Accelerate your transition from traditional IT to the Cloud

Accelerate your transition from traditional IT to the Cloud

Question, all right we're talkingabout how you accelerate your transition from the old to the new.

How many people hereare responsible for the old? Good show of hands, thank you.

How many are responsible forthe new? There we go.

All right,I think you're in the right place.

Now honestly I didn't wantto call this session, How do you Accelerateyour Transition.

I really wanted to call it How to Avoid the Three Horsemanof the Cloud-pocalypse.

All right, anda lot of you are wondering, why aren't there Four Horseman.

It's cloud right,we already saved 25%.

So the focus here ison the challenges.

Just, I'll just give it a second,because we've got, come on folks.

There's plenty of seatsin the front here.

Come on down.

So we're gonna havean interesting session.

We're not gonna do a lot of talking.

We’re gonna do a lotof demonstration.

And when I say demonstration, we're not previewing some futurerelease of some future product.

We're gonna do live productiongenerally available product that we are currently using,all right, in our own business in orderto avoid the pitfalls and the challenges ofthe Three Horsemen.

What are those? So if you're moving fromthe old to the new, there are three big challengesthat most if not all organizations face inthis shift to cloud.

Number one, how do you migrate? Not just migrate a workload, buthow do you migrate at scale? So we're talking not justtens of workloads a month, we're talkingthousands of workloads.

How do you acceleratethat transition? That's number one,that's the first horseman.

The second one is, modernize.

So, you can lift and shift workloadinto the Cloud, but the reality is, is that you want to improve thatworkload, you wanna make it cloud-ready, you want it toleverage Cloud attributes.

So, how do you modernize yourapplications for the Cloud? So, that's the Second Horseman.

The Third Horseman is manage.

So once you're in the Cloud andyou have a large estate, what are the best practices,what capabilities? And so as we, somebody's laughing.

So as we go through this, y'all have different rolesin your organization.

And so there are peoplewho are responsible for driving Cloud migration.

I'm sure there are people herewho are solution architects, who are responsible for whatthose Clouds solutions look like.

And lastly, there's a new crop offolks, new roles that are emerging, and organizationsaround Cloud admins.

How do you manage that new estate? And so we're gonna show you and demonstrate based on the dayin the life of those people.

People responsible for migration.

People responsible to solution andpeople responsible for that eyes on glass day to daymanagement of that new estate.

So I hope you're allin the right room.

And you're excited to hear aswe are about what we're doing in this space in order to tackle that.

Now, I've waited untileverybody was settled.

I'm Michael Lebeau.

I have the pleasure of working forAccenture and leading our Accenture CloudManagement Platform efforts.

And so I have a large team, andwe've been driving that for a number of years now.

I have my partner in crime here,Rodrigo Flores, who leads our platform innovation.

I think of him as our CTO, and so I'm gonna askRodrigo now to take over.

And again, we're just gonnakinda drive through this.

>> [INAUDIBLE]>> Thank you, Michael.

Thanks for joining us.

You could have been anywhere.

You chose to be here with us.

I appreciate that.

By the way,that's a quote from Jay Z.

>> [LAUGH]>> So, we'll try to keep it entertaining.

Really, this is a great presentationbecause we're gonna really try to get a lot of time for demos.

So I want to start by setting thestage, what is the challenge that we all face, particularly between whatI would call the traditional IT, right and what we know howas digital business, and many of us I thinkare on both sides.

How many of you actuallyare managing both the old and the new at the same time? All right, how many of youwould like to walk off a cliff? >> [LAUGH] Right? So it's difficult.

You have on traditional side whatI sometimes call the IT list does, right? Change control,Friday meeting, go to the cab.

We wanna have everything ina service catalogue with request processes.

On the other side, you have peoplewho really want agility, right? They wanna be deployinghundreds of times per day.

They don't wanna go to a meeting,right? Cuz there's an old saying,I'm quoting myself here, which is if it's not automated,we gotta have a meeting, right? And nobody likes meetings.

So, to bridge this gap, to makethis thing work on both sides, you really need a cloudmanagement platform.

I know it's hard toread from the back.

We've proven that we are onlyeffective for the first ten rows.

After that, the learningdissipates very quickly.

So what we mean by a cloudmanagement platform is a set of tools thatprovides the automation.

Provides cost management,provides the migration tools, provides the blueprinting that let'sus set up Azure resource models and then deploy them consistently and automatically which we'llbe showing you later today.

It means also the managed services.

Just because you putit in the cloud, doesn't mean patchinggoes away right? Who loves patching? One person.

Thank you sir.

We all want to talkto you after that.

[LAUGH] So, sowe need is this management platform.

And what are the essential elementsof a management platform for cloud.

Number one,you need freedom with control.

And what do we mean byfreedom with control? The fact is that on the right side,the people who want agility, the DevOps model, right.

They really wanna be able to deployevery time at the push of a button without going throughcomplicated processes.

On the other hand, the complianceofficers still wants their logs.

They still want their bit of flesh,yeah.

And somewhere along the way somebodysays you need an approval process.

Sir, you agree with me? You would need an approval process? [LAUGH] Nobody needs an approvalprocess in the Cloud.

So we need this balance.

So we need a platform thathelps us do this balance.

Cost management.

After security, the biggest concernpeople have is cost management.

Why, because it is easyto deploy in the cloud.

It is easy for some one to fatfinger,type a thousand instead of hundreds VM's andall of a sudden you get a big bill.

Or this auto scaling stuff, rightwhich sometimes works fantastic when your website is being hit, butother times you get a big bill.

Should be delivered as a service.

If you're really using the cloud,and you're really going to public cloud first then I don't know ifyou heard our CIO on Monday right.

Accenture today is 60% in the cloudfor our internal applications.

Public cloud, not private cloud.

I'm talking about public cloud, 60%.

So once you're up there, you needinsights, because things go bump.

The vendor moves you tothe wrong write card.

You get an email we're gonnahave to vacate your host.

So things happen, andwhen those things happen, you need those actionable Andintelligence and expert guidance.

You need somebody to help you,and that's what we provide.

And finally, of course you havesecurity and service operations.

You still need tointegrate with service now.

You still need to integratewith Microsoft System Center.

You still need to haveties to security.

None of that goes away.

That's what we mean bycloud management plane.

A consolidated cloudmanagement plane.

And the basics of this are, ofcourse, on the other side you have the public cloud and you stillwill have the private cloud.

And you have your people,process, and technology, as you've always had.

And on the other side, we now have to be able tointegrate with those clouds.

So you need to have systemsthat let you create and integrate with multiple clouds.

You need to have your processes.

That's where orchestration,automation, configuration management,all that good stuff, Puppet and Chef and PowerShell, etc.

The cloud management services whichis your backup, your patching, antivirus security,the management platform, the cost provisioning, etc.

And then, of course, the help desk.

We have to run a consolidatedhelp desk for all of Accenture.

So when somebody has a problem,they know where to go and to get help right atthe proper level.

And all of this is in orderto enable three major things, migrations as Michael said,modernization, and then of course, finally when you're steady state,management.

So what I want to here is introducemy colleague, Sean Peterson, and he's actually going totalk about migration and we will have a demo of that.

All right.

Thanks.

>> Thank you Rodrigo.

>> [APPLAUSE]>> So I have the pleasure of introducingthe first horseman, migrate.

So when we're talking aboutmigration, as Michael had said, this isn't about migrating oneworkload, one VM into Azure, we're talking about migratingbusinesses on the cloud.

And that journey involves fullevaluation of the business case, it involves evaluation of theapplications in the infrastructure to be able to rotate the businesspredictably, safely, securely, reliably into cloud that takethe benefits of that cloud.

So, when we talk about that,I've had the pleasure to work with a number of companies tohelp them on that journey.

And one such company we've workedwith is a global resources company.

They have sites all over the globe.

They have 2,500 VNs.

They have 800 applications.

And they need to rotatevery quickly into the cloud to be able to take advantage ofcost, agility, and innovation.

And so we worked with them toevaluate what goes into Azure IS, what goes into O365, what goesinto a maybe a modernization and uplift that into Azure Pass.

And as you can imagine,that journey is complicated.

We need to structure that in a waythat we don't disrupt the business.

We move efficiently.

We do it safely.

We do it securely.

We do it reliably.

The clicker, so when we do that,we certainly need structure in industrialization and that startsfor us with the methodology and that methodology starts withengagement with the business.

What does a business need for this? How is cloud to kinda deliverthat cost optimization? How is it gonna deliverthat efficiency? How is it gonna deliver agility andinnovation? And then it's a run through of allthe applications that company has.

And understanding how thoseapplications whether as they are, modernized, new applications, new SaaS services can enablethat agility, that efficiency.

And then we work through with IT todetermine what are the systems that support that today.

And then what are the cloudcomponents that need to support that tomorrow.

And we need to structure that,bundle that, understand not just which VM moves,which application moves, but how we do that in a bundle in a waythat doesn't disrupt the business.

We can't, through theseindustrialized migrations, kill the business.

We can't take an outage.

You're in closed, has to occur.

We've gotta get the migration donein Windows that don't disrupt that and it goes throughthe full life cycle.

We determine whatare the right dispositions for all those applications in bundlesand drive that whether it's a remediate the app and move,whether it's move then remediate, whether it's modernize the app,light touch, whether it's rewrite the application to take advantageof Pass services or SaaS services.

And then certainlyit results in run.

We can't just migrate and walk away.

We have to do the migrationin a way that allows for clean cloud operations as wecontinue to drive the value of that.

So that in itself requiresindustrialization, as I mentioned.

That we can increasingly automate,organize, and drive through industrializedsolutions that we've developed.

So, got my friend David Millshere who's gonna talk through some of the capabilities that wehave developed to really drive what I just described that wholemethodology from planning to execution in industrialized fashion.

>> Thanks Sean.

>> Yeah, take it away >> All right, so I'll apologize up front.

This isn't necessarilya really sexy demo, but I think it is an insightfuldemo and it's an important demo.

And if you think about what Seanwas talking about, there's so much involved in an applicationmigration, especially when you've got a lot of them and you need tokeep the business up and running.

So, generally speaking you'vegot the inventory process, that can take a long, long time.

You're collecting all the data.

There's high level planning once youkind of map the dependencies and figure out what the business useis for the different applications.

Then you get into detailed planning.

You're trying to figure out, whatis it worth moving or not moving, what should you lift and shift? And then finally,you get to the execution phase.

But you really need a place tocollect all of this information and make sure it's organized and make sure that everybody hasthat same source of information.

And they're staying up to date asthey work on this migration and go through the process.

So the tool that I'm sharingout now is what we use.

We've been using this for I think we've done over 200,000application migrations doing this, some in the thousands, some inthe ten thousands that we've used.

So this really has tobe a repeatable and industrialized process.

I'm not gonna bore you by walkingyou through a questionnaire and here's how we gather informationto take inventory and all of that because I think manyof you are familiar with that and it would take a long, long time.

But I really wanna key on someparticular areas where we've taken those learnings and the thingsthat we've industrialized.

For instance, just templatizingsome of the processes because you really need to map out theprocess that needs to happen to move an application prior to during andafter that migration.

What's the right point to cut over? What are the dependencieson that application? And so this is just an example ofsome different templates that we have that we've put into the systemthat are reusable over time.

These can obviously be combined.

But you can see that there'sall the different steps and dependencies and at what point doyou need to start those activities to meet your T0 date whenyou're gonna start migrating.

So needless to say,you can use all of this and it's probably gonna cut about, I think we've learned 50% of thetime it takes to do that migration if you can just codify some of thisstuff that you can templatize it and industrialize it and then makesure that you can reuse it.

So given that we have all ofthat stuff in the system, if we go to some of these migrationbundles that Sean was talking about, let's say I have a wave that I'mgonna migrate in Q3 number 4, there's some bundles of resourcesthat are associated with that.

And I can search and pull those up.

Now if I click into these, I cansee that I have a lot of different applications, somedifferent instances.

And even a service thatI need to move over.

And what I've considered in all ofthis is what do I need to do when step by step to not bringthe business down or to have minimal downtime forthe business.

Now if I want to see and actually go to the next step,I can actually select these and put these on a design surface andsee what the relationship is.

So if I drag these over, and this isthe kind of sexy part, all right, so here's where you get some UI,that's a little bit of eye candy.

If I hit expand,you can see all of these things, all of these activitiesthat have to happen and they have to happenin a certain order.

Now they're mapped out here forobviously reasons, so we don't bring things down.

We've also built someintelligence into this.

So let's say that I drag anddrop something that doesn't belong, it's automatically gonna tell me,you can't do that.

So you need to go and back that out.

So I'll just take this cursor anddrag it back where I got it from and then everything is fine.

So this is really useful forpeople to be able to go in and get a visual representation of allthese steps that need to happen in this wave with all thesedifferent applications and resources and make sure thatthose are gonna happen on time.

So, once I get over to the executionphase, let's say that I've got everything tied up and sequencedcorrectly according to the correct processes, now I can go toa place called My Tasks.

So if I'm the person working on themigration, like I said it's really important to have that one sourceof truth, I have that here now and I can go through and see all theactivities that are assigned to me.

I can tell whether they're active,I can also tell whether there's prerequisites that haven'tbeen completed yet.

And there's predecessor tasks andso I can go in and figure out what I need todo to make that happen.

I can also drill down and finddetailed notes in the same place, related to this task, if you need tohave screenshots that somebody takes and owns that application, youget it from them, insert it here.

If there's an issue with that task,if it was causing a problem, you can also log an issuein that location.

So, if I'm a project manager, nowstepping outside a little bit, and I want to know how things are going, this is also the onesource of truth.

So, if I go and look at, let's say I've got some bundlesthat I have in a wave here and I'm a project manager,I've got my bundle activity.

I can see a yellow line.

I know some of thesenumbers are small, but just look at that yellow line.

That's T0, right? So that's when youkick everything off, that's when the migration starts.

And all I have to do is look lefthere and I see all these blue boxes, if everything blue is over here,I can at a glance as a project manager go in andsay things are ready to go.

If I do have issues, I might see orsomething I need to follow up on, I'll see a maroon box over here.

Somebody needs to disablemonitoring, right? So I can go back in andfollow up on that task.

But it's a really goodhigh level way for project manager tosee what's going on.

Now, if I'm somebody who'smore executive level and I wanna see how things are going,I can look at a report that really gives me kind of atimeline and maps my progression of actual migration activity againstwhat I'm supposed to be doing.

So this is a really quick andeasy way for me to go in here and say okay, the dark blue line orthe more bright blue line is how I'm tracking and the expectedprogression of this migration is the kind of grayish linethat you see up there.

And I can go down some moredetails below this report, but at the very top,I can look at this and see, okay,things are pretty much on track.

Everything's fine.

So, that's really where I'm doingall of this planning, if I'm working on the migrations, if I'm somebodywho's in project management, even at an executive level.

And you're tracking all thoseactivities that have to happen before, during, and after.

We also are connecting withautomation around the actual migrations themselves.

And one of the places we dothat that I'm gonna show you here in a second,is over in the management platform.

Where we have somethingthat's called Cloud Mover.

You can go in here and actuallyselect something to do one off migration, this can be triggered bythe migration planning tool as well.

So, if I go into the catalog and Isee Workload Migration, I can go in here and fill out a simple formof some information about, where is this resource coming from,where is it going to? All I have to do isclick on move my VM.

I've got some forms here andwe won't sit here and wait for it to happen, but I can put in somebasic information about charge code, where is this gonna be going.

I'm going to send it up to Azure,what's the source server name, put the source server name in there andwhether this is gonna be manual or automated, meaning it's gonna betriggered by an external system, and I can kick that whole thing off andget it running.

And once I do that I can actuallylook at how I'm tracking against the progress and it's showing methe image of origin is being copied, there's a new instance being createdon the cloud you're moving to.

And you can have it actuallysynchronize on an ongoing basis so you keep both servers up and runningand you have the least down time possible with the server that you'regonna be migrating and then it does a final sync right before cutoverand spinning up on the new server.

So basically I've gotan application that's running, a web application that's on mydata center, and I'm moving it.

You're keeping that up and runningas long as possible, it cuts over, and then it's gonna pop up in Azureonce it goes to the other side.

So it's a real quick tour ofsome of the tools we're using.

But hopefully this gives you an ideaof some of the ways we're taking the methodologies and we'retaking some of the expertise and we've actually builtthem into a platform that helps you reduce the timeto migrate workloads.

And it's gonna reduce the chance oferrors once you get things over.

Great, thanks David.

Yeah, thanks David.

>> [APPLAUSE]>> [INAUDIBLE] >> Yeah and so as David had just shown, this isjust a little bit of an insight of how we work to manageall of that complexity.

If you think about what ittakes to do this, we're talking all of the IT admins, we're talkingall the application owners, thousands of workloads, hundreds ofapps, dozens of business units and being able to migrate thatwithout disrupting the business.

That's what,through our experiences, through our lessons learned overthe years, that's what it takes.

And so these are the kindaresults that we've seen.

We can cut work loadmigration by 50%.

That again is not justthe automation on migration or the workload that's important,but all that people coordination, all that time coordination, status reporting, all of that donethrough these kinds of solutions.

And the industrialization aregetting up to thousands of workloads per month, because the idea is toget from the old to the new as quickly as possible,to get to those cost savings and flatten the overall time it takes toget there, but do so in a reliable, predictable manner.

So now we're gonnatalk about Modernize.

That's great that we can do allthat migration, migrate into what? We're gonna be migrating intothe cloud but the cloud is not just a generic thing, it has to be veryspecifically built for purpose.

And sohad an opportunity to work with yet another customer very recently whois a global beverage provider.

Somebody you'd all recognize,probably partook of it this morning, and so, orsome of their products this morning.

And so through that what we'velearned in working with companies like this is they had a specificscenario of needing to deploy some kind of collaborationapplication to hundreds of retailsites across the globe.

And they've been doing that ina manual way inconsistently, and that really broke down a lot ofthe collaboration that they were attempting to achieveacross these retail sites.

So, working throughthe methodologies that I'd introduced before,we landed on the need for SharePoint extranuts in those sites.

But we had to do so again ina way that was cost efficient and very quickly to be able to deploythem, to be able to update, to be able to manage them.

And so we learned through thatprocess and, again, in working in other scenarios, how do you do thisin a much more enterprise fashion.

In ways that are configurable,dynamic, and be able to do that, though, with consistency and not lose that trade-off ofreliability standardization and do so in a way that bringstogether the power of the cloud.

And so, digging just a bit deeperinto some of the methodology that we talked about previously.

It involves evaluatingthose applications for that business use case todetermine what to do with it and we just simply titlethose as the six R's.

Do I remove the app? Do I really need that application orcan I kill it? Rationalize, when wetend to go through and look at all these applications,we find there's a lot of redundancy.

And so sometimes it'sa combination of remove, collapse some of the functionality.

Maybe move toa consolidated platform.

And then relocate and some of thetooling that David talked about is the methodology and capabilities weuse to really industrialize that.

I'm gonna take advantage oftechniques of doing virtual to virtual, physical to virtual,sometimes I just do lift and shift.

A lot of times that takes the formof location to cloud-adjacent co-location to be able tomigrate something if I can't yet get it into cloud butplan to get it there very soon.

And then we start to moveinto some of the more higher business value activities butwith some work and tradeoffs.

So those tend to bethe Cloud rewrites.

Do I do a light touch to beable to take advantage of some Cloud capabilities like Paths,where I have to remove maybe some synchronous communications forasynchronous or take advantage of queueing servicesall the way to cloud native.

These tend to be your applicationsthat you're looking for high degrees of scalability, you're lookingto take advantage of containers, service fabric, andthose types of capabilities.

And you will get the benefit of thathigh degree of dynamism, that high degree of agility with thatapplication to support the business.

And that's what we then do,as we talked about before.

We engage with the line of business.

What are you trying to achieve? Are we just trying to optimize? Are we trying to delivernew capabilities faster? So in the case of this beverageprovider that was exactly it.

We gotta get collaboration better,we gotta get it done globally, and we have to get it done fast.

So they were looking to harness thepower of Cloud to take advantage of that from a business opportunity,not just cost takeout.

And then we start to evaluatethat very systematically.

What's this goingto cost to move it? To modernize it? What are the benefitswe get out of that? What's the riskassociated with that? Are we gonna disrupt the business? Is there a high degreeof technology risk? And then what's the timelines? We have to fit this insidethe business window.

These modernizations and migrations don't take place whilethe business takes times off.

Everything increasingly is movingmore quickly, and so we have to fit these into the right windowsto execute it very carefully.

And then lastly, it's the benefit.

I seem to be stuck here,there we go.

And it's the benefit.

So it's being able tounderstand what are the cost, risk, time, trade offs to the overall benefit align the business.

And that's how we determine the ROY,we determine the dispositions and we set the road map that we canthen execute it in tooling like David's short.

And so what it take to be ableto pull all this together.

So what we've learned is it takesa lot of cloud architecture acumen.

How do I pull togetherthe infrastructure, the application platforms to be able to tie the best of AzureO 365 together to be able to provide a safer level platform that'sgonna be able to execute this.

It takes understanding ofthat application workload, the characteristics, the IOPs.

Is it sensitive to outages andreliability? Do we have synchronous orasynchronous communications? When the business drives workload,how does it scale up, how does it scale down? All of those things need to befactored in to how I architect that in the cloud.

Then lastly,it involves the cloud lifecycle.

How do I do deployments of thatinfrastructure and application.

How do I do updates to that? How do I change it? How do I modify it? All these things have to cometogether in an increasingly codified way with infrastructures,code and configurationist code to be able to do this veryquickly at enterprise scale and a very dynamic configurable way.

And I'm gonna introduce mycolleague here Ron Gilpin.

Ron's gonna share with us ways thatwe've industrialized that beyond just subject matter expert in eachof those areas with script showing up but in ways that we've workedto quotify that using the best of the some of the tools we have in theAzure platform to do this at scale.

Okay, hey, thanks Sean,appreciate it.

Okay, so we're gonna walk througha demo that Sean just pointed out, which is for a global copy company, it's a project we delivereda little while ago.

And we used a solutionframework that we put together.

Up on the screen, we have a coupledifferent marketplace items and I know we've all seen marketplaceitems out there that have solutions like this,you click on it.

But basically they're sandboxes.

Our team, our projects we'reworking on, we don't do sandboxes.

We do enterprise platform solutions.

So we have to go way beyond justinstalling a single machine that has all the surfaces on it.

So that adds a lot ofcomplexity to your solution.

So we need to come up with an endto end cloud platform design and deployment solution.

So that's why I'm gonnawalk you through, and I'm gonna walk youthrough SharePoint.

If you know these othersolutions like Sitecore, SAP, or think about something else,think about the process as I walk through this that you wouldhave to do for that solution.

We have this built forthese other solutions as well.

Okay, we're gonna kick it off here, stick with me on this one becausethis is the really cool part.

In fact there are actuallya couple of cool parts.

David was mentioning where his demowasn't sexy, this is a sexy demo.

>> [LAUGH]>> Okay the first thing we do is we have actually have twocomponents to our solution.

We have a designed engine componentwhich is since she's gonna help us design the NDN solution andit's gonna sign a base on reference architecture thatwe work with Microsoft, and also the application platform SMEsto actually come up what should the cloud architecture bewhen you're gonna go in.

So it's taking all Microsoft's andthen best practice and built this in.

The second piece is wehave a deployment engine.

The deployment engine is basedon Microsoft best practice for how you deploy solutionsto the clouds.

Basically, templates and scripts, sowe're doing infrastructure as code and we're doingconfiguration as code.

So we're using templates and DSC within our framework toactually do our deployments.

On every project we startout we're gonna identify.

This one was a retail client.

And this was fora SharePoint Extranet portal.

So the solution is SharePoint.

But we could click down ifwe're gonna do something else we can select another solution here.

It's kinda small there.

I may have to blow this back up.

And so, anyway what we're gonnado next is we plugged information because we're have to deploythis into a subscription.

So we have the subscriptioninformation here.

What we would do is int heclient is we would go ahead and put in what theirsubscriptions were.

If you do a drop down pickthe subscriptions that you want to deploy to.

The next one is we're gonnaset up a solution itself.

So the specificationsthat we actually need for the solution, load that back up.

Is that we know that we're gonna doit for the UAT environment first.

So we're gonna designthe UAT environment, we know it needs highavailability to it.

And we have to give itsome information to help the configuration part of this.

So we'd need to know whatthe network class is we're gonna use to set upour cloud environment.

The next thing is over onthe left hand side, for those of you who know SharePoint, we have to identify what serviceapplications we're gonna install.

So for this instance we're gonnado Search, Management of Data, User Profiles, and State andDistributive Cache.

I could turn any of this on or off but, for this solution,we needed all of these.

This is where one of the cool partsof our design engine comes in is that we're using the usage capacityfor performance metrics that Microsoft has provided for helpingyou design your platform to support the user and the applications thatyou're gonna be deploying on.

So in this case, we have 20,000users that we need to support.

So SharePoint, if you guys knowSharePoint, it's all based on user count, requests per seconds,and concurrent users.

So we've identified andbuilt that into the design engine.

So we come down here where we'llhave what the user count is, what our total, andalso what our peak is gonna be.

So what it spits out isat the bottom here is what is the web we need? The minimum to support the solution.

So we only need twoweb front ends for a minimum because wehave high availability.

We couldn't do it with just one.

But we only need two to support thisparticular solution just based on the requirements we put in there.

So next what we need to do is we'regonna move onto the configuration, we need AD information, we know thatSharePoint has to connect to AD, we know that we have tocreate service accounts, we know we have to do domainjoined servers in there.

So we need to have atleast privileged access to the Active Directory domain, we getthe credentials in here, and then we identify what our debate is gonna beand put our credentials down here.

Next, we're gonna move on to isthe SharePoint configuration cell.

This is the information you have togather whether you're gonna deploy a single farm with 1 or 2 servers oryou're gonna do 1 that has 20 or 30 servers.

You have to have the baseconfiguration that's gonna set up your web servers, your app servers,and your SQL Servers.

So on the left hand sidewe have our product key, a pass phrase, and then what theversion, we have CU, so we actually do support different CUs that needto be supported for each deployment.

Then, identify what the portwe're gonna use for this.

On the credentials side, SharePointneeds specific credentials for specific service accounts.

So again, we have to set these up.

When you move downinto the service and the database names, SharePointhas a lot of databases and a lot of people try to shortcutwhich databases they're gonna use.

This is based on Microsoft'sbest practice where, whats gonna get setup.

Same thing for the services,we have the user profile services, the management data and the search.

Then on the left hand side hereyou can see these other services.

So once we have allthis other information, it actually kicks out what the nodesare in the configuration's gonna be, so we know that on a note count wehave two servers, two app servers, two database servers.

This is high availability group sowe also need a server in here.

So this is what's gonna be deployed.

This was put together bythe design engine based on the input that we've created sofar, so the next thing we need to moveon to is the SQL configuration.

Cuz SQL is a little bitseparate from SharePoint, we do need to do a specificSQL set up first.

So we have to provide allthe information that's needed to actually do the highavailability server, the instances, the availabilitygroups, and then also the server.

So this next one is a summaryof what we just did.

So at the top here itshows this DNS summary.

So we know that we're gonnause an address space, class C address space.

We have the DNS serverswe're gonna use.

We have the domains thatneeds to be joined to it.

On the right hand side here,this is our network.

So you're gonna notice thatthis has kicked out three, actually four separate subnets.

The best practice recommendationis that we isolate each one of the server roles intodifferent segments.

So that's what this referencearchitecture is actually built.

So we have one for the web,app, database, and then we have gateway subnets forcommunications.

Then next here we havebasically a bill of materials, we have the web front ends, tellus how many we have of each one.

We have what the sizeof the server is.

We can change that if we need to.

And then, we have the IP address and the storage typethat's gonna be used.

And then, down at the bottom, we have our service that'sgonna be installed.

Now, this is kinda key here,because these services, it knows which serversto install it on.

It knows that it's gonna installsearch in metadata on the web frontends and it knows userprofiles from application server.

You don't want to installservices on the wrong servers.

So once we have this, the nextthing we can do though is for those of you that have built strips, whilethe strips need deployments and you do full network builds andthen application deployments, you click on it, three, four hoursdown the road, it fails on you.

So what we did is webuilt validation in here.

So when we do this validation, it'sactually running this against Azure, all the configuration fromthe resource group, the VNET, the subnets, the servers, the OS s,the SharePoint, it's running that against Azure to make sureeverything is actually gonna work.

So that we don't get three, fourhours into it and have this fail.

So we have a completesuccessful pass here.

The other thing that's kinda cool isthat it's great being able to see this on a page, but architects, welike to see a visual representation.

So what we're gonna do, when weclick Visualize, it's gonna bring up a visual design of our solutions.

So it's drawn out exactly whatis being built and how it's gonna be built in Azure, so again,we have a resource group up here.

I can validate this as our VENT,I have separate subnets for each of these.

I know that since it'sgonna be high availability, we're also buildingin availability sets.

So we have that foreach one of the server roles.

We have our database andour witness server.

We have a load balancer that'sgonna go for the web front ends.

And then, we have our gateway forcommunications.

This is everything that you needto deploy an inter price solution.

So this is justa reference architecture.

And a lot of people get out andsay, well, this is great but I have clients orwe have internal requirements that We need to do thingsa little bit different.

So what I'm gonna do is I'm gonnasave this configuration to my reference architecture.

And I'm gonna come out here andjust visualize this again.

So we're gonna see what we have.

So for this particular client thatwe had, they didn't wanna have segregation, and isolationbetween their app database, their application servers andthe database servers.

They wanted those in the same VNET.

But they did want an isolation forthe web finance servers.

So what we wanna do is we wannatake these application servers and we wanna move them to the otherVNET, actually to the other subnet, and then we wanna clean them up.

So I'm gonna go into the nodes here,and this over here on the left-hand sidegives us the ability to go through to customize whatever weneed to in this design, whether it be resource groups,the virtual networks, subnets.

We can go through and customize allthis using this left hand column.

So I'm gonna go intothe app servers.

I'm gonna click this.

This is all the information thatgoes in for designing and deploying those app servers from the VM thedisk settings, the network settings.

So the network settings, I'm gonna go down to the subnetsthat it was attached to.

And I'm gonna change this andput this into the database.

Now that I have that in the databasewhat I need to do is I need to go clean up that subnetthat's out there, so I'm gonna go to my virtual networks.

Go and edit that.

Gonna go down, find my applicationsubnet, again I'm gonna remove that.

I'm gonna save it.

I'm gonna come back and I wanna takea look and see what I just did.

So as I visualize you can see now.

We've actually takenthat application subnet, moved all the application serversover to the other subnet, removed it and cleaned it up,may be able to validate this.

This looks good,this what I wanna deploy.

I could click ondeploy right now and it'll deploy to Azure cuzit's also validated it.

Now, I'm not gonna doa deploy right now because, it's lot that going on here.

Deploying all the network,the resources and the SharePoint takes about four orfive hours to deploy.

But think about it, 4 or5 hours to do this and we just designed thisin under 15 minutes.

It's pretty cool stuff, isn't it? >> [APPLAUSE]>> Another challenge that we run into is that,sure, we've done a UAT.

We need to create production.

We know that production UAT aregonna be very similar to each other.

So what I'm gonna do is I'm justgonna make a copy of those.

I'm not gonna go throughthe whole design aspect again.

This was already the referenceddesign we've already optimized it for the customer requirements.

Now I have this copy andI'm just gonna go and rename this.

I know I have to makethis production.

Not only am I gonna make production,I'm gonna version this.

So the full featurewe've added in is we can now version all our environments.

So when I save this.

Now if I were to visualizethis again just so we can see this is a production.

So we got this configuration.

Thing about production thoughis that we got down the road a little bit,we figured out that, okay, two web printed servers toapp servers was not going to handled the application load thatthey had designed for the solution.

So we did someadditional calculations.

We needed four web servers.

We needed three app servers.

So what we did is we wentin here to the nodes, Let me go to the web nodes.

We have a count of two,I'm not update to this to four.

And then application servers,I'm going to go to those, and I'm going to make three app servers.

Okay, I'm gonna come back out nowand I'm gonna visualize that.

So this is what I have now.

So over here you can see I havefour web servers been added and up in the application service Ihave three application servers.

So what I'm also goingto do with this though.

I'm going to go back here andI'm going to edit this.

And this is actually Version 1.

1.

So what this allows me to do isonce I've deployed Version 1.

0.

and it's out there.

When I have Version 1.

1 andI'm ready to go.

I can then go deploy this andall it's gonna do is it's gonna go out andnot deploy the whole solution.

It's gonna go out andsee what's already there.

We used both partial andincremental deployments.

It's gonna take a look what's thereand only add what's been changed.

So we added web servers.

We added app servers.

But the really cool thing,we didn't just add a web server and app server, it's configuredwith the desired state for that specific role.

So we don't have any inconsistenciesin our environment now.

With servers for the same roleshaving a different configuration.

And that's actually an areawe found in the cloud.

When you deal a lot of scripts whereyou have a lot of discrepancies.

Where you are finding a lot ofservers roles that are inconsistent.

And what happens there,you go to deploy the application.

The application starts to fail.

This is to ensure that yourinfrastructure solution that support the business applicationis designed consistently, repeatable, and efficiently.

So we can go ahead and deploy thisand it would just version it in.

And if we needed to, if there was a problem,we just pull up the version.

We could fall back in a matter ofhours, because instead of going out and trying to fix things,we'll go back and redeploy it.

We actually have thatwith this client, where we actually did deployment,the configuration is well set.

They had gone in andmade some adjustments, and it blew up their servers.

They called Microsoft support,it has taken hours for them, in fact it looked likeit was gonna take days.

We just went back in, did a redeployment, five hours,they're back up and running.

I mean this is really valuable beingable to do this to an environment.

And it's gonna show youwhat the resiliency of your environments gonna be.

Pretty cool stuff isn't it? [APPLAUSE] Where do you seein deployments like this? [APPLAUSE] So the thing I wannawrap up on this, is summary.

So we have the designengine that does the whole reference architecture.

We have our deployment framework,but notice we didn't go out write anymore templates, we didn't have to write any scripts, this is alldone through configurations.

So we can do end to endjust configuration driven.

Thanks guys.

>> [Applause]>> Yeah.

[APPLAUSE]>> Great, so, as we lead into this part, Cloud architecture,you saw Ron walk through the design.

Not just an initial deployment butthe configuration of that.

You saw Ron walk through bringingthe SharePoint workload into it.

How we size it with the actuallife cycle of it and then you saw how wecan do a deployment.

We can do an update.

We put something in the wrong spot,let's move it.

All at enterprise scale,all industrialized.

And what we have found is, in theexample that we talked about here, all that architect time,all that engineer time, the mistakes made,we reduce it from weeks to days.

75% reduction in overalleffort to get it done.

Dramatically improve quality.

It's pretty powerful stuff,and so if you imagine, you're trying to do,in this case, what was it, hundreds of extranets around theglobe in a very consistent fashion, this is what we believe ittakes to get that done.

Yeah.

[APPLAUSE] So with that, I'm going to now turn it over to, I believe,Rodrigo to talk about manage.

>> Great, thank you.

,>> Thank you for, >> That was pretty cool, actually Ron.

So in terms of managementwhich is the third horseman of the apocalypse,of the cloudpocalypse.

So we've migrated, we've modernized, we have our nice templatesdriven off automation.

And now what do we need to do? We need to start thinking abouthow we're gonna have on-going management of the stuff.

Right, so we some sortof centralized dashboard is what we found toadminister our cloud state.

Which is, I'm not allowed to saythe number but it's tens and many more of thousands of virtualmachines and past services running on as well more in other clouds.

So we need to have a unified portalsomeway to have a centralized dashboard to keeptrack of the health, see the state of those machines,be able to do what we call QADA, questions, answer,decisions, actions, right? So something happened, what are thequestions that we should be asking.

Actually one of the new thingsabout this cloud stuff is because it is new.

We don't really sometimesknow which questions to ask.

Things happen.

The cost went up 25%, what happenedwhat happened on Thursday.

Did we deploy more machines? No, no what happened nothinggot deployed that day why did it go up 25%? We need some analytics tool right.

And by the way that's a realuse case, the answer was the vendor moved us to a differentprice list temporarily by mistake.

So that bill that youget in Cloud might not be the bill that youactually should be paying.

Little things that youmight want to know.

So we also want you to seeresource and data transparency.

You know,people want to know who did what.

How did that get done? And at the end, it's really comesdown to a very simple topic, which is how are wegonna match this cloud? How we're gonna apply the policies, what kind of governanceare we gonna have, right? How do we know thisis fair consumption? How do we know peopleare staying in budget? As I said earlier,cost is one of the big issues, and my colleague David here is gonnashow you a little bit of that because the problem is from a CFOperspective they need to allocate, you get a cloud build fora couple hundred thousand dollars or a couple million dollarsthe question is who ordered what? Is it valid?Right? To which cost centershould this be assigned? Why is that important? Well, because it's this little thingcalled Sarbonnes Oxley, you know, the go to jail card, which is oneway you can cook the books is by assigning cost to places thatmakes a business look good by having less costs on businesslooks bad, but it's okay.

So it's very important to be ableto assign cost to the proper cost center, so with that I want to kind of turnit to my colleague David here.

Whoops, I was a littletoo quick on the.

>> [INAUDIBLE]>> And do a demo here how wemanage our cloud right.

So David here is playing the roleof the cloud admin right.

Go ahead David.

>> Thanks Rodrigo.

So I think you said I wasthe third horseman and one of those horseman washeadless Michael said.

So I'm the headless cloud admin.

And I'm riding around.

Can you hear me? >> Yep.

>> So, All right, so I'm the headless cloud admin and myjob is to manage my cloud resources, control my costs and make sureeverything is secure and protected.

So I'm gonna show you how you can do that with the cloudmanagement platform.

So one of the things, I can't managewhat I don't know what I can't see, so in our Cloud Management Portal, we have what we calla Resource Dashboard.

This provided me insightinto my Cloud estate.

I can see, let me do a quickoverview in here, and I can see whatresources I can manage.

I can manage VMs,I can manage Resource Groups, Storage Accounts, And images.

And we're continually addingsupport for more users.

I can manage my users, my policiesand also other features like an audio trail that I can see who,who's doing what.

A couple things Iwanna point out here.

I can see more thanjust one provider.

So this is a crops platform portal.

I can see, by different providers.

If I was also managing privatecalls, I could do that as well.

I can see summary information,additional detailed information from my resources, I can alsosearch the filter and look for the different resourcesthat I can manage.

If I select a particularresource I can see details and I can also perform actions foreach of those resources.

So we'll go into more detail on eachof these as we go through the demo.

All right, so one of the things thatI wanna be able to do in my portal is organize things in a waythat makes sense to me.

On my cloud called managementplatforms allows me to do that.

So, say I'm accountable for myhuman resources, One of my client's human resources, so I'm accountablefor management infrastructure.

I can lay a benefits apps andpayroll apps and I can organize things ina way that makes sense.

So they have health insurance, longterm disability applications, so that I can create this structurein a client project and sub-project way thatmakes sense to me.

Right, so when I do that then allmy cloud resources are gonna be allocated against thoseclients' projects.

And so projects andthey're gonna be discovered.

They're gonna appear tomy resource dashboard.

All right?All right, so I also talked about users.

So I've got a whole corralof cloud horsemen that are, it's not just memanaging my resource, I have people that I also needto control what they can do.

NACP are cloud management portal.

I could control who can do what foreach of these resources.

So here I have a dashboard.

And here I have all the usersthat have access to my projects.

I can see what role they have, whatthey are looked like admin role.

But they have a project role,sub-project roles.

And I can click on their names andsee what they can do.

So our cloud management platformhas rule based x control system where I can control what my userscan do for each of my clients, projects and sub-projects.

>> So one of the things that isvery important here and sorry for interrupting you,is in the issue of freedom and control, policy andgovernance, right? The question is becomeswho can do what? So our clients and our projects,we need to know which organization, which people, what role do they haveand what can they do in this cloud.

And by allowing that, right,we're able to get agility to them.

Because they get to see anddo exactly what they need, but without getting inthe way of other people.

>> It's correct.

Right, so I've got my projectsfigured out, got my sub-projects.

I've figured out who can do what and I've gone in my cloud managementplatform and set that all up.

I've got all these provider accountsthat I also need to manage.

In my cloud management portal, I'vealso can see what cloud provider accounts are associatedwith my project.

So here you can see, I have several,again they're cross provider.

And I can also associatewhat we call quotas.

So Rodrigo mentionedfreedom of control.

So what I can do for each of myproviders, I can set quotas.

And so on the server, somebody goes in andprovisions a particular resource.

If I've set a threshold, I wannabe notified when that threshold reaches 80% andalso when that threshold is reached.

So- >> Right, so this is an Accenture bestpractice that we recommend.

Rather than having approvals,we recommend putting quotas and budgets to specific people.

Quotas and budgets forspecific people and that way, it's kinda go wild, this is yourstate, right, this is your budget.

Do whatever you need to do.

You don't need any approvals, right? And we will tell you,the system will tell you where you're approachinga particular threshold.

So if you're 80% and it's the first week of the month,you've got some trouble, right? Now it's 80% andit's the third week of the month, you're okay kind of thing, right? >> Yep, so right,this can help me by avoiding surprises at the end ofthe month when I get my bill.

When I get my bill at the endof the month, it's too late.

So this is a way I canprevent that from happening.

All right, so what else can I do? So one of the key features of ourcloud management platform is you can also go out and tag my resources.

How do I provides you a wayto keep track [INAUDIBLE].

What I can do is create a datamodel, provide some metadata that's gonna help me manage see whatI've got out through my cuffs.

So, here in the portal, I can see the different kindsof what I can do with tags.

So I wanna point outa couple of things here.

So you'll see inthe resource tag column, I cannot tag differenttypes of resources.

So it's not just VMs, I can tagresource groups, storage accounts, images, all the resources that I canmanage are called management portal.

So let's say I wantedto create a tag and I want to organize my resourcesbased on environment.

And say I created a tagcall environment.

I can create that asa different type.

I can kick it as a tag orI can just sent through a string.

But I can also create it in a waythat there's specific values that must be provided.

And I can call it, say for example,I have an environment of demo.

And so when the users provisions a VM,they must select one of the tags.

So when I look at consistent datamodel, and this is a cross provider, it's not just one.

I can designate thingsas a billing tag.

So when I get my bill from ouraccount management portal, all my resources that I've tagged inthat way are gonna show up within that category, and I could see allof them align, all the resources and the costs that are associatedwith my environment.

I'll manage my tag.

What I could do just manage my tags.

I can designate tagsthat I care about.

So when I indicate so that a tag ismanaged whenever somebody goes out and provisions a resource throughmy cloud management platform, they're gonna be prompted toenter a value for that tag.

So this is how I can use my cloudmanagement platform to manage my procedures that I use tocontrol my environment.

And I can also specifywhat tag is mandatory.

So I can make a tag up here butI can't make people fill it out.

I can do that if byspecifying as mandatory.

So if Rodrigo is going outon vision to go VM and they had tag isdesignated as mandatory, he's required to putthe value into that tag.

So I have a consistent datamodel caused by writers.

>> Thanks.

So let me kind of summarize that,right? Tagging is critical strategy formanaging your cloud state.

And it's vastly underestimatedbecause you need to have a consistent tagging model.

You need to be ableto enforce your tags.

There's no point if somebodycalls the machine dev, somebody calls development,the other guy calls it D, etc, etc.

You won't get the reports, youwon't be able to take the action, so having really strong controlaround your tags is critical.

In fact, some of the best practicesthat I've seen with one client, basically, if your machineis not tagged properly, a reaper bot will just turnit off in five minutes.

Go ahead and provision and dowhatever you want, but we're gonna turn it off in five minutes if itdoesn't have the proper tags, right? And soit's a strategy I also put for my architecture groupafter I got the bill.

[LAUGH]>> And these tags are synced with providers, so if I define these tagsin ASP and I go look at my provider portal, I'm gonna see these tags andfind out there as well.

So, all right, so tags are great.

What else can I do with them? I can go out intoour portal here and I can search based on a tag value.

But what else can I do,how can I use those tags to further manage my costcontrol of my environments? So we have this whatwe call our policies.

So I can go to policies, andI'll talk a little about some of the different policiesthat we can use here.

So based on tag values,I can provide different policies.

So let's say I have my,I tag a VM as my demo environment.

I know I'm gonna use it forthe next two months.

So then after thatI'm gonna be done.

I don't wanna remember to haveto remember to go out and decommission that or power it off orshut it down so I don't pay for it.

So I can define what's calleda lease expiration policy.

So any VM that is tagged with PMdemo, I can set an expiration date that on that date, that VM is gonnabe automatically decommissioned and then no longer I'm gonnahave to pay for it.

I can create a automaticpower cycling policy.

So any VM that has a tag of PM demo, I know I'm not gonna be workingon weekends doing demos.

So I can set a policy that on Fridaynight, this is gonna go out and shut down that VM.

And on Monday mornings, this isgonna bring it back up for me.

I can also use those tags tocontrol my access to those VMs and who can do what.

So that's called an automaticrestriction policy.

So based on the take value, I can control who cando to what to what VMs.

We also have somethingcalled governance policies.

So I can go out and createa governance policy and say so I don't want anybody to go out andrequest a new provider accounts which is something I can't doto my call management platform.

So something I wanna control.

Whenever somebodysubmits request for a new provider account that I canrequire that it have requires my approval before it's provisioned andcompleted.

Okay? >> Sothose policies are really important.

The policies to return things off,the policies to assign rights.

And all of thishappens automatically based on the tagging system.

So this is why I kind of said,tags are a really critical and underestimated part of managingyour cloud state, right? It's really havinga good data model for what you're gonna be runningon the cloud, right? Yeah, once you have thatthen automation becomes a lot simpler, a lot easier.

>> So everything we're talking aboutare things that exist today and we're co-working on addingadditional policies, policies that I'llbe able to identify, or resources that are notbeing able to use anymore.

So say if I have disk that'snot attached to a VM, it gets all the policy that it'sautomatically be deleted in script.

But after for a long, pay for that,I can further control my costs.

Right, another thing on here iswhat we call an audit trail.

So I wanted to kind have someinsight into who is doing well, what they're doing, or somethinghappened why it might have occurred.

So we have this featurecalled an audit trail.

So I can go out here and I can see.

What events have occurred,you know, in what period of time, you know I can do by day, by months.

I can see who's doing itover here on the left, you know, the top tenevents that are happening.

So, it looks like a lotof people are viewing and discovering the infrastructurethat they have, over on the right, I can see who's doing it.

So, it's looks like somebody namedEd Contini's doing a lot of work so if we know that weknow that he does so.

Then down at the bottom I've gota list of all the different events that have occurred.

And I can go out and search andfind out what's happened and do some analysis of what mighthave caused the problem to occur.

All right, sothat's our audit drill, all right, let's talk about analytics.

So- >> My favorite part.

>> What's that?>> You know that's my favorite part.

>> This is the sexy part, I think.

So, Rodrigo mentionedsomething about QADA in one of our best practices andour methodology.

Which stands for questions,answers, decisions and action andanalytics is the A in the QADA.

It provides you the answerto your question so he make decisions and take action.

All of which you can do throughour cloud management platform.

So we talked aboutthe tag values and so let's look at a coupledifferent ways that our cloud management platform usesthose tags as part of our analytics.

So I'm gonna select the ResourceType, so we have analytics for different types of resources.

I'm going to choose the client orproject I care about, I'm going to select the tagthat I'm interested in.

So here I've got a summary ofmy consumption by tag values, so at the top here it saysat a client summary level.

So I talked about the client,project, sub-project hierarchy and this is at the client level, so forall my projects and sub-projects.

I have some insightinto my consumption, so if I was to hover over these bars Iwould see here on the left I've got nine application servers,two database servers, three others.

I've got a number ofservers that are untagged.

Over on the right I can viewthose by provider, by project and down at the bottom it's broken downfor me by individual project or sub projects that are in myclient project structure.

I also have something that'scalled cost drill down and Susan's cost drill down feature,I can see what my costs are and drill down into them andsee where I'm spending my money.

And then I can take actions in ACPand those actions might include employing those policies that wejust talked about to save me money.

So here I, see a graph of my currentconsumption, I can show forecast, see if I expect to be over or underbudget at the end of the month.

And down at the bottomI see some more information I candrill into further.

I can drill into, I can click onCloud management pilot project here, I can see this informationspecifically for that project.

In the center, I can get some information based onthe tags that I've defined.

So, if I was to click on a tag,the tag called primary use, I could see that for my VMs thatare tagged with resources and I spend $230 for those VMs.

Over in the right I cansee those by category, and down in the bottom I cansee the information by service.

So analytics provides the answersto the questions I have so I can make decisions and take actionto reduce my costs, control and manage my environment.

>> So hopefully, this is making itclear what we mean by a controlled plane for your publiccloud consumption, right? It's being able tounderstand what you have, being able to take actions,being able to understand, analyze, and then decide what'sthe best course of action.

Including things like,have you over provisioned of the VM, have you underprovisioned of the VM? Are you paying too much? Do you have stranded resources,disks that are sitting orphaned somewhere andyou're paying for them? Right, so I think we're gonna finishnow with the final piece, Manage.

>> Final piece is Manage, so we talked about how Ican manage my cost, control my resources, who does what.

And we're gonna talk about how Ican secure, protect and back up my resources in the Cloud andprimarily we're talking about VMs.

So for my dashboard, I could go inand I could select a particular VM and there's also a catalog where youcan go out and order these services.

I have different actions thatI can perform for this VM, I can start/stop them but I also have something calledCloud management services.

Rodrigo mentioned patching, Cloudpatching is one of our services that we have set up, that I cango out and subscribe my cloud resources to to make surethey're protected and secure.

>> Right, so if you will,David migrated some workloads.

Ron [LAUGH] put ina SharePoint site.

Now we have managed the statusthere, and we need to add patching, right? We need to add patch in,so very simple here.

Point to the machine, click it,and it will add all the agents, do all the right things,put it on the schedules.

And now you have patching forthat machine and you can do antivirus andof course backup.

>> The provisioning is automated,I get it quick, there's no long term approval process thatI need to have completed.

And also on our platformhere we have what we call a CMS dashboard where I cango out and make sure and see that all my VMsare subscribed to backup patching.

And monitoring our antivirusservices, so that's our cloud management platform, andhow me as a headless horseman.

>> Thank you, David.

>> [APPLAUSE]>> Okay, we can turn [INAUDIBLE].

>> Great, so what we just saw is.

>> Basically, we can reduce the costto compute by 40%, and lower the management cost by a significantamount, more than a third, right? Why through standardization, automation, propermetadata management.

Those are the secrets ofmanaging in the cloud, and the things we've learned managinga significantly large state in the cloud overthe last three years.

So with that,I'm going to turn to Michael.

>> All right, so hopefully wehad three horsemen come up, take you through the challenges andhow we solution these challenges.

It's not just the tooling,some of the little bit of sexy, tooling Ron's excitement there.

I mean we have, you know a lot ofcapability, a lot of visualization, integration, configuration andautomation behind the scenes.

But what you're doing isrendering best practice, you know codifying that andthen scaling it.

And from our perspective, dealingwith larger organizations I mean, this is absolutely critical inorder to be able to leverage cloud.

The new operating model versusthe old capital intensive model in this new day and age.

So we announced yesterday a wholebunch of new features and capabilities in our cloudmanagement platform.

These capabilities havea lot of fidelity for Azure.

They provide a lot of insight and capability around cost management,as well as kind of the organizationalconstruct capabilities.

To leverage Azure Resource Modelsand whatnot and resource groups.

And so that capability is what we'rebringing out to our clients world wide in order to help them kind ofleverage this journey to the Cloud.

And we're also wrapping variousservices around that so it's not a point solution, per se, it's not something yousubscribe to or download.

Its something that, its a managed service, most of thesethings we have migration factories, we have commitment to certificationprograms and training.

So that we have these cloud solutionarchitects and cloud admins and these, you know, teams of migrationspecialists that can work at scale.

Because our belief isthe faster you move and transition from a traditionalstate to the new state, the better the economics are foran organization.

It takes out a lot ofthe underlying cost and helps people become moreagile organizations and more flexible to the needsof the business.

So we have, I think, a coupleof minutes here for questions, I really thank you forstaying through this entire thing.

There's a lot here,there's a lot of IP, a lot of investment in assets andpeople.

But it is that combination ofpeople, process and technology that makes cloud real for a lot oforganizations, any questions? No, okay, well thank you,really appreciate your time today.

>> [APPLAUSE].

Source: Youtube