HybridCloud

harder to optimize your existing infrastructure

legacy DBMS-based apps.

compliance ? legacy stuff ?? good question I'm not sure what the top use cases are

A5. I don't think it matters except for governance and regulatory needs

its a connected world DCs will be in public/hosted clouds (on the internet), no more server-rooms & ROBO

maximize data locality, performance and protection

any traditional monolithic app should not really be on a public hyperscale cloud

What you can/ can't do is heavily influenced by what you will/ won't do.

It's a moving target. Public cloud is innovating faster than private cloud so that uses cases that are private only is shrinking.

Predictable demand apps are going to be less expensive on-prem?

Active Data will be close to its source, e.g. IoT, Exchanges, .. but preferably close to the internet backbone

apps that the physics gods don't like. IoT edge; complex apps with large write/read ratios.

Low latency apps (manufacturing, fin serv compute), many high memory, networking on disk i/o, data compliance reasons

The difference between public and private is often about the particular business and their perception of the cloud. For example, we see firms keep data on prem for better security, and we see others move data to the cloud for better security.

@yaronhaviv wait until the last question :-)

Customize hardware infra, networking models, meet data sovereignty needs

need to distinguish "on-prem" from Hosted, many private will move to hosting facilities

Today, just about anything *can* be done in public cloud today. The question is: why you would/ would not or should/ should not.

GPU, film/media rendering

At large scale that is definitely true. Reserved capacity in public clouds could narrow that gap but it's too difficult 4 most users right now to get a handle on.

@tcrawford bandwidth charges/latency

COST!

isn't the cloud dead? Edge computing is why on-premises will be very interesting (more than some of the legacy reasons) - here's the A16z video https://vimeo.com/19...

Reverse innovation - one can try new/innovative hardware approaches on-premises far faster/more efficiently than in public cloud. Big breakthroughs via this path are possible.

@nikiacosta u mean on-prem is cheaper?

@stu some think so!

customize hardware directly to software needs, e.g., google building hardware specifically for gmail ... most companies are not going to get to this size or need for a long time though IMO

At a certain scale, on-premises is far cheaper in many cases. TCO models are plentiful. Key is how one treats/manages data.

Saying on-prem is cheaper than public cloud or vice versa is not helpful without proper context. Would like 2 reports that show for every dollar spent for on-prem & for public cloud, how much money does it generate in revenues.

Excellent chat!

Excellent *chart*

Mostly private - lots of public and hybrid is new(ish) - depending on the defn

that was posted in reference to others asking me things

Interesting to see ERP on that list.

the @futureofcloud survey looks at some of the same private/public trends - here's the recent data http://www.slideshar...

@nikiacosta thanks for joining Niki...awesome comments

Fun to see authN and authZ on there, allowing 3rd party services to provide access to systems that span on and off prem.

@stu thx, I didn't have it right in front of me

what is the source for that table you posted please?

2017 State of the Cloud Joe Emison author
Report http://reg.interop.c...

@jgershater Looks like UBM State of Cloud Computing Survey Oct 2016

Workloads likely the most important area to watch given it's how many of this co's make/save $$

@lcooney Lauren is in the house

some workloads cost lots of money...so yeah

I've been watching for 15m :)

Data...data...data...

Data is a big one

driving customer engagement

Customer engagement, first.

A4: Product Functionality, Customer Engagement, Analytics, IoT not necessarily in that order

Or you can look at the types of workloads that have migrated already -- email, CRM, etc.

If it's digital its data and if it's digital it's often something to do with customers

#analytics and data are now a feature of all digital distruption

@krishnan but a pretty good stab at ordinal rank!

#IoT definitely. Machine learning on all kinds of data will accelerate

The most critical apps I see are focused on Customer Engagement. The other classes either touch CE or one of the other biz objectives.

Enterprise wants customers engaged, and with smooth UX, drive satisfaction, loyalty and up-sell

product functionality and the flexibility of the organization to use that functionality to adapt.

Many enterprises target streamlining biz ops as low hanging fruit.

@LawrenceHecht why those workloads lawrence?

Things like analytics and IoT may be more strategic but harder to tackle.

@plburris :-)

Seeing tons of Big Data, retailer apps (e-commerce), financial institutions, test/dev, and media streaming/delivery

It depends on what cloud your are referring to, for SaaS email and CRM lead the way... IaaS can be a different story

Analytics, AI, IoT are the ways to keep users engaged and happy

@tcrawford Not seeing systems of record moving over

those are workloads I have seen with lots of adoption in the public cloud. @dberkholz has seen data more than I

#IoT is key behind global GDP explosion from $4.2 trillion (2016) to $11.1 trillion (2025)

@nikiacosta Nope. Those are going to be some of the hardest for a number of reasons.

@tcrawford agree! \cloud reach facilitates engagement, opening up other business options.

SaaS is a big one. Almost all of our @ciscocloud customers are going hybrid

@Scott_Sin Lots of general business apps moving to #cloud. Those are less strategic functions.

Several CE and analytics apps which require simple scale-out to function well; or back-office apps which are very static. Agree that systems of record are staying on premises. However some apps are 'returning home' as on-premises s/w and scale get easier

Starts w internal apps -- big data and business operations

Seeing a ton of service provider activity around NFV/SDN

i personally think saas is the biggest game changer...it's not only disrupting internal systems, but it's also permeating into application design / microservices.