The path to cloud computing adoption persistently appears complex and risky to both government and enterprise IT leaders, recent surveys show.
This next BriefingsDirect managed cloud methodologies discussion explores how tackling complexity and security requirements upfront helps ease the adoption of cloud architectures. By combining managed services, security solutions, and hybrid cloud standardization, both public and private sector organizations are now making the cloud journey a steppingstone to larger business transformation success.
Listen
to the podcast. Find it on iTunes. Read a full transcript or download a copy.
We'll
now explore how cloud-native apps and services modernization benefit from prebuilt
solutions with embedded best practices and automation. To learn how, we welcome Raj Raman, Chief
Technology Officer (CTO) for Cloud at Unisys, and Jerry Rhoads,
Cloud Solutions Architect at Microsoft. The discussion is moderated by Dana Gardner,
Principal Analyst at Interarbor
Solutions.
Here are some excerpts:
Gardner: Raj, why
are we still managing cloud adoption expectations around complexity and
security? Why has it taken so long to make the path to cloud more smooth -- and
even routine?
Raman: Well,
Dana, I spend quite a bit of time with our customers. A common theme we see -- be
it a government agency or a commercial customer – is that many of them are
driven by organizational mandates and getting those organizational mandates in
place often proves challenging, more so than what one thinks.
Cloud adoption challenges
The
other part is that while Amazon Web Services
(AWS) or Microsoft Azure
may be very easy to get on to, the question then becomes how do you scale up? They
have to either figure out how to develop in-house capabilities or they look to
a partner
like Unisys to help them out.
 |
| Raman |
Cloud security
adoption continues to be a challenge because enterprises still try and wish to
apply traditional security practices to the cloud. Having a security and risk
posture on AWS or Azure means having a good understanding of the shared security
model across user level, application and infrastructure layers of the cloud.
And last, but not least, a
very clear mandate such as digital transformation or a specific initiative, where
there is a core sponsor around it, oftentimes does ease the whole focus on some
of these.
These are some of the reasons
we see for cloud complexity. The applications transformation can also be quite arduous
for many of our clients.
Gardner: Jerry,
what are you seeing for helping organizations get cloud-ready? What best
practices make for a smoother on-ramp?
Rhoads: One of
the best practices beforehand is to determine what your endgame is going to
look like. What is your overall cloud strategy going to look like?
 |
| Rhoads |
Instead of just lifting and
shifting a workload, what is the life cycle of that workload going to look like?
It means a lot of in-depth planning -- whether it's a government agency or
private enterprise. Once we get into the mandates, it's about, “Okay, I need
this application that’s running in my on-premises data center to run in the
cloud. How do I make it happen? Do I lift it and shift it or do I re-architect
it? If so, how do I re-architect for the cloud?”
That’s a big common theme I’m
seeing: “How do I re-architect my application to take better advantage of the cloud?”
Gardner: One
of the things I have seen is that a lot of organizations do well with their
proof of concepts (POCs). They might have multiple POCs in different parts of
the organization. But then, getting to standardized comprehensive cloud adoption
is a different beast.
Raj, how do you make that leap
from spotty cloud adoption, if you will, to more holistic?
One size doesn’t fit all
Raman: We
advise customers to try and [avoid] taking it on as a one-size-fits-all. For
example, we have one client who is trying – all at once – to lift and shift
thousands of applications.
Now, they did a very detailed POC
and they got yield from that POC. But when it came to the actual migration and
transformation, they were convinced and felt confident that they could take it on
and try it en masse, with thousands of applications.
The thing is, in trying to do
that, not all applications are one size. One needs a phased approach for doing
application discovery and application assessment. Then, based on that, you can
determine which applications are well worth the effort [to move to a cloud].
So we recommend to customers that they think of migrations as a phased
approach. Be very clear in terms of what you want to accomplish. Start
small, gain the confidence, and then have a milestone-based approach of
accomplishing it all.
Gardner: These
mandates are nonetheless coming down from above. For the US federal government,
for example, cloud has become increasingly important. We are expecting
somewhere in the vicinity of $3.3
billion to be spent for federal cloud in 2021. Upward of 60 percent of
federal IT executives are looking to modernization. They have both the cloud
and security issues to face. Private sector companies are also seeing mandates to
rapidly become cloud-native and cloud-first.
Jerry, when you have that
pressure on an IT organization -- but you have to manage the complexity of many
different kinds of apps and platforms -- what do you look for from an alliance
partner like Unisys to help make those mandates come to fruition?
Rhoads: In working
with partners such as Unisys,
they know the customer. They are there on the ground with the customer. They
know the applications. They hear the customers. They understand the mandates.
We also understand the mandates and we have the cloud technology within Azure.
Unisys, however, understands how to take our technology and integrate it in
with their end customer’s mission.
Gardner: And security
is not something you can just bolt on, or think of, after the fact in such
migrations. Raj, are we seeing organizations trying to both tackle cloud adoption
and improve their security? How do Unisys and Microsoft come together to
accomplish those both as a tag team rather than a sequence, or even worse, a
failure?
Raman: We
recently conducted a survey of our stakeholders, including some of our
customers. And to no surprise security -- be it as part of the migrations or in
scaling up their current cloud initiatives – is by far a top area of focus and concern.
We are already partnering with
Microsoft and others with our flagship security offering, Unisys
Stealth. We are not just in collaboration but leapfrogging in terms of innovation.
The Azure cloud team has released a specific
API to make products like Stealth available. This now gives customers more
choice and it allows Unisys to help meet customers in terms of where they are.
Also, earlier
this year we worked very closely with the Microsoft cloud team to release Unisys
CloudForte for Azure. These are foundational elements that help both
governments as well as commercial customers leverage Azure as a platform for
doing their digital transformation.
The
Microsoft team has also stepped up and worked very closely with the
Unisys team developers and architects to make these services native on
Azure.
The Microsoft team has also
stepped up and worked very closely with the Unisys team developers and architects
to make these services native on Azure, as well as help customers understand
how they can better consume Azure services.
Those are very specific
examples in which we see the Unisys collaboration with Microsoft scaling really
well.
Gardner: Jerry,
it is, of course, about more than just the technology. These are, after all, business
services. So whether a public or private organization is making the change to
an operations model -- paying as you consume and budgeting differently -- financially
you need to measure and manage cloud services differently.
How is that working out? Why
is this a team sport when it comes to adopting cloud services as well as changing
the culture of how cloud-based business services are consumed?
Keep pay-as-you-go under control
Rhoads: One
of the biggest challenges I hear from our customers is around going from a
CAPEX model to an OPEX model. They don’t really understand how it works.
CAPEX is a longtime standard
-- here is the price and here is how long it is good for until you have to then
re-up and buy new piece of hardware or re-up the license, or whatnot. Using
cloud, it’s pay-as-you-go.
If I launch 400 servers for an
hour, I’m paying for 400 virtual machines running for one hour. So if we don’t
have a governance strategy in place to stop something like that, we can wind up
going through one year's worth of budget in 30 days -- if it's not governed, if
it's not watched.
And that's why, for instance,
working with Unisys
CloudForte there are built-in controls to where you can go through and ping
the Azure cloud backend -- such as Azure
Cost Management or our Cloudyn
product -- where you can see how much your current charges are, as well as
forecast what those charges are going to look like. Then you can get ahead of
the eight ball, if you will, to make sure that you are actually burning through
your budget correctly -- versus getting a surprise at the end of the month.
Gardner: Raj, how
should organizations better manage that cultural shift around cloud consumption
governance?
Raman: Adding
to Jerry’s point, we see three dimensions to help customers. One is what Unisys
calls setting
up a clear cloud architecture, the foundations. We actually have an offering
geared around this. And, again, we are collaborating with Microsoft on how to codify
those best practices.
In going to the cloud, we see
five pillars that customers have to contend with: cost, security, performance,
availability, and operations. Each of these can be quite complex and very deep.
Rather than have customers
figure these out themselves, we have combined product and framework. We have codified
it, saying, “Here are the top 10 best practices you need to be aware of in
terms of cost, security, performance, availability, and operations.”
It makes it very easy for the
Unisys consultants, architects, and customers to understand at any given point
-- be it pre-migration or post-migration -- that they have clear visibility on
where they stand for their review on cost in the cloud.
We are also thinking about
security and compliance upfront -- not as an afterthought. Oftentimes customers
go deep into the journey and they realize they may not have the controls and
the security postures, and the next thing you know they start to lose
confidence.
So rather than wait for that,
the thinking is we arm them early. We give them the governance and the policies
on all things security and compliance. And Azure has very good capabilities
toward this.
The third bit, and Jerry touched on this, is overall financial governance. The ability to think about -- not just cost as a matter of spinning a few Azure resources up and down – but in a holistic way, in a governance model. That way you can break it up in terms of analyzed or utilized resources. You can do chargebacks and governance and gain the ability to optimize cost on an ongoing basis.
These are distinctive
foundational elements that we are trying to arm customers with and make them a
lot more comfortable and gain the trust as well as the process with cloud
adoption.
Gardner: The good
news about cloud offerings like Azure and hybrid cloud offerings like Azure Stack
is you gain a standardized approach. Not necessarily one-size-fits-all, but an
important methodological and technical consistency. Yet organizations are often
coming from a unique legacy, with years and years of everything from mainframes
to n-tier architectures, and applications that come and go.
How do Unisys and Microsoft work
together to make the best of standardization for cloud, but also recognize
specific needs that each organization has?
Different clouds, same experience
Rhoads: We
have Azure Stack for on-premise Azure deployments. We also have Azure
Commercial Cloud as well as Azure
Government Cloud and Department
of Defense (DoD) Cloud. The good news is that they use the same portal,
same APIs, same tooling, and same products and services across all three
clouds.
Now, as services roll out,
they roll out in our Commercial Cloud first, and then we will roll them out
into Azure Government as well as into Azure Stack. But, again, the good news is
these products are available, and you don’t have to do any special
configuration or anything in the backend to make it work. It’s the same
experience regardless of which product the customer wants to use.
Unisys
CloudForte works with Azure Stack, with Commercial, and Azure for
Government. For the end customers it's the same cloud services that they
expect to use. The difference is just where those cloud services live.
What’s more, Unisys CloudForte
works with Azure Stack, with Commercial, and Azure for Government. For the end
customer it's the same cloud services that they expect to use. The difference
really is just where those cloud services live, whether it's with Azure Stack
on-premises, on a cruise ship or in a mine, or if you are going with Azure
Commercial Cloud, or if you need a regulated workload such as a FedRAMP high workload or an IC4, IC5
workload, then you would go into Azure Government. But there are no different
skills required to use any of those clouds.
Same skill set. You don’t have
to do any training, it’s the same products and services. And if the products
and services aren't in that region, you can work with Unisys or myself to
engage the product teams to put those products in Azure Stack or in Azure for
Government.
Gardner: How
does Unisys
CloudForte managed services complement these multiple Azure cloud environments
and deployment models?
Rhoads:
CloudForte really further standardizes it. There are different levels of
CloudForte, for instance, and the underlying cloud really doesn’t matter, it’s going
to be the same experience to roll that out. But more importantly, CloudForte is
really an on-ramp. A lot of times I am working with customers and they are
like, “Well, gee, how do I get started?”
Whether it’s setting up that subscription
in-tenant, getting them on-board with that, as well as how to roll out that POC,
how do they do that, and that’s where we leverage Unisys and CloudForte as the
on-ramp to roll out that first POC. And that’s whether that POC is a bare-bones
Azure virtual network or if they are looking to roll out a complete soup-to-nuts
application with application services wrapped around it. CloudForte and Unisys
can provide that functionality.
Do it your way, with support
Raman: Unisys
CloudForte has been designed as an offering on top of Azure. There are two key
themes. One, meet customers where they are. It's not about what Unisys is
trying to do or what Azure is trying to do. It's about, first and foremost, being
customer obsessed. We want to help customers do things on their terms and do it
the right way.
So CloudForte has been
designed to meet those twin objectives. The way we go about doing it is -- imagine,
if you will, a flywheel. The flywheel has four parts. One, the whole consumption
part, which is the ability to consume Azure workloads at any given point.
Next is the ability to run commands,
or the operations piece. Then you follow that up with the ability to accelerate
transformations, so data migrations or app modernization.
Last, but not least, is to transform
the business itself, be it on a new technology, artificial
intelligence (AI), machine
learning (ML), blockchain,
or anything that can wrap on top of Azure cloud services.
The beauty of the model is a
customer does not have to buy all of these en masse; they could be fitting into
any of this. Some customers come and say, “Hey, we just want to consume the
cloud workloads, we really don’t want to do the whole transformation piece.” Or
some customers say, “Thank you very much, we already have the basic consumption
model outlined. But can you help us accelerate and transform?”
So the ability to provide
flexibility on top of Azure helps us to meet customers where they are. That’s
the way CloudForte has been envisioned, and a key part of why we are so
passionate and bullish in working with Microsoft to help customers meet their
goals.
Gardner: We
have talked about technology, we have talked about process, but of course
people and human capital and resources of talent and skills are super important
as well. So Raj, what does the alliance between Unisys and Microsoft do to help
usher people from being in traditional IT to be more cloud-native practitioners?
What are we doing about the people element here?
Expert assistance available
Raman: In
order to be successful, one of the big focus areas with Unisys is to arm and
equip our own people, be it at the consulting level, a sales-facing level, either
doing cloud architectures or even doing cloud delivery, across the stripe, rank
and file. There is an absolute mandate to increase the number of certifications,
especially the Azure certifications.
In fact, I can also share that
Unisys, as we speak, every month we have a doubling of people who are across
the rank of Azure 300 and the 900. These are the two popular certifications, the
whole Azure stack of it. We have now north of 300 trained people, and maybe my
number is at the lower end. We expect the number to double.
So we have absolute commitment,
because customers look to us to not only come in and solve the problems, but to
do it with the level of expertise that we claim. So that’s why our commitment
to getting our people trained and certified on Azure is a very important piece
of it.
Gardner: One of
the things I love to do is to not just tell, but to show. Do we have examples
of where the Unisys and Microsoft alliance -- your approach and methodologies
to cloud adoption, tackling the complexity, addressing the security, and looking
at both the unique aspect of each enterprise and their skills or people issues
-- comes all together? Do you have some examples?
Raman: The California
State University is a longstanding customer of ours,
a good example where they have transformed their own university infrastructure
using Unisys CloudForte with a specific focus on all things hybrid cloud. We
are pleased to see that not only is the customer happy but they are quite eager
to get back to us in terms of making sure that their mandates are met on a consistent
basis.
Our federal agencies are usually
reluctant to be in the spotlight. That said, what I can share are representative
examples. We have some very large defense establishments working with us. We
have some state agencies close to the Washington, DC area, agencies responsible
for the roll-out of cloud consumption across the mandates.
We are well on our way in not only
working with the Microsoft Azure cloud teams, but also with state agencies. Each
of these agencies is citywide or region-wide, and within that they have a health
agency or an agency focused on education or social services.
In our experience, we are
seeing an absolute interest in adopting the public clouds for them to achieve
their citizens’ mandates. So those are some very specific examples.
Gardner: Jerry,
when we look to both public and private sector organizations, how do you know
when you are doing cloud adoption right? Are there certain things you should
look to, that you should measure? Obviously you would want to see that your
budgets are moving from traditional IT spending to cloud consumption. But what
are the metrics that you look to?
The measure of success
Rhoads: One
of the metrics that I look at is cost. You may do a lift and shift and maybe
you are a little bullish when you start building out your environments. When
you are doing cloud adoption right, you should see your costs start to go down.
So your consumption will go up,
but your costs will go down, and that’s because you are taking advantage of either
platform as a service (PaaS) in the cloud, and being able to auto-scale out, or
looking to move to say Kubernetes and
start using things like Docker containers
and shutting down those big virtual machines (VMs), and clusters of VMs, and
then running your Kubernetes services on top of them.
When you see those costs go
down and your services going up, that’s usually a good indicator that you are
doing it right.
Gardner: Just
as a quick aside, Jerry, we have also seen that Microsoft
Azure is becoming very container- and Kubernetes-oriented, is that true?
Rhoads: Yes,
it is. We actually have Brendan Burns,
as a matter of fact, who was one of the co-creators of Kubernetes during his
time at Google.
Gardner: Raj,
how do you know when you are getting this right? What do you look to as chief
metrics from Unisys's perspective when somebody has gone beyond proof of
concept and they are really into a maturity model around cloud adoption?
Raman: One
of the things we take very seriously is our mandate to customers to do cloud on
your terms and do it right. And what we mean by that is something very specific,
so I will break it in two.
One
is from a customer-led
metric perspective. We actually rank ourselves very seriously in terms
of Net Promoter Score. We have one of the highest in the industry
relative to the rest of our
competitions. And that's something that's hard-earned, but we keep
striving to
raise the bar on how our customers talk to each other and how they feel
about
us.
The other part is the ability
to retain customers, so retention. So those are two very specific customer-focused
benchmarks.
Now, building upon some of the
examples that Jerry was talking about, from a cloud metric perspective, besides
cost and besides cost optimization, we also look at some very specific metrics,
such as how many net-net workloads are there under management. What are some of
the net new services that are being launched? We especially are quite curious
to see if there is a focus in terms of Kubernetes or AI and ML adoption, are
there any trends toward that?
We
rank ourselves very seriously in terms of Net Promoter Score. We have
one of the highest in the industry, but we keep striving to raise the
bar on how our customers talk to each other and feel about us.
One of the very interesting
ones that I will share, Dana, is that some of our customers are starting to
come and ask us, “Can you help set up an Azure
Cloud center of excellence within our organization?” So that oftentimes is
a good indicator that the customer is looking to transform the business beyond
the initial workload movement.
And last, but not the least,
is training, and absolute commitment to getting their own organization to
become more cloud-native.
Gardner: I
will toss another one in, and I know it's hard to get organizations to talk
about it, but fewer security breaches, fewer days or instances of downtime
because of a ransomware attack.
So it's hard to get people to talk about it if you can't always prove when you don’t
get attacked, but certainly a better security posture as compared to two years,
three years ago would be a high indicator on my map as to whether cloud is
being successful for you.
All right, we are almost out
of time, so let's look to the future. What comes next when we get to a maturity
model, when organizations are comprehensive, standardized around cloud, have
skills and culture oriented to the cloud regardless of their past history? We
are also of course seeing more use of the data behind the cloud, in operations
and using ML and AI to gain AIOps
benefits.
Where can we look to even
larger improvements when we employ and use all that data that’s now being
generated within those cloud services?
Continuous cloud propels the future
Raman: One
of the things that’s very evident to us is, as customers start to come to us and
use the cloud at significant scale, is it is very hard for any one organization.
Even for Unisys, we see this, which is how do you get scaled up and keep up with
the rate of change that the cloud platform vendors such as Azure are bringing
to the table; all good innovations, but how do you keep on top of that?
So that’s where a focus on
what we are calling as “AI-led operations” is becoming very important for us.
It’s about the ability to go and look at the operational data and have these
customers go from a reactive, from a hindsight-led model, to a more proactive
and a foresight-driven model, which can then guide, not only their cloud
operations, but also help them think about where they can now leverage this
data and use that Azure infrastructure to then launch more innovation or new
business mandates. That’s where the AIOps piece, the AI-led operations piece,
of it kicks in.
There is a reason why cloud is
called continuous. You gain the ability to have continuous visibility via compliance
or security, to have constant optimization, both in terms of best practices,
reviewing the cloud workloads on a constant basis and making sure that their
architectures are being assessed for the right way of doing Azure best practices.
And then last, but not the
least, one other trend I would surface up, Dana, as a part of this, which is we
are starting to see an increase in the level of conversational bots. Many
customers are interested in getting to a self-service mode. That’s where we see
conversational bots built on Azure or Cortana will become more
mainstream.
Gardner: Jerry,
how do organizations recognize that the more cloud adoption they have, the more
standardization, the more benefits they will get in terms of AIOps and a virtuous
adaption pattern kicks in?
Rhoads: To
expand on what Raj talked about with AIOps, we actually have built in a lot of
AI into our products and services. One of them is with Advanced
Threat Protection (ATP) on Azure. Another one is with anti-phishing
mechanisms that are deployed in Office 365.
So as more folks move into the
cloud, we are seeing a lot of adoption around these products and services, as
well as we are also able to bring in a lot of feedback and do a lot of learning
off of some of the behaviors that we are seeing to make the products even
better.
DevOps integrated in the cloud
So one
of things that I do, in working with my customers is DevOps, how do we employ DevOps
in the cloud? So a lot of folks are doing DevOps on-premises and they are doing
it from an application point of view. I am rolling out my application on an infrastructure
that is either virtualized or physical, sitting in my data center, how do I do
that in the cloud, why do I do that in the cloud?
Well, in the cloud everything
is software, including infrastructure. Yes, it sits on a server at the end of the
day; however, it is software-defined, being it is software-defined, it has an
API, I can write code. So therefore if I want to blow out or roll out a suite
of VMs or I want to roll out Kubernetes clusters and put my application on top
of it, I can create definable, repeatable code, if you will, that I can check
into a repository someplace, press the button, and roll out that infrastructure
and put my application on top of it.
So now deploying applications, especially with DevOps in the cloud, it's not about I have an operations team and then I have my DevOps team that rolls out the application on top of existing infrastructure. Instead I actually bundle it altogether. I have tighter integration, which means I now have repeatable deployments and I can do my deployments, instead of doing them every quarter or annually, I can do deployments -- I can do 20, 30, 1,000 a day if I like, if I do it right.
Listen
to the podcast. Find it on iTunes. Read a full transcript or download a copy. Sponsor: Unisys and Microsoft.
You may also be
interested in:
- How containers are the new basic currency for pay as you go hybrid IT
- HPE strategist Mark Linesch on the surging role of containers in advancing the hybrid IT estate
- The venerable history of IT systems management meets the new era of AIOps-fueled automation over hybrid and multicloud complexity
- How the Catalyst UK program seeds the next generations of HPC, AI, and supercomputing
- HPE and PTC Join Forces to Deliver Best Outcomes from the OT-IT Productivity Revolution
- How rapid machine learning at the racing edge accelerates Venturi Formula E Team to top-efficiency wins
- The budding storage relationship between HPE and Cohesity brings the best of startup innovation to global enterprise reach
- HPE’s Erik Vogel on what's driving success in hybrid cloud adoption and optimization
- IT kit sustainability: A business advantage and balm for the planet
