[MUSIC]
Welcome to Cloud at Microsoft, subject matter expert, or
SME roundtable.
Hi, I'm Sheri Bettine, and I'll be your host.
We are the IT Showcase team, and we love to talk to you, our customers,
about how Microsoft does IT.
Today, I'm here with many of our IT experts
that work on cloud-based solutions here at Microsoft.
This is your opportunity to ask direct questions of our SMEs, and
receive candid answers.
We'd also like to learn a little bit about you.
So we've posted a question in your Q&A window that we'd love to
hear the answer of.
What percentage of your portfolio have you taken to the cloud?
As you answer that question,
we can see your responses on the screen in front of us.
And with that, it'll help shape how we answer your questions.
I'd like now for our experts to introduce themselves.
So, we'll start with you, Rob.
>> Hi, I'm Rob Polly, and
I run a Cloud Security team here at Microsoft IT.
It's a team of architects, engineers, and program managers.
And we focus on helping Microsoft essentially digitally
transform in a safe and secure way.
>> Hello, and I'm Joe Mazzotta, I also work in Microsoft IT.
And I work in the Cloud Management Services team.
And the Cloud Management Services team is a central infrastructure,
operations, and help desk function.
Specifically, my role is a program manager on the Azure strategy and
consulting side.
So I work with internal and external teams, helping them adopt Azure.
>> Hi, I'm David Johnson.
I'm a program manager architect accountable to our Office 365 and
collaboration services portfolio.
The team that I'm in within Microsoft IT drives
how we collaborate as a company, everything from Exchange Skype for
Business, SharePoint, our portal ecosystem, Enterprise Search.
Basically the technologies that enable our employees
to work together.
>> Hi, my name is Arun Mannengal.
I'm a Senior Software Engineer at Microsoft.
My team is responsible for the strategy, design, and
development of hyper-scale services that power cloud scenarios,
which are crucial for functions in our organization.
>> We are also doing one more thing with this SME roundtable today.
We are collecting up your feedback for the product team.
So, in addition to questions for our SMEs, if you have some
feedback you would like for us to pass along to the product team,
please enter that into the window as well.
And we'll collect those up offline, and
give them to the teams that they go to.
We'll make every effort, in the next hour,
to get through all of your questions.
However, if we don't, we'll stay behind in the studio,
continue answering questions, and we'll post that extended footage,
with this video, at microsoft.com/itshowcase.
With that, we'll get started.
And I first want to ask our SMEs the same question we asked you guys.
What percentage of our portfolio have we taken to the cloud?
>> It's time for me?
So it's a very good question.
So when we talk of the cloud, there are multiple nuances to that aspect.
There's a SaaS aspect to it, there's an IaaS aspect to it, and
there's a PaaS aspect to it.
So when we touch upon each of them separately, we find various levels
of maturity, and adoption in each of these spaces.
For example, in the SaaS space, I'm sure David would also touch upon
additionally on that, we are seeing at least upwards of 95% for
services that we rely on managed services.
The Office 365 offerings.
When we talk of the PaaS capabilities that we have migrated
over, that is where your custom development comes in.
And this is where different teams within Microsoft are at different
levels of maturity.
Having said that, any greenfield operation definitely starts off with
an on-cloud offering only, because there is a strict mandate that
there's no on-prem development that can happen on a greenfield thing.
On legacy and brownfield applications also,
there is a gradual cut over that takes place.
So it is a cohesive mechanism which ensures that it's business
continuity, while at the same time, having focus on the cloud.
So that's how we are modeled within Microsoft,
in terms of adoption of the cloud.
>> Yeah, exactly.
With SharePoint, for example, as you say, we're over 95% in.
We had a huge portfolio on-premises with my sites,
team sites, group collaboration environments, portals.
Almost all of that's now on the cloud, with a couple of exceptions.
Same thing for Exchange, obviously other on-cloud release services like
Yammer clearly only in cloud, Microsoft Teams is in cloud,
Skype for Business is now partially in cloud.
So we're moving more and more portfolio.
But things like SharePoint for example, which was a huge footprint
for us, that are almost completely there now.
>> And then I think from an OS instance perspective I know that we
have pretty aggressive goals in moving up to,
I think our goal is 90% of having our on-premise
OS environments moved into Azure by the end of June, I believe.
So that deadline's fast approaching, huh?
It's coming up soon. >> Yeah, and we're currently at 50%
virtualized in, well, 50% in Azure from a OS instance perspective.
And as Rob mentioned, yes, our goal in Microsoft IT specifically,
is 90% by the end of the fiscal year, June.
>> Wow.
With OneDrive and Office 365,
can you explain the role of SharePoint and Microsoft Strategy?
>> So effectively, think of OneDrive for Business as the place for
the individual files.
The files that the person may have formerly stored in their My Site or
their machine or other replication devices
where they're storing them locally for them.
They may be sharing them, but it's really first and foremost for
the individual.
We think of SharePoint as the team collaboration environment,
the portal environment.
Where you've got a new group, new project, you spin up a new group or
SharePoint site.
That SharePoint site is where those team collaboration files live.
And that's fundamentally the distinction.
And the reason for that, by the way,
is fundamentally from a permissions model, we want the simplicity of
a SharePoint permissions to basically handle the team.
That I want to make sure that when I'm sharing with my team,
I've got a space I can do that safely.
When I'm sharing it with just my manager, great,
go do that on my OneDrive for Business site.
>> Okay.
I'm wondering if there's an aspect of on-premise versus cloud
because Office 365 has SharePoint as part of its products, yeah.
>> Yeah, I'll add to that, so
SharePoint Online is effectively part of Office 365, so OneDrive for
Business and SharePoint are just different site types on Office 365.
So, if you go into Office 365 and you see within the suite navigation,
you've got a ribbon which includes things like your OneDrive for
Business, includes your SharePoint portfolio and
your SharePoint home to show you all the things the employee cares about.
And yes, there's still, of course, have SharePoint on-premises for
things that you want to keep on-prem.
In fact, in our case,
we have a SharePoint hybrid environment in place.
So that SharePoint on-prem content shows up in
the Office 365 search index.
So when someone for example finds content in Office 365,
it includes all the content that maybe someone still has on
an on-prem SharePoint site.
>> Okay, very good.
What is the best approach to migrating to Azure?
>> So I talked to many internal and
external teams in terms of migration.
There's lift and shift, there's refactoring,
rehost, and so there is various ways you can get to Azure.
In terms of the best or the optimal strategy,
I think if you're taking a look at enabling agility and
cost-optimization, from a Microsoft IT perspective and
across the company, we're taking a look at the SaaS and PaaS offerings.
Because in a SaaS and PaaS environment, you're able to leverage
more of Azure's native capabilities in terms of scaling,
in terms of management or manageability, in terms of security.
So that you're not having teams on-prem having to manage
all of those levers or drivers in the Cloud.
And so, in Microsoft IT there
are many teams actively marching towards refactoring workloads.
The other aspect of refactoring is also starting to decouple
traditional tightly coupled applications on-prem, so
that you can scale components of the application independently.
>> So you can move some to the Cloud, have some on-prem?
>> That's certainly one approach.
So that's certainly on approach, we're actively driving,
in Microsoft IT, 90% of our environment to be in Azure,
and so ideally, then We can start decomming more of our traditional
on-prem infrastructure and truly leverage Azure's strengths.
>> But that helps you also just take it piecemeal and
not have to do big bang, right?
If you can migrate.
>> Correct. [CROSSTALK]
>> This is an important
perspective here.
Microsoft is uniquely positioned in that that it has a hybrid story
which can ensure your business continuity.
It's not like, hey, tomorrow I have to be all in cloud.
That's not a practical reality, nor a pragmatic approach to follow.
So, the assets that we have and the product offering that spans and
allows our customers to go from being completely on-prem to
the cloud in a very predictable and pragmatic fashion,
which adds value to the offerings that we have today.
And that's helped us lot within our systems as well,
when we have done especially critical invoicing systems and
legal and compliance systems.
Because not all cloud data centers may be compliant on day one to all
the regulations in that country or the province that it is hosted in.
So that's a critical aspect that helps.
>> Yeah. I think you guys hit on a couple,
I don't know if there is a right way, right?
Like a single way.
But I think you have some characteristics that you brought up,
at least in what I've seen and experienced.
I think having a very pragmatic approach
that's very specific to your organization's business and
needs, right, if they understanding kind of what is that
from a tolerance perspective, what risk do you feel comfortable taking?
Where do you have skills?
Where do you have processes that you're able to kind of experiment?
So I think being pragmatic,
understanding kind of what you're trying to get out of it.
I think starting small with something, that's greenfield. Right?
Something brand new.
That there's really not a lot of business risk in place.
I think that's another- >> Easy way.
>> Great way to kinda get in there and take advantage of some of that.
And I think through that you'll be able to kind of test
some of those connectivity pieces that you have,
whether it's service management, performance, availability,
cost, your security pieces, how your developers interact.
I mean, there's all these kind of things that you can begin to
experiment with and understanding.
Because, I think the one thing that I do know that the technology,
it's always gonna be there, and
it's changing faster than we can even keep up with.
But the part that's really the most impacted, I think, are people and
their skills and their willingness, from a cultural perspective and
a personal perspective, to actually change,
cuz sometimes the way you do things in the cloud is different.
And then also the business processes that you have that support that
kind of migration, those things may change significantly as you move. Right.
>> And the advantage of starting
small, like Rob mentioned, is that you are able to reuse your learnings
back when you date back to a bigger project that is of
important significance to your organization.
You don't have to peel the onion when you're doing a big project.
>> [LAUGH] >> You have already encountered
the problems, you are aware how to resolve them, and that also builds
in a motivational culture in the team that they have had success,
they have tasted success, which they can replicate into bigger projects.
>> And their skills are growing, right?
So as you're doing the small, medium, big,
you're growing the capabilities in your organization.
>> Absolutely.
>> Yep.
>> If I could add,
so from a Microsoft IT perspective, a couple workloads,
types of workloads that really make sense to look at, identify first,
and then take a look at learning in the cloud are applications that
have burst capability, maybe once or twice a year.
And so some of the applications that,
line of business applications that Microsoft IT started with migrating
are tools like our HR review system, where it's used twice a year.
Previously, we would have massive infrastructure that was permanently
on-prem.
By migrating it to the cloud, we're able to burst when we need it, and
then scale back down.
And so our cost savings are,
we're realizing ROI on our application portfolio.
>> So we're not paying for the whole box when we're only using it 2,
3% [CROSSTALK] >> Throughout the year.
Correct.
>> Right.
>> Yeah, absolutely.
>> Another use case on similar lines-
>> Yeah.
>> Is the dev test environment.
I mean- >> Yep.
>> Most of the dev test and environments are just lying vacant
when you're not really pounding that with your workload so to speak.
So, these environments could be spun off, and this is where automation
comes in and helps you gain that financial benefit.
And it's not only financial benefit, it is also a discipline that gets
to the team about use your resources when you need them rather
having that hogging mentality and keeping it for you, forever.
Which also introduces another forecasting issue,
which we'll touch on as we go.
>> Okay.
>> Yeah, I think that last piece is something that we would certainly
file under a big lesson learned.
I think that's one of our biggest lessons learned about how to build
that snooze or turn it off kind of mentality, right?
Back into the- >> The culture.
>> The culture, yeah.
>> Yeah, turning the lights off when you leave the building.
It's something you have to encourage and
build in as a muscle in your organization, right?
>> Well, but if you think about the journey from a data center
perspective where a lot of companies had big data center build-outs,
bought a lot of physical hardware, and
we found that it was underutilized, then came
virtualization technologies that was intended to kinda up level that.
But it's funny when you kind of move to the cloud environment,
it's almost back to starting- >> Yeah.
>> With what you had in the data center because buying those skews at
whatever you think you might have is so easy to do but it's not
probably what you need if you really looked at your overall utilization.
>> So you can overbuy in the cloud environment just like you can-
>> You can.
>> Overbuy in on-prem.
>> Right. I think that's another lesson that
we've learned too, but I think back to the way that
you don't have on-prem is actually you have a really easy.
I don't know, from our perspective, we think that we give our developers
the obligation- >> Yes.
>> Actually of taking those things- >> Absolutely.
>> Down and holding them accountable for that cause.
>> And with the cloud that the management tools that you have,
you can also incentivize that culture,
which would not have impossible traditionally.
>> Right. >> Because there are management
tools which help you to do it.
There's also this is a philosophy of,
hey, developers might have good intent but they are not
provided with the necessary tools to help accomplish their objective.
They wouldn't be able to do so.
But what cloud has enabled is that there is a mechanism in place for
you to set a process in place, and
also monitor whether that process is being followed.
So, the dual benefits accrue.
>> Okay.
What are some of the biggest challenges that your organization
has had to overcome as you moved from hosting IT services in your own
data centers to those run by cloud service providers.
Rob?
>> The security one, that's a big one.
I think for us, I kinda go back to this concept
of traditionally in security we've kinda played gatekeeper on-premise.
You had to buy your hardware from somebody.
You had to go through some security checks,
whether is to making sure that the operating system was configured
correctly, whether the network stuff was configured right,
whether the apps that were gonna run on it were done properly.
But the power of the cloud, is certainly
democratizing the access to compute, network and storage, right?
And so, essentially those gates were removed.
And I think a lot of the folks, at least in my discipline,
it's still a big leap as we try to apply the same principles in
the same way that we did on-premise with those in the cloud.
The philosophy that we've tried to take to kind of enable
innovation to occur in Microsoft, which is from an IT perspective,
our goal is to simplify and accelerate, right.
And if we're gonna accelerate the innovation that occurs here
We've gotta get out of the way of enabling that to happen, but
we also have to have, we have to be the backstop, or
the guard rails I like to call it, guard rails instead of gates.
That we're kinda helping people from kinda making some of the mistakes
and getting everyone in a little bit of trouble.
>> I could add to that, that's great point Rob.
So, our organizations work closely together with security,
and our teams work with Rob Polly's team pretty closely around
a couple of things that customers could do today.
One of them is, and as Rob mentioned, thinking about
alleviating gatekeeper and thinking about guard rails.
Leveraging ARM templates, so
Azure Resource Manager templates in Azure to pre-provision environments.
Our teams worked with security very closely to publish
ARM templates out to the company so that Microsoft internal teams can
consume these ARM templates, build-out environments within
the ARM templates we've embedded our IT standards and security controls.
There is also another construct called dev ops toolkit,
which developers can then import directly into Visual Studio.
As they're building their code, the dev ops toolkit enables
the health and welfare of the code itself so that IT nor
security is getting in the way, but we've enabled an experience that
makes the business more agile, allows them to go to the cloud,
from the developer or the business perspective, they don't see gates.
What they see are IT and security helping them get to the cloud in
a way that's standard for Microsoft and in a way that's secure.
You make security easy for the developer, make it the default path.
>> Absolutely, that's our.
>> If it compiles, it complies, that's our goal.
>> [LAUGH] All right, good deal.
What are some proven practices in outsourcing
to Microsoft SaaS offerings, David?
>> So I'd say, first of all, understand, I guess,
understand the services available to you,
understand the capabilities you can take advantage of.
For example, if I'm building a storage layer for
a bunch of files, for example, for my application,
when can I use SaaS services to back that up?
For example, maybe storing files, for example, in Office 365,
SharePoint Online or OneDrive, and consuming them from my PaaS
application or other applications it's using against that.
Because, of course, in 365 the SaaS services are there as
a fully packaged product, or it can go into SharePoint Online,
OneDrive for Business or Exchange, and use the UI completely.
Or it can use the APIs from these services and embed them in my own
experiences or applications, and therefore using the data.
For example, you might not think about why would I ever wanna build
a calendaring tool for an application when Exchange, clearly,
already has a calendaring system built into it?
I can simply call that.
I can simply call Mail, I can call Yammer for social connectivity.
I can call SharePoint for file management, for example.
So understanding what the services are available to you, and then
figuring out how they best apply to help you build your applications.
And that's a question in addition to just simply
getting people to onboard from core services, themselves.
Which there's a whole other conversation about onboarding.
>> And to add to it, the real paradigm shift that is
happening in the SaaS space is the application of analytics,
which we couldn't do traditionally if you had your data dispersed on
premises across multiple assets, right.
The cloud, what you're essentially having is
one view of all the data across your organization.
And you could apply analytics and derive meaningful information and
predict the trends around what is happening.
It could be sentiment, it could be a lot of things that are critical for
your business needs.
>> Yeah. >> In that aspect.
>> So hold on, so, David, it sounded like, sorry about that.
It sounded like, it's kind of amazing what you just said,
it kind of just struck a chord in my brain.
It sounds like Office 365, very similar to what Azure's done,
kind providing these, we call them paths.
But, essentially, things like web apps, and Azure Machine Learning.
I mean, those are just other types of SaaS that you're leveraging.
So, similarly, Office 365 also provides
these kind of building blocks by which you can kind of go and
begin to assemble these kind of apps.
>> Yeah.
>> Yeah, that's really cool stuff.
>> Arun, tell us a bit more about application insights we have today
that helps application developers.
>> Absolutely, so one of the major things that happens when you deploy
your applications in the cloud is the importance of telemetry and
awareness of how your application is performing.
Because unlike the on-prem world,
you do not have the flexibility of getting into the cloud as you wish.
Though you could do it on a need basis.
But the frequency at which you would be aware of what the application is
doing, you have to rely upon log analytics and telemetry.
And this is where Application Insights, as an offering from Azure,
comes in very handy.
Because not only does it help you understand what are the code
parts that has been taken as part of your application.
It also helps understand how the customers are using it because,
essentially, the scenarios get lit up.
And that in turn helps us focus the developers' attention toward
features that are heavily used, driven by data.
As opposed to an assumption that we believe that users are using it
in so and so fashion.
So App Insights in that fashion.
This also alludes back to the DevOps kit that was mentioned earlier.
So it embeds into your application and you're able to derive analytical
information about how your application is used,
where an application is crashing, if at all,
if there are some code patches are crashing, even when you deploy.
So Application Insights helps us in that fashion tremendously.
Also, it aids the DevOps culture to be interpretive more.
You don't have to rely on an ops engineer or
a service engineer to go fetch some logs for you.
You have it delivered on your portal in a clear fashion.
>> And the nice thing about it is we use App Insights,
not just on Azure applications, but we even embed App Insights, for
example, just even sometimes on SharePoint applications.
Cuz I want to understand the results in performance, or where potentially
there are some bottlenecks in the page load, for example.
And App Insights helps provide that.
>> Absolutely, and the beauty is that the SDK is prepackaged for you.
You are only providing additional hooks,
specific to your business scenarios.
And the telemetry piece is taken care of automatically, in that case.
>> So we get the intelligence to be proactive about the support of our
applications, too.
>> Yeah. >> If your application's running
sluggishly, you can see that, right?
>> Absolutely, absolutely.
And then time to reaction is proactive rather than reactive,
which is very important.
>> Right, I think one of the key things from a security perspective,
I mean I know we've spoken about the performance with the availability of
the application.
But the power of having something in the hands of the developer
that tells them whether their app is behaving in a manner that
is different than what they expect.
>> Yeah.
>> That's a really powerful thing from a security
kind of awareness perspective, right.
>> Absolutely. >> And
I think as we move away from kind of this traditional,
we look at security kind of as prevent, detect and respond.
And as we move to more of the guard rail approach,
we want to embed great security practices, help people along.
We're not gonna have the gate, so prevention kinda goes to the side.
But being able to detect, and then how quickly can we respond to that?
>> Absolutely.
>> Such as something like you just talked about,
whether the page is not loading fast enough, etc.
I mean there's just a ton of great telemetry that's out there,
whether it's audit logs, the App Insight stuff.
I mean, all of that really enriches, I think,
what the developer experience is.
It also, on the other hand, puts a ton of accountability back, and
responsibility back,
to the developer to know what to do with that information.
But, overall, I think from a security perspective,
and a performance perspective, it certainly helps everything run much
smoother from an IT perspective.
>> And the beauty is, in context of App Insights as a product, you
really don't have to be on the cloud to gain the full advantage of it.
Yes, if you're in the cloud you get more features.
You also have a pathway so
that when you're prepping up to move to the cloud, you could start by
inserting these building blocks into your applications.
It would still emit out the telemetry,
which is stored in the cloud,
and you get the information and analytics about your application.
>> So this builds a little bit on that, or parts of that.
This is a question on cloud service management, especially for SaaS.
With 95% SaaS,
what is the best way to proactively monitor that services are working?
Can you rely solely on service communications or do you need more?
>> Well, I'd say that it's a mix.
And I think it depends also on the service, in particular,
you're talking about.
I think the communications back from Microsoft, for example,
would be alerting you.
You can use tools from Microsoft that let you see from a dashboard
of how your services are performing and where you stand around that.
We do, as Microsoft IT,
we do a couple other things in addition to that.
So we have tooling, for example, with App Insights, lets us watch
when are things performing, potentially, a little sub par.
So we even know if a core service will tell us, great,
the 365 is up and available.
And all of these core offerings are here and
this is where the server stands.
And I can go into the dashboard or
I can get proactive reports about that.
But I also have our own monitoring in place with the App Insight stuff
we were talking about earlier.
And so from the service management perspective,
Microsoft obviously communicates to the customer, the enterprise,
themselves, around basically what's going on, our contacts.
Obviously, then, it's up to the IT shop,
to how do you distribute that within your own organization?
We do things like our service management,
if there is a service outage, for example, or
something's going on, we use, for example, services like Yammer,
where we communicate broadly to our employees.
Here's what's going on our services,
here's something new on our services.
We do things like have what we consider a users groups.
For example, we have an Office 365 and
SharePoint users group on Yammer, whereby we can have a customer
conversation around things that are working well for them.
They can ask us questions.
We can kind of do help desk almost,
plus here's the how to solve a business problem with tooling.
And here's things you should be aware of because here's how
the service is changing.
Or things you should be aware of because
maybe there's a negative impact for some reason.
>> If I can add to that.
So one of the other things that the organization that I work in,
Cloud Management Services, provides,
in terms of best practices internally, is also think about,
when you're in SaaS or PaaS, think about your retry logic.
Think about the health of the APIs, themselves,
as you're interacting with SaaS and PaaS.
And so understanding the health performance.
Often a lot of the APIs have,
they do have telemetry that's provided back.
So you can query the API and
understand the overall health of the API.
So our organization ingests that information,
we ingest that information in a couple of different sources.
One of them is Operations Management Suite.
And then we provide that service back out to the rest of the IT
organizations.
>> And live site, from a service manageability perspective, live
site is a huge focus on Microsoft, especially with cloud services.
Because we understand that our success is
dependent on the usage of our services.
And for our usage of our services to happen more frequently,
it has to be performing,
available and when the customer needs it, it should be there.
And anytime beyond a certain threshold if that
characteristics are not being met, the developer responsible for
that piece of code is alerted real-time and
that is made possible through monitoring tools and what not.
So the thresholds could be for latencies, it could be for
the liabilities like if you start seeing a bunch of arrows in
a cluster and if it is beyond that threshold the developer is alerted.
This has two impacts.
A, the time to mitigate is reduced drastically, because the person who
is aware of that functionality intimately is fixing that.
And the other aspect is that developers incentivize to enforce
those retry logics and the secure uncoding practices, and the best
practices in quoting that he should have done that in the first place.
And this is really moving the needle in terms of quality
of the application overall in an iterative fashion, and
that is the beauty at which we operate.
So service management as a whole,
the bar just keeps rising as we go out.
>> Now this, the, I mean, being a cloud service provider,
you're providing a utility.
>> Absolutely. >> Right,
you've gotta be really good.
The water's gotta turn on when the faucet is cranked, right?
>> That's right.
What are the plans for a hybrid solution?
A solution whereby you can mix on-premises with
cloud infrastructure and services?
So we have hybrid.
We're a hybrid.
Who wants to comment on that?
>> I could talk about it from a security perspective.
I think we certainly have been running,
in the Office 365 space a lot longer in kind of the hybrid arena than
we probably have in the kind of traditional line of business,
app developments space that we do see in Azure.
And I know that David probably has a lot of insight about how those
pieces kinda matured over time.
My perspective is definitely more focused on the Azure site,
and how we moved in.
It's definitely I'd say that the connectivity options from a security
perspective that we've been able to provide,
it's taken us a while to get to the ones that
we feel comfortable with from a security perspective.
A little bit's our mindset and some of it's the capabilities and
some of it's how we wanna get it implemented.
But certainly there's been a ton of challenges and
a lot of lessons learned about what's the right way to do it.
I think it's back to the thought of the right way to do it.
I mean if you just start completely with Greenfield types of things, and
I know that may not be possible because of data anchors that may
exist.
But if it's at all possible to start with the Greenfield, and just work
on that piece to get that right, I think you are so much better off
than trying to keep that all in on your on-premise environment.
That's just been from a security perspective one of our traditional
struggles.
>> Yeah, as I think about hybrid in Microsoft IT at least.
So we do have hybrid where we've got
portions of the application, line of business application on-prem.
Particularly like data anchors.
Data anchors because there's multiple dependencies with other
data sources on-prem that's simply are not feasible to move right away.
Or we want to move it in the right way to Azure and so
we have to decouple and componentize the application and
move the parts that are easier and provide an ROI immediately.
As you mentioned, connectivity plays a part.
Now Microsoft itself has a product called Azure Stack which provides
the ability to run your on-prem environment similar to Azure and
then leverage capabilities of both.
Leverage Azure type services both on-prem and in Azure so that,
the goal is that you can move, you can shift between on-prem and
into public Azure seamlessly.
The other aspect of hybrid is thinking
about how to componentize, how to decouple your applications,
and then how to get the benefits of both worlds, if you will.
If you've got already sunk investment, on-prem that you wanna
continue to kind of ride until end of life and then there
are portions of your applications that you wanna move to Azure.
We've got a couple of teams in Microsoft IT that have taken very
novel approach in terms of taking a look of their front-end websites and
moving those to Azure websites or Azure web apps and
then doing that very seamlessly.
And then either over service bus or over ExpressRoute being able to
maintain the data connection back to on-prem.
>> Yeah.
>> So, there's a many different ways you can achieve hybrid.
At the end of the day, Microsoft is in a great position where
we provide multiple options to customers.
Our products strive towards seamless integration.
>> And at various levels of maturity,
depending on your organization so that is a huge plus for
someone looking into moving to the cloud.
>> Yeah, absolutely.
>> Yeah, and as Rob noted for 365 hybrid we have things like hybrid
services for SharePoint, for search for example or hybrid taxonomy.
So I can have a single enterprise taxonomy,
a term sets that span both on-premise and the cloud.
So no matter where I'm managing my content, for
example, I've got content consistency.
No matter where my content is stored I've got the ability to have it with
single search index.
So the core 2016 for example provided these, for
the on-prem product provided these additional hybrid
capabilities to work against Office365.
And so between that and exchange hybrid we've got a pretty good,
we've had for years this ability to be left kind of in both worlds,
on-prem and cloud.
>> And it's easier today.
>> Yeah. >> Because of that.
What is the roadmap for Azure Site Recovery?
Like Azure to Azure replication etc.
I'm gonna caveat initially by saying we speak to what we know in IT.
And we don't always know roadmap questions, but
is this one that we have any insight on?
>> We could have this definitely post it back on the product team.
>> Yeah, we'll take that back to the product team, okay.
>> It's an important feature,
we are not discounting that because how do we recover from disaster.
There are multiple facets to it and we just don't wanna speak for
the product team because they are still In
various stages of development.
>> We do use Azure Site Recovery in Microsoft IT,
we use it in a couple ways.
One is for site recovery, the other aspect is for simple migration.
And so we can take an entire environment and
migrate an entire environment.
But as Arun was saying, it roadmap and product related questions, future-
>> Go back to the product team.
>> Future questions on product capabilities.
Yeah, let's have the product group answer.
>> Is the backend storage for OneDrive and OneDrive For
Business files the same?
As I understand that One Drive for Business is for sinking SharePoint
document libraries, and I'm assuming would be stored in SQL in the cloud.
Does the one terabyte of One Drive
personal storage I get with Office 365 get stored in SharePoint also?
>> So let's think of it from this way.
You're as a tenancy you get a certain amount of total storage
available for user allocation of which your defaults are actually now
25 terabytes for One Drive for Business.
25 terabytes for SharePoints site collection, but
up to your maximum purchased storage for the tenancy overall.
And then you can decide if you want to allocate that and
maybe limit how many, what your users can use.
First of all, One Drive for
Business is a different storage mechanism than One Drive Consumer,
but you don't need to think about from that perspective.
Think about it from a policy and manage and storage perspective,
all the 365 protections that SharePoint has in place,
that SharePoint One Drive have in place are all there for
you on the 365 Service Stack.
So everything from the encryption at rest,
the identity management infrastructure in place,
for sharing models and protections and policy management, they apply to
both the Share Point online sites as well as One Drive for Business.
We think of the two as the two sides of the same coin of I've got
One Drive for Business where I've got my personal storage and
I've got my team SharePoint site for my group storage.
But both are fundamentally using the same storage models underneath and
both are using against my tenancy quota.
I can go up to 25 terabytes of my OneDrive and I can go up to 25
terabytes on my each individual site collection on SharePoint.
>> One important point that this question raises is about
how should we perceive manage services going forward.
Traditionally, we had to know what is happening in the guts.
What manage services allow is the other internal boiler plate is
shielded so today it could be SQL Server tomorrow it could
be Azure Table or Azure Blobs or what not.
It's just services and the policies like how David mentioned.
The policy is a series that you have been Contractually obligated,
that you should be concerned about, rather than the internal guts,
which keep changing frequently, by the way.
We experiment, we do AB tests, we find out what is in the best
as various technologies mature, and so on and so forth.
>> Yeah, cuz SharePoint, OneDrive, would store content in SQL, yes.
But they're moving a lot of that content, especially blobs,
into Azure storage, for example.
Always encrypted and encrypted at rest, but still stored there,
as opposed to just stored directly in SQL.
But whether it's stored in SQL or Azure, to your point,
it doesn't matter, because at the end of day, all the same controls
are in place for you, all the same protections are in place.
>> Yeah, I know it's hard sometimes as a digital technologist,
you have to be curious, right, to stay abreast of the latest thing,
and to keep at it.
Sometimes you do wanna know what's going on underneath the covers, and
that's something that you should keep asking those types of
questions.
One of the things I actually talked to a customer,
a very large customer, the other day.
And one of the things that they talked about was the need to know
that kind of stuff was essentially pivoted on whether
that would provide themselves with some kind of advantage in
the marketplace with their specific industry.
>> Absolutely.
>> And if it was commodity, like sharing files or storing stuff, or
getting email, unless they were in the business of doing that, right?
It's not gonna be a differentiator for you in the marketplace.
And that way you can kind of focus that curiosity back to the stuff
that really makes a difference to help your organization transform and
innovate, right?
>> Yep, scenarios have to be the front and centerpiece, and
then from there, your curiosity can help you derive the differentiator.
>> Yeah, great question.
[LAUGH] >> This may be one we can't answer,
but I'll ask it.
What is the roadmap for Microsoft teams?
Will this replace SharePoint Groups and/or Planner down the road?
>> So, I can answer that one,
because that's actually a common, I think customer misperception.
Cuz here's the thing with Microsoft Teams.
At the end of the day,
Microsoft Teams sits on top of the Office 365 Groups fabric.
Let's think of Groups as an infrastructure.
It sits on top of the Azure Active Directory for
the directory construct for the group.
It sits on top of a SharePoint site.
Every group effectively in Office 365 gets a SharePoint site, gets,
you can get the mailbox and calendar and other artifacts for that,
you can get Planner and everything else.
And definitely gets an Azure Active Directory security group
effectively for that group.
Teams sits on top of that.
So, think of teams as the kind of persistent chat based teamwork
collaboration paradigm on top of Office 365.
You can choose to use that and collaborate with Microsoft teams
if you want kind of that rapid collaboration where you're
kinda almost in the hallway, I want a quick conversation back and forth.
I've got my channels, I can IM about quick things, and
get information and see what's going on.
But that doesn't preclude the fact that underneath Microsoft teams
is that group.
That group contains a SharePoint site.
So, when I save files in Microsoft Teams,
that's going into the SharePoint site for that group.
If I'm using Outlook, and I'm having a conversation in Outlook, and
I click Files, those files are going to that same SharePoint site.
So it's not that Teams replaces Groups,
it's not that Teams replaces Outlook,
it's a different collaboration model on top of the same infrastructure.
And so when they don't preclude each other,
it's about how your team likes to collaborate.
By the end of the day, you're protected, because
your files are managed consistently irrespective of how you work,
what collaboration workload you do.
>> Okay, and when you set up a team, you get a group.
>> When I set up a new Microsoft team,
I get an Office 365 group underneath the covers.
I get the SharePoint site, I get the Outlook components to that including
a mailbox, including a calendar.
I get Microsoft Planner, so
I get all the groups constructs with the Microsoft teams on top for
the persistent chat and the nice new interface for
that kind of team quick collaboration that teams provides.
>> Great.
Azure Stack appears to be a solution.
Can you tell us more about it?
We've mentioned it a few times.
>> So, Azure Stack,
there's plenty of information up on the Microsoft website.
But, correct, yeah, so, Azure Stack provides the ability.
It's currently in preview, or technical preview.
But Azure Stack provides the ability to transform portions of
your on-prem infrastructure to appear and act just like Azure.
And so it creates resource providers for compute and
storage, whereby you can then create VMs that are similar to or
identical to what you would get in Azure.
You can then land your workloads on-prem in facsimile, like Azure.
And then you the goal would be to have the ability to seamlessly
migrate those resources as you need, or burst to Azure as you need it.
>> So, you can develop on-prem like you're developing for
Azure with this technology, and
then when you want to move to the cloud, it's easy.
>> Correct.
>> Or you could also alternatively think about it for, say, a large
bank, does not want to really put it on public cloud infrastructure,
they want a private cloud.
It's your private cloud functioning just like how Azure would in your
datacenter on your hardware.
>> With the auto scale?
>> Absolutely, all of those baked in.
>> So, the only part, you mentioned compute and storage.
So you're just utilizing your own network.
>> Correct, yeah, and the goal, my understanding is the goal,
I'd have to take a look at the roadmap.
My understanding of the goal is that Azure Stack would also provide PaaS
type services as well in the future as well, yeah.
So, things like Azure Web Apps.
>> Makes sense.
>> So, how does this panel feel Microsoft Azure stacks up
against Amazon Web Services?
>> Frankly, we don't want to be talking for competition at this
point in time for or against in that context, right?
We are here to share our experiences, and
we'll be happy to share our experiences.
>> A very diplomatic answer.
>> Specific to Azure.
>> Thank you.
>> I will add to that, that Office 365,
if you're building a solution on top of it, it doesn't really care where
you host that solution it's consuming from 365.
I can build provider hosted applications running in AWS,
or running in Azure that connect to and enhance Office 365.
>> So our APIs are well-documented,
they just rely on the transport mechanism, I mean, your protocol.
It doesn't care about who owns the hardware where your compute runs.
Whether it's Azure or AWS or Google Cloud or what have you.
>> There's plenty of professional reports.
Gartner I know does an in-depth report every quarter, I believe,
comparing Azure and AWS.
There's plenty of reports out there that also compare
the individual services within AWS and Azure.
You get very much a lot of the comparable services.
There's compute, and storage, data stores, SQL or
NoSQL in AWS, just like you'd get in Azure.
You can get Linux, Ubuntu,
various flavors of OS in Azure and AWS.
And so yeah, I mean.
>> So now that we've come out of the diplomacy shell,
let me talk to this on the Gartner report, specifically.
>> The gloves are off.
>> Exactly. So, specifically,
Gartner has four reports for the quadrant report that they have for
IaaS, PaaS, SaaS, and the data storage layers, right?
Microsoft Azure figures on the top quadrant in all the four of them.
AWS does not figure in the top quadrant for all four of them.
So, that's one way to, if you wish- >> Okay.
>> To look at it, yeah.
>> Let me hurry through the next few.
Let me know whether Azure has any road map to enable
VNet peering between two regions.
Is that something we know anything about?
>> I think we'll have to leave that to the Azure networking team to
disclose.
>> We'll pass that on.
>> Microsoft IT has provided a very similar requirement.
It's a requirement that we need internally, and so,
it's a requirement that the product group has.
When they deliver that should be addressed by the product group.
>> And we'll pass that along that you desire that capability as well.
Can you speak about mixing and
matching multiple Azure Office 365 offerings?
For example, I wanna use Power BI to report on a new Intune
bring-your-own-device programs adoption, and then use machine
learning cognitive services to see patterns in the mobile device usage.
Microsoft Flow Azure Logic Apps has a good way to integrate
lower level tasks, but I don't see use cases like I described.
Do you have anything to offer up there?
>> So that's a great question.
That is classic stating that hey, without developer help,
how am I able to stitch together scenarios?
But with Power BI, you are able to supply it with your data source,
and you could have lot of analytics around it.
Now, hooking Power BI to Intune, certain Intune dashboards,
I guess lower-level dashboards are possible,
that's what the question is alluding to.
But advanced analytics, how you really take that out,
build your own model, build you own machine-learning model.
Because end of the day,
it's your model that determines how you're able to predict.
What is the accuracy of prediction and whatnot.
Today, definitely the logic apps are designed for a developer.
Logic apps may not be for the developer experience.
But you would still have to augment that with some developer, custom
developer component so that it feeds in the machine learning model.
And then as you go and- >> [CROSSTALK] That's
data science, right?
>> Yeah, that's data science.
>> That's exactly what.
And we've done something very similar.
There's certainly pockets here inside of Microsoft where we're
doing things like financial forecasting, right?
Where we're talking data from On-Premise, using a gateway to
pump that data up into, maybe SQL DB or maybe the Azure Data Lake, right?
And then running Azure Machine Learning types of things that are,
treasury department and others, right?
Go build as a competency, right?
Cuz, back to the thing,
we think that that might be a competitive advantage for us.
So we just go and do things like that to go and
invest in those areas to derive some intelligence.
>> Absolutely, and the machine learning model gallery definitely
has a lot of pre-canned model which others have built and
published for- >> Yeah.
>> Which is used across industry.
So that is something you could use off the shelf or
you could finetune it according to the needs of your business.
Or if you feel that there is a certain bias to certain
variables which is not assigned properly, we could again modify it.
And again republish it.
So that's the model that is taken forward.
Certain glues that need to stitch together to accomplish this
[INAUDIBLE] scenario, yeah, does not exist today.
>> Yeah.
>> Okay, a couple of hopefully quick questions.
Can we get an updated image of the architecture described for teams?
Is there a place online where they could get the view
of the world that you painted with the team creates group?
>> Yeah, I'm sure there is.
I'd have to go look it up and find where that URL is now.
The Office organization is providing a lot of content around Microsoft
team and now fits in a broader Office 365 ecosystem and groups and
where you collaborate with what.
I'll have to find that URL though.
>> Okay, and we can post it with the video.
How can I get resource utilization report of my subscription on Azure
arm like number of cores, number of storage accounts, network data?
>> I guess there is a partial script that existed definitely in
the old portal.
I think there's an equivalent for the new portal as well.
This is same thing, if I get that URL, I'll-
>> Okay,
we'll pass it along with the final video.
Okay, well, we are near the end of our hour, and
I do want to ask one final question of our SMEs.
What is the one tip you wanna leave our audience with today about Cloud?
We'll start with you.
>> Yeah, so today every industry has to undergo or
has already undergone digital transformation.
And that becomes a key differentiator when they go to
market with it.
Having said that, when your computer environment or your infrastructure
behind the scenes has to be the power that pulls you through.
So when you go to the cloud with that in context,
you just don't want to be lifting and
shifting your old legacy stuff in the same manner as it were.
You want to add intelligence into it.
That is where analytics or machine learning systems come into play.
Which can help you predict better for your scenarios or
your customer scenarios.
And help them succeed better.
Traditionally, machine learning and things like these
which had heavy reliance on compute, it would be costly.
Because the cost would exorbitant to get ahold of PCs and whatnot.
Whereas now, these are workloads that you could do on demand, which
is why you should be embedding lot more of that into your applications.
So that you have that edge in your business or
your customers' business going forward.
>> I would say you start with the self-service, easier workloads.
When you're starting to go to cloud,
especially status services in Office 365, it's open up for net new.
We talked about green field a lot in this conversation.
Cuz I think when we think about this, when we you think about team
collaboration, group collaboration, or even personal collaboration.
You want to of course maintain what you have and carry it forward.
But part of that is at what point do you actually
start creating net new into cloud?
When I have a net new project created as an Office365 group,
you then get all the benefits of the workloads.
And we talk about teams earlier.
You can use Microsoft Teams.
Start to kick the tires, let people try this things out,
how do people like to collaborate in the cloud.
Do people like SharePoint Online for file management?
Do people in the cloud,
do people use the group capabilities in Outlook?
Would they rather use the instant messaging capabilities in the chat
services with Teams, for example?
So just opening up at self-service, in some ways,
people consider it a scary thing of opening up self-service.
Because, hey, I'm letting people do what they want to in the cloud.
But at the same time, you can still have the appropriate controls in
place to manage who does what and how you manage it.
So you can still protect yourself while you're opening up
self-service in the cloud.
>> Those are great points.
And I like the idea of being able to do green field.
And many customers have,
they've got VMs that they simply need to lift and shift.
They have workloads that they have the opportunity to refactor for
the cloud.
My tip would be for enterprise customers,
leverage your Microsoft technical account manager.
Or leveraging your Microsoft cloud solution architect
to fully understand the capabilities of the cloud,
whether it's Office 365 or whether it's Azure.
What you'll find is that you can definitely decouple applications or
services and be able to host them in the cloud and gain significant
ROI by leveraging the cloud provider's native capabilities.
So that you can really focus on making your business agile.
>> Those are all great points.
I think the thing that I would like to leave behind,
I use this phrase a lot when I talk to customers,
is that it's easier to make your datacenter look
like a cloud than it is to make the cloud look like a datacenter.
And a little bit of that speaks to we've definitely failed many times.
We've had to retry because we've tried to take everything that we did
On-Premise and tried to do it like like in the cloud and
it's just not the same.
Now that's not to say that the 20 some plus years of ITIL
needs to be thrown out the door.
No, quite frankly it needs to be shoved into your developers' brains
in the matter of the timespan that you have to get there.
Which I think is the second challenge,
is don't underestimate how much it takes the people and the process and
the culture to get around, get their heads around how they do live site.
Or what their responsibilities and
accountabilities are as you migrate, go forward.
Or even understand what the power that's been put in their hands to
do great things for their company.
I think it's a completely empowering concept, but
there's a lot of stuff, I think that people need to get through in order
to fully realize the benefits.
>> Great, thank you.
Those were great tips.
We are at the top of our hour and I wanna start by thanking our SMEs for
being here today, for stepping away from your day jobs.
It truly is important to do this.
And thank you, customers, for joining us today.
We truly love the dialogue and the questions and
the opportunity to come talk to you.
You will find this video posted on microsoft.com/ITShowcase.
And there is a wealth of information.
Other videos, technical case studies, even readiness guides.
How we prepared our organization to adopt new technologies very quickly.
We provide those guides to you to use in your own organization.
So visit us at microsoft.com/ITShowcase.
We do live shows weekly.
Please join us again and bring your friends.
Thank you.
>> Thank you.
[MUSIC]
For more infomation >> Chuck Schumer ADMITTED He Tried To TRICK President Trump! – You Won't Like This! - Duration: 2:47. 
For more infomation >> [FREE] Drake x Quavo Type Beat | Finesse ( Prod. by SammieSosza) - Duration: 3:15. 
For more infomation >> How to Keep Your Blog Running [My BEST 3 TIPS] - Duration: 10:37. 
For more infomation >> #EasyWins Episode 8: Batch Content - Duration: 0:34. 
Không có nhận xét nào:
Đăng nhận xét