I thnik people have adotoped multiple clouds, each having specific tooling and capabilities which adds complexity - the goal is driving multicloud

@BrettLey2 Totally agree, Brett! Multiple clouds != multicloud

For resources and content to help you navigate to a multicloud infrastructure please visit our multicloud page http://bit.ly/2zHYQC...

In this short video we walk through everything you need to consider on the road the multicloud http://bit.ly/get-re...

There's a lot more people involved in determining the goals and needs while at the same time things are moving faster. So the engines of IT have to provide the needed functionality and guardrails for the new order of business.

good stuff dude, thanks alot !! @alamontagne

If joining a bit late, please hit play . to watch the video.

Good morning from early in San Francisco

it was a really informitive interview

As always, the answer depends. The equipment would need to be capable of some sort of IP Fabric features, I imagine. Perhaps EVPN VXLAN as that's the way the industry is going. Legacy also means applications, keep in mind. We need to bring forward legacy in all pillars.

It depends on how you define legacy. Legacy could be as simple as connectivity BMS servers to your network and extending there connectivity via a gateway to a cloud service

Does that answer your question?

We have clients with their core applications still running to legacy systems of more than 20 years old if more. Therefore for those clients we have first to move them from their local network to a virtual network and then upgrade to cloud.

Yeah, totally, which is why EVPN VXLAN helps us bridge that gap between legacy and future. We can handle layer 2 applications with the scalability of a layer 3 infrastructure