[music playing] Good morning,
my name’s Dave McCann, and it’s a pleasure to be
speaking to you at re:Invent 2020. In the next session I am going to be
discussing the experiences that we have seen from a number
of our large customers in their transforming
key business applications migrating a substantial portion
of their application portfolio to the AWS Cloud
and then the approach they have taken to governance and operations
of their application portfolio that is running on the Cloud. Later in the talk I will be here
with Paul von Autenried, the CIO of Bristol Myers Squibb, a major US corporation who are doing
pharmaceutical delivery worldwide. Let’s talk about the evolution and modernization
of the application portfolio. Today, companies are evolving
faster than every before and they are having to so
for multiple reasons. At one level there are billions
of consumers now around the world
on smart phones whose interaction with their
institutions is done over a phone. There are millions of software
developers more than every before who are innovating on behalf
of corporations and suppliers, transforming the customer
experience of the supply chain, taxpayer for government
and in addition, AWS is introducing
disruptive Cloud technology for changing the economics and the innovation opportunity
for companies around the world and, as a result,
companies are developing applications in new ways faster
than ever before. Now, those drivers translate
into corporations and governments in different ways
and as we have studied over a thousand migration
projects around the world, today we’re going to be
discussing about how these innovations
are considered. Now the application portfolio level,
the CIO, has a broad portfolio of applications
and let’s say we’re seeing on average over a thousand applications
in a large enterprise, in a mid-sized company
or organization that could be three
to four hundred applications, and in a small business it might be
fifty to a hundred applications, but there’s a portfolio
of applications that have been built up
over the years, some of which are very new
some of which are a little older, and the CIO is assessing
with the head of their organization which particular applications
need to be modernized first, which ones need to be
completely rewritten and very often the priority is over
customer experience or supply chain. While the CIO is looking
at the application portfolio, the CISO of the organization
is looking at the security posture, and as the applications
are modernized, there’s a lot of change
going on around network, end-point and threat detection. The head of infrastructure
and operations used to running resources
on-premises now has a set of resources
running in the Cloud and there are over a dozen domains
that are up for refresh across storage
and network and security and we’ll talk about that later and then an entity
has a VP of big data who is looking at databases,
analytics, visualization and the tool refresh possibilities
in the journey to Cloud, and all four
of these ways of change are being considered
by companies simultaneously. Now, as this process goes on,
the development methodology is also changing to modern tools
to an agile methodology approach and this involves small teams
continuous innovation and this results
in a velocity of change applied to applications
more than ever before, and this is causing both
the opportunity for automation and a change in the operating model
that many companies have. At Amazon for example,
we deliver over 15,000 changes a day and compare that to the old days when perhaps only one change was
applied to an application a month, and so this change
in velocity of development requires a change in tooling
and a change in approach. Now, many companies are well
into this journey and we’re going to be talking
about learnings that we’ve been observing
from leaders around the world. Let’s start with $11 billion
Aon Insurance who are a global
reinsurer headquartered in both the UK
and the US and with development
centers around the world, and I recently had the pleasure
of talking with Rajeev Khanna the CTO of Aon. Rajeev would talk about
the most impactful thing being about moving
to the Cloud that they had been able
to cut their ideation time from concept
to an application by half, and that ideation time reduction means that they are able
to introduce new applications far faster that can benefit
their customer experience and how their financial
advisers provide advice to their tens of thousands
of customers around the world
in over a 120 countries. They have also completely
modernized their data analytics and big data strategy
and are leveraging AWS Redshift to analyze volumes of data
far faster than ever before. Now in another sector
which is government, a good example of a customer well
along this journey is the US Navy, and they have a major supply
chain system setup and they have actually moved
72,000 users with multiple incidences of SAP
onto the AWS Cloud, and there they have been able
to leverage the compute power and compute scalability of AWS to reduce some of their
compute workloads by up to 80% and they have moved hundreds
of applications onto the Cloud. So, taking that as a start of
the learnings that we’re seeing, we actually did a study with IDC and we’re done an internal study
in over a thousand migrations and the common pattern that we
can see that companies can yield, is on average an infrastructure
savings of over 31%, in addition a reduction
in security incidence on their application portfolio, and an increase in staff
productivity of over 62%, and a velocity of innovation
where changes apply to the application portfolio
are up by over three times and these are tremendous
changes in productivity that are made possible. Let’s talk in a little more detail about how customers
are making that happen. So, let’s click down
a little on lessons that we have learned
from a number of customers who have been successful
in the migration and modernization of those applications.
There are three particular lessons that we are going to give you
feedback on this talk today. The first one is going to focus
on how we approach the planning
of your application portfolio and the classification
of which applications to be given which treatment. The second pillar of learning
that we want to discuss is the building
a governance framework and an operating model for the
applications that are running on AWS, and the third pillar of learning
that we want to bring you is insight across a number of
the cases we’ll quote today where our key learning is to work
with a consulting partner either AWS Professional Services or
one of our many partners worldwide. So, let’s start by clicking down
further on the application
portfolio planning. As we work with customers today to modernize
their application portfolio, the starting state that we’re seeing
as a common pattern, is that in this application portfolio
of a thousand applications for a large enterprise of 500
from a midmarket organization, is that most applications are still in the server-based
architecture running on-premises. There are a number of applications
that we’re noticing are now modernized
on some form of containers, and then many companies have
a small number of applications that have moved to software
as a service delivery model, but today that ratio of the portfolio
is perhaps at 5%. This is the starting state
that CIOs begin with. Now the choices are many
to move onto AWS and some of those choices
that customers have to consider are whether to bring
their application over in its existing
server-based approach, whether to refactor
and modernize their application onto a container-based model, or to actually rewrite
the application and leverage some of the power and capability
of a Lambda Serverless Model in order to give them
less operating burden and ability to focus on Core execution, or to increase the ratio
of software as a service, and in AWS Marketplace
for example today we have over 800 software
as a service providers available and operating on our API
for provisioning and entitling into your AWS accounts.
So, this is the kind of journey that many of our customers
are now going through. Now in planning the application
portfolio there’s some key observations
that we’ve looked at recently, as we have analyzed these thousand
migration projects worldwide. Our first learning is that 15-20%,
sometimes as low as 10%, but 15-20% of your applications
are probably going to be retired. These applications are old or out
of date and in some cases, as you build a new application, you’re just going to retire
that application and the operating cost will go away and you have to think about
that dimension of your portfolio. We also recently talked
to our customer advisory board with the intention
of many companies is to move 20-30%
of the application portfolio out to software
as a service provider. Companies have already started using
solutions like Salesforce or Workday and so they are moving to a confident
model of Software as a Service. So, if you’re retiring 10%
moving to 20-30% SaaS, what do you do with the remaining
60% of your portfolio? There the decisioning is
Lift & Shift, modernization, or to prioritise the complete
rewrite and transformation of a few critical business processes
and that’s kind of the model that we are seeing
prevalent around the world. Now, we started out doing migrations as a professional services
practice with customers in 2015, in the North East of the US we built
a global migration practice, we have now partnered with
hundreds of consulting firms, and really in 2020, we have
a proven repeatable methodology. There’s an assessment phase,
there is what we call a ‘mobilization phase’ and there’s a migration
and modernization phase, and now that we have done
this many times and we have observed
and learned from the experience, we’re going to be clicking
into what that means but we’ve got a pretty
proven methodology that gives us confidence now
that customers can do this in as little as six months, or they can take it over
a longer period of time. One of the key learnings
that we have got is that post assessment
you want to move quickly into the action
of early migration and there what we’re doing
in the mobilization phase is focusing on getting teams
at the customer to quickly commence the migration
of up to ten applications, and we’re also focusing
on the automation of what we call a ‘Landing Zone,’ and we used to do that
on a custom basis but we now productize that into
a service called ‘AWS Control Tower,’ and just recently
we rolled Control Tower out to five more regions
around the world, where a secure Landing Zone
can be established in as little as a couple of days
or at least a week, and then on top of that Landing Zone you migrate
and modernize applications. So, this is a proven journey
it’s more repeatable and easier
than it ever has been before. One of the reasons
that makes it easier is that the toolset
that is available to you today is far richer than it has ever been
and frankly globally, available. In the assessment phase we bought
two companies at the end of 2018. We bought a company
in Canada called ‘TSO Logic’ and we bought a company
that we have called ‘CloudEndure.’ We have now turned TSO Logic
into a new service called ‘Migration Evaluator,’ we have made
that globally available this year and Migration Evaluator
is a site experience where customers can do
an initial directional assessment of the business case
to migrate an application let’s say from
Windows Server onto AWS. We’ve also rolled out Migration Hub into multiple regions
in Asia Pacific, EMEA and the US, and Migration Hub
is where you can plan and control the migration of the application,
and in the mobilization phase we have done a lot of work
on AWS Control Tower. We have recently integrated it
with 20 third-party vendors and with Control Tower you can set up
a very powerful multi-account model for governance
of your resource, governance of who has access
to what accounts, and the use of Control Tower
and Service Catalog accelerates the migration project,
and on the migration side of the move we can now enable you
to use CloudEndure, we have a CloudEndure
migration factory and we also have
a set of software partners. We have over
19 software tool partners that we’ve established work
really well in enabling you to get
the application migrated onto the AWS Cloud. So, tools and methodology
are making this a very predictable
and easy process. Now as we do that we have
also observed that in some industries while customer scan more the entire
portfolio in seven months, a number of customers are telling us
that given the size of the portfolio, particularly in financial services,
pharmaceutical, public sector, and perhaps oil and gas,
the application portfolio is actually measured
in thousands of applications and some customers anticipate
being in an on-premise’s and AWS Cloud model for
perhaps two or three years, and across AWS in its entirety,
you’re seeing us respond to that with a series of service responses
that give the customer the ability to have a common environment model
whether on-premises or on AWS. Recent examples of that
would be VMware on AWS, where we launched with VMware
the ability for you to bring the familiarity
of your tools to running VMware on AWS. A second example of that would be
that we launched AWS Outposts, which is an AWS appliance
that can be writing your data center and has a number of AWS services locally controlled down
on your network but connected
into the AWS network, and we’ve even taken our managed
database service RDS and put it down
on top of VMware. Now just this week,
Andy has announced two days ago EKS Anywhere and ECS
Anywhere, and this is a further motion
that we responded to customers who want to be using
our container experience on AWS and want to have that
same container experience on on-premises deploy model, and a couple of weeks ago
we announced with IBM Red Hat the availability of Red Hat
OpenShift on AWS, ROSA, such that if you’re using
an OpenShift environment you can experience
the OpenShift environment both on-premises and on AWS. Now we’ve not only done this at the compute
and development environment level but at the storage and data movement
level we have been innovating further with our Snowball appliances
and while we sent out Snowball appliances
to move data up onto AWS, we’re actually finding
that many customers keep a Snowball appliance
on prem to have that same AWS experience
on prem as well on AWS, and on the management side
of the house you’re going to hear us make continual announcements
about AWS Systems Manager being able
to apply management to incidence resources
not only on AWS but on-premises and that’s been
an evolution of our systems
management tool. These all make it easier for you
to migrate your applications and rescale your teams
as they move to the Cloud. And another good example of this
in a regulated industry and a customer using these kinds
of approaches, is Sallie Mae. Sallie Mae is
a US financial institution, the service or consumer
whose persona is that of student, and they originate loans
for students going to university, and Sallie Mae wanted to reduce
the data center cost, spend more of their resources
on application development, and contracted with a service
called ‘AWS Managed Services’ in order to be able to quickly migrate a set
of applications on the AWS Cloud. Now, in so doing, they have migrated
71 applications up onto AWS and AMS
has taken over the automation of the patch management
of their service fleet and that patch management
has caused a 30% efficiency in their operating velocity and it’s allowed them
to innovate faster applications that service
their student consumer. So, this is an example
of how this is playing out. Okay, so what I would like
to do now is move on to the next pillar
of our learning, which is around
the operating model that we want you as a customer
to establish in order for you to manage
your applications well and also to accomplish
cost efficiency to free up more resources to focus
on application development, and in accomplishing that one
of the tools that we provide, and one of the tools that
our third-party partners provide, that we can recommend. Invent I gave a talk
with Tristan Baker who’s the chief architect
at Intuit and Intuit are a media software vendor in the US
who provide multiple packages to both taxpayer consumers
and small businesses, and they moved hundreds
of applications onto AWS, and his observation last year was that in order to have
great confidence on migrating all the applications,
they had established with our team a service management framework
including tools like AWS Config, CloudFormation
and Service Catalog, in conjunction with
other third-party tools, and they have been
pretty early and fast on the migration
of their portfolio. We have since taken
a further look at that learning and we’ve seen this across
a number of our customers. So, as we have studied
1,000 migrations worldwide, there are a number of
IT domains that you’ve already established
in your organization and we’ve seen that
in particular five or six domains require
a refresh of tooling, some standardization
of decisions. What are those? The first one are your
OS environments and your patch
management approach, and whether you are standardizing
on a mix of Linux, Windows Server,
or Red Hat OS’s, that whole OS standardization
and patch management approach can yield
tremendous efficiency. A separate team will very often
look at security, identity, and network as it is modernized and that will cause opportunities
for both cost rationalization and increase your ability to handle
the velocity of CI/CD model. Storage and databases become
another domain of consideration and we’ll take about
all your choices there, and then monitoring observability,
control mechanisms, those all become very relevant in how they get modernized
as you move to the Cloud. And if you modernize all of these
domains staff have to be retrained, new skills have to be acquired
old skills have to be lost, and we are working hard to enable
all of this to made possible so that you can go faster, spend less money
and time on operations. Now as you do that we have
built out our own management
governance portfolio and in 2018 we changed the console
to management and governance as a title area of the console, and we continue to innovate
with new services in our console to solve
for your operational requirements. AWS Control Tower and organizations
become central mechanisms that we want you to be using
in a multi-account environment, where you will have dev accounts, test accounts and prod accounts,
foundation accounts, audit accounts, and we want you to have
an account model that works well for distributed teams
in a small team environment. We also want to give you
the capabilities for laying in governance
and compliance controls using things like Service Catalog and for provisioning
and automation with CloudFormation and Marketplace provisioning
third-party software and lay in tools
with CloudWatch and CloudTrail along with third-party tools
for observability, and ultimately to have
operations controls both in terms of instantiating
and running resources and spend management
controls and visibility tools over budgets and resources. So, we are innovating
on all of these tools that we want you to use
and as you migrate, modernize, this becomes
a collection of tools that accomplish
your operating efficiency. Now, we’ve learned in the last year that customers want more
and more prescriptive guidance and the way we deliver
prescriptive guidance to customers is through our
Well-Architected Frameworks. Our Well-Architected Frameworks
on a program designed and delivered by a Solution Architecture organization in conjunction
with Professional Services, where we give you
prescriptive guides, web pages, documentation,
prototype code, and I am pleased
to announce today Well-Architected Management
and Governance Lens. This is a particular prescriptive
set of guidance around management and governance
of your application portfolio. We have looked
at best practices that we have seen in these
migrations around the world. We have documented integrations
to 20 third-party vendors including vendors like ServiceNow,
Atlassian and many others, and we have snippets of code
and connectors to these tools to compress the time that it requires you to set a highly efficient
operating model for your teams. What we learned with
Tristan last year at Intuit was that Intuit stood up what they call their
own service management framework and what we now see
is that every customer has their own private
mix of tools that that customer
wants to use with services from AWS laid out for different IT functions, whether it’s the CIO
looking at the portfolio, whether it’s the Cloud
center of excellence managing resources
and infrastructure as code, whether it’s the monitoring
or operation tools embedded into your DevOps pipelines,
or whether it’s IT, finance and procurement looking at spend management
on all your assets. Every customer is building their
own service management framework and in
the Well-Architected Lens we’re going to become
far more prescriptive to shorten the time for you
to make those decisions and stand up a framework, and as we do that we’re working
with over 130 management and governance partners
around the world. We have expanded that portfolio and we’re working constantly
with these ISVs to guide them on which
APIs to integrate with and which products best
interoperate with which products. We work with these partners
on many dimensions including the AWS
partner organization, with the Marketplace organization, and over 70 of these vendors
have products in Marketplace where you can automate
the contracting and provisioning. All of this results in you
being able to modernize your software that becomes part
of your operating model, and what we’ve noticed
during the migrations is that software
is getting modernized across four decisioning centers
just in the same way that customers are looking at modernizing
the application and security, vendors are being selected
as best in class for that Cloud journey and this modernization
is happening in four different buying centers and procurement,
legal and IT and finance are also
adopting their processes so that they can go faster and allow you to innovate
more quickly, and as we do that Marketplace
is our service for provisioning all of that software
into your multiple accounts and I am pleased announce today that we’re releasing four more
capabilities on top of Marketplace including support for private
Marketplace APIs, Professional Services as a new
offering type in the Marketplace and an integration
of Marketplace with AWS License Manager that we call
‘managed entitlements.’ We’re also publishing into
Marketplace 200 new vendors to expand the selection
that you have so that you can bring the appropriate
vendor modernize the contract and attach that software
to your running resources in the appropriate account. This is all simplifying
modernization. Now third-party software,
homegrown software, infrastructure as code, gets packaged up ultimately
into a running application and I am pleased
to announce today, that making it easier
for you to manage applications, we’re launching
a new capability in AWS Service Catalog called ‘AWS
Service Catalog AppRegistry.’ Now AppRegistry
is a central hub for declaring
your application resources and we have always enabled you
to manage your resources, now what we’re giving you
an ability is to declare and register and associate
all resources to an application and that application noun
in the Service Catalog API, and we are allowing you
the customer and third parties
to be able to register and manage all the resources
associated with an application. This is going
to increase observability, it’s going to simplify
management and governance, and it becomes a connecting point
for many third-party ISVs that you’re already using
in your service management framework. We’ve already started the work of working
with leading software vendors and I am pleased to say
that we’ve been working with six particular vendors early on
in the integration to AWS Service Catalog AppRegistry. So companies such as New Relic
and AppDynamics in the monitoring
and performance space, Splunk on the SIEM space,
Apptio on spend management, Cutover in the compliance
reporting for banking space, and cloudtamer are also in
the spend management space. These vendors were already
integrating into this API and we anticipate working
with all 130 vendors in the management
and governance portfolio. This reduces your risk
and increases the interoperability that you can accomplish. Now we’re also going to have
other AWS services work with AppRegistry
and I am pleased to preannounce today that in the coming weeks
AWS Systems Manager is going to release
a new capability called ‘Application Manager. Application Manager is
a new module of Systems Manager which is able to discover
and be aware of all the resources that connect an application,
and Systems Manager will both query and refer to AppRegistry
on Service Catalog and use that association rule
to give you the ability to apply Systems Manager verbs and actions
to your application resources. There are over a hundred
functional actions that can be performed
with Systems Manager and you the customer
can also extend Systems Manager as can partners so we’re increasing
the automation surface of Systems Manager
and AppRegistry to all the resources that can be
described using CloudFormation JSON. Again, reducing
your cost of operations and allowing you to move
at a faster pace. Let’s turn to a customer
that is well down the journey of automating
management of applications. We started working with BP,
British Petroleum in the UK, back as early as 2014. I remember actually first
meeting with them in 2015. BP were one of the first
companies in the world to implement a Landing Zone
working with one of our top consulting partners
UK based Cloudreach, and in building out
their Landing Zone they have now moved to a point
where they have migrated over 1,800 applications
onto the AWS Cloud. They have integrated dozens of
instances of their SAP implementation and it has made substantial impact on both the operational efficiency
of their applications and they have actually moved
the resource consumption model to be variable
on many of their application workloads moving from a fixed
cost model to a much more flexible
variable cost model giving them both agility
and cost efficiency as they migrate their application
portfolio to the Cloud, and this was one of the first
companies in the world that made that move. So, this is the importance
of a Cloud operating model and I next now actually want
to introduce Paul von Autenried from Bristol Myers Squibb in a conversation I had
a couple of weeks ago to discuss the learning
from another major customer in the pharmaceutical industry. Today, I am delighted to say
that I have with me Paul von Autenried CIO
of Bristol Myers Squibb, and as many of you probably know,
Bristol Myers Squibb are a global pharmaceutical leader with thousands of employees
around the world developing innovative medicines
for people with serious diseases, and I will let Paul speak more
about what BMS does. At AWS we always learn
from our customers. Bristol Myers Squibb have been
through a pretty amazing journey over the last several years moving and modernizing a set of applications
onto the AWS Cloud, accomplishing both business agility
and cost savings, and so we’re here today to discuss
transformation modernization and governance
of the application portfolio and to learn
some of the great lessons that Paul as a practitioner
can bring to other AWS customers around the world, and to discuss where he thinks
he is going to taking BMS next. So, Paul, good morning! Dave, good morning,
great to see you again and really delighted
to have the chance to share the story of our company’s
digital journey with the Cloud. You know this is a journey
that’s enabled us to fully leverage the rapid
cycles of innovation that we are seeing across
the technology industry for the last five to ten years, and our progress is
especially gratifying when we’re placed in the context
of our company’s mission advancing science and transforming the lives of
patients fighting serious diseases. So, Paul, AWS has been working
with you for many years, I have had the privilege
of meeting you now several times and I have learned a lot
every time I talk to you. At AWS right now we’re talking
a lot about transformation and transformation is a word
that a lot of people use. It can mean dramatic change, we tend to think of dramatic change
of a business process, and then in using the disruptive
capabilities of the AWS Cloud we believe that we can
help companies like yourselves transform a business process,
modernize your applications, talk about how you built
a transformation agenda at BMS and what the drivers were
and what you decided to transform
and modernize at BMS. You know, Dave, our transition
into the Cloud felt a bit more like
an evolution than a revolution, and really it started
15 years ago before I think any of us
were using the term ‘Cloud.’ We recognized that our industry
and our company needed to have increasingly
reliant digital capabilities and we knew that
those digital capabilities are only good as the infrastructure
upon which they rely. So, we recognized
that infrastructure benefits from scale and our mindset
already starting in 2005 was to find ways to leverage
external infrastructure instead of building and maintaining
our own bespoke infrastructure. The service provider relationships
we started at the time were built on this concept
of relying upon their infrastructure instead of having them
manage ours. So, by 2010, we were already
using AWS services for compute and storage, primarily in drug discovery
and development areas. The business driver at the time was the need for massive amounts
of compute for short and unpredictable bursts
of time and frankly, the Cloud was perfect for that. By 2012, we had already adopted
a Cloud First Principle and this Cloud First Principle was an accelerant in the use
of single global instances of Core systems for our company
like ERP, CRM and HR, but the real accelerant
to move most of our Core compute and storage assets
from on-premise to Cloud was a company-level decision
made in 2016 to close a large campus
that happened to include our primary main
global data setter. So, speed of exit
was the top priority, and we used a lift
and shift model. The thought of rearchitecting
the applications is ongoing, but we knew that that would happen
after the lift and shift. Our work there involved
moving thousands of servers over a period of several months and we had to do that
with zero business disruption. But probably the milestone
that I am most proud of in our journey to the Cloud just really occurred
in early part of this year. When we migrated our large
global instance of SAP HANA into the AWS public Cloud,
this is a large system for us, thousands of concurrent users
and interfaces all relying on a single
in memory multi-terabyte database and I would say with that
a moment we pretty much said, “We’re all fully reliant
upon Cloud services.” So, Paul, tell me some of
the business results that Bristol Myers Squibb
has been able to achieve from the transformation
of your application portfolio as you both modernized
and moved to Cloud? You know, Dave, we’ve been
operating predominantly Cloud based infrastructure
for so long it’s almost a bit hard to imagine
operating without it, and I should clarify that the phrase
‘Cloud based infrastructure’ for me is all forms of Cloud,
infrastructure as a service, platform as a service,
software as a service and even BPaaS as a business process
as a service as we say. So, as you know, the true business
benefits require more than just using Cloud services they do require
changes to information flows, technology architecture,
business process, but for us we have seen benefits in a number of different
ways over the years. First of all, our scientists working
in drug discovery and development have been at the forefront of peak
demands of compute capacity, things like molecular modelling
and genomics, clinical pharmacology,
and proteomics, these are the things
that need access and ability to analyze large
volumes of data. We would get requests for compute
clusters of 25 hundred Cores, but they only need it
for a couple of minutes. So, the elasticity of Cloud
based infrastructure enables them to do this. Beyond that initial and I might say,
dare say almost obvious benefit, is that the health care
industry itself has reached tipping point in the adoption of electronic
health and medical records. This is the world’s corpus
knowledge about disease, treatments, and outcomes, and it’s becoming
pervasively digital, which is exploding the size, depth, and complexity of data
that we have access to. So, Cloud based infrastructure
has enabled us to ingest this type of data
often thought of as real-world heath care data and rapidly put it
in the hands of researchers. Beyond that, in the last ten years, our company has seen a significant
amount of acquisitions and investiture activities. Having a Cloud based infrastructure
has enabled us greater agility in the execution of those business
development transactions, and maybe the fourth benefit
just the top of mind here in 2020 is that with the unplanned arrival
of SARS-CoV-2 pandemic and the very sudden shift of our
global workforce to remote working, our predominant Cloud
based infrastructure enabled a relatively
smooth transition. There has been significant
progress in our ability to continue to deliver life-saving
medicines for clinical trails and for continuing care
to patients all around the world. If I had to summarize all these
benefits in a simple phrase, it’s speed of execution. Paul, thank you for that, that’s an impressive synopsis
of pay back and ultimately Andy says, “We want all of our customers
to be able to innovate like a start-up and yet at the same
time operate at scale.” I think some of the benefits
you summarized there really show
the dimensionality of how you can go faster
with development teams, with science teams and yet you’re
scaling out many of the applications on the AWS Cloud
in many different ways with the Cloud Center
of Excellent Team. So, thank you
for that feedback. Sure. It’s interesting that a lot
of our customers talk about the transformation
and the application portfolio and some customers have said that up
to 30-40% of their applications are actually becoming
a software service and as the operator
of AWS Marketplace, we have almost now a thousand
software service vendors that are in Marketplace
that can be provisioned where centrally you’re externalizing
your infrastructure because you’re consuming
the resources of the compute or from the Software
as a Service provider. While you were modernizing
your application portfolio, whether you are using SaaS or whether you’re writing serverless
with your own developers, the top concern
we do hear from customers is you still have to control cost
and you want to optimize for cost. How have you approached
the whole topic of cost management as you transform
that application portfolio to your point amidst the serverless
and SaaS and how is cost efficiency
impacted the organization? We are perpetually focused on cost
not just the initial one- time cost of building
or enhancing an application, but even more concerned with the ongoing maintenance
of that digital capability, the hosting, and the support
what we sometimes call ‘run costs.’ I’ve seen countless spreadsheets
attempting to accurately compare the cost
of on-premise’s hosted infrastructure with Cloud and Colo models,
my perspective is this, almost all those analysis
that I have seen overlook the most difficult cost to measure
which is the cost of having our employees focused
on infrastructure concerns instead of focused
on helping our sciences discover a transformative medicine. Now, with respect to specifically
managing and optimizing our use
of the Cloud and our Cloud cost, our approach has been to fully
leverage all of the tools and go big on automation
and standards. By doing that it’s helped us
create and efficient and what I would refer
to as a ‘fast spin up’ model
of infrastructure components. This is quite a contrast
to the days of provisioning our own infrastructure
in our own data centers which sometimes felt like each server
was a unique work of art. In the Cloud it’s a bit
of a green field opportunity to pour a better foundation
and it’s really only possible because of a strong
central Cloud team. There are many benefits
from standardization that a central Cloud
team provides, but among them speed, supportability,
reliability, and scalability. I’m glad you brought up this issue
because to me cost also translates also into ongoing what we call
management and governance and in the ongoing management
and governance of resource, one way to reduce cost
and be more efficient is automation and I know that
your Cloud engineering team have worked diligently with us
around using one of my service lines which is Service Catalog
in the automation and provisioning of resources
out to your teams and we’re never the only tool. We clearly integrate with
the serverless tools that you use,
our Lambda service, we integrate with AWS Config
and other management tools. As you’ve built out
this operating efficiency, this Cloud flexibility at AWS, how do you think about
the primary tools that you want to have to automate
that management framework? I am conscious you use
a couple of other vendors in conjunction with us,
we think of this as something that we call
a ‘service management framework,’ and our view is we want
to offer you the best tools but we also work with some of
the world’s best software vendors. What other tools might you be using
in your service management framework? Yes, so in our world, if you’re
for example a data scientist you care the most about the time
from your first query to start your project. When we’re bringing
about a new system we care about the time to be live,
to go live for customers. If you’re a commercial
application team bring forth web solutions
for patients and doctors you care about how fast the internet
facing solution can iterate, and none of those are factors that should really be worried
about in the heavy lifting and undifferentiated work
of managing servers. So, what we try to do
is deliver a capability that enables them to have
the environments they need relatively fast. So, with serverless solutions
for example has been a major focus for us.
We’ve pre-packaged governance kits to help teams simplify their journey
through serverless and keep up to date, and in the era of software
as a service applications, serverless kits, help our internal
IT solutions a bit look like mini SAPs because we don’t have
to worry about managing the underlying IT infrastructure. It’s always up
and if no workload runs due to a holiday or an off time,
we’re not paying, it’s the closest thing
to a SaaS experience. Yes, it’s interesting
we actually talk at AWS about abstraction services, so some of my services
sit on top of our other AWS services. We have a 75 plus services
that you can pick from, but we want to be able
to abstract your business users away from the technical complexity
and at the same time we do want to give you
that flexibility and so things
like Service Catalog integrated
with your Lambda functions, do let your Cloud engineering
team provision resources out. We work with all the SaaS vendors
to integrate into a set of APIs so that we can give you
a secure identity management model that you can run
across your organization and we’ve even recently
done things like integrate Service Catalog
with vendors like ServiceNow or other vendors that
other customers may choose because each customer has their own
service management framework to minimise
the operation burden that we put on your scientist,
your researchers, the people who are creating
the shareholder value for BMS. So, I think we are working
very closely with you there. If we think about where you want
to take the digital transformation next of Bristol Myers Squibb and with all the great progress
you have already accomplished, let’s kind of go
from where you’ve been, because now you’re very much Cloud first
you’re a heavy user of AWS, you’ve abstracted away
the complexity, you’re automating your scientists,
what’s next on the Paul von Autenried vision
of where you want to go? You know, Dave, when I think
about all forms of Cloud computing, again infrastructure,
platform, software to service, our world today is characterized
by a lot of Clouds, and so if you will allow me to use
the metaphor of real clouds, I’m pretty sure that
when clouds collide they create thunder and lightning,
so probably high on my wish list viewing this world through
that lens would be a greater methods of frictionless integration
across Clouds. But another thought
that comes to mind is that so much
of what we’ve talked about is the role of technology
enabling new and faster ways
of our company operating, but frankly in my mind, we can and we should expect
that advance technology combined with great therapeutics
in science are going to actually change the way patients experience
and access medicines. But third and probably most important
on my wish list for the next three years, is that all of these great technology
things that we’re doing, is actually going to result in
fundamental breakthroughs in science, in medical science, so that patients are able
to prevail in their fight against serious diseases. I would love to focus on the patient, speaking as a human being
you’re ultimately we all have a persona
of being a patient, so we want Bristol Myers Squibb to continue to do
the tremendous innovation that you have accomplished.
I think your comment on clouds and lightening to use that term,
clearly I think you see AWS doing a lot of work
on identity federation, both with things like our SSO, our partnerships with companies
like OneLogin and Okta, so when you have applications
coming from multiple SaaS vendors, we’re working with
thousands of SaaS vendors and in fact last year
I set up the health care Marketplace where we’re working with dozens
of health care software vendors and we clearly want
to be able to bring you the best selection
of software to service and allow you to easily
configure it and control it and govern it in the BMS well governed management
framework that you have. So, I really want to thank you
for the innovation that you have accomplished with us,
rightly we learn from you, your Cloud Centre of Excellence Team
have been great contributors to rightly improving
one of my services, Service Catalog, and we really want
to continue to work with you and help you solve those patient
challenges all around the world. Paul, thank you. Dave, thanks very much
for having me and giving me the opportunity to tell the story
of our digital journey. It’s very exciting the things
we are doing at BMS and the impact it’s having
on health care, thanks. Thank you. What a great talk with Paul
from Bristol Myers Squibb and a lot of lessons there
that I think customers on this talk will be very interested
to want to accomplish. You know, Paul moved out
of data center at speed, stood up the largest
SAP instance that we believe we have
running on AWS globally, and is clearly that
in moving to a serverless based application development
methodology is innovating faster for the benefit of Bristol Myers
Squibb patients and customers and is also
accomplishing governance by provisioning out infrastructure
and applications in a service management framework heavily centered
around Service Catalog, CloudFormation and integration
with other third-party tools, very much the kind of thing
that we’re recommending that any customer should be doing
as you transform migrate and govern
your application portfolio. Paul also worked with several
consulting partners in that journey and still today is using
several major consulting partners who are AWS partners, and so the last pillar that I want
to talk about on our talk today is the importance
of working with a partner. In studying over 1,000 migrations that we have seen done
around the world, we have really observed that
where customers involve a partner there are multiple benefits. The journey and modernization
of your application portfolio is also some culture change
and organizational change and involving an external partner you’re not only bringing
in external skills, perhaps to tackle
one part of the project, but you’re also very often
bringing in a consulting partner that can engage with you on
the educational and cultural changes that you probably want
to bring to your own IT staff. So, as we look
at these different pillars, we have seen the companies
that really want to drive
new business outcomes, engage the partner in different
parts of the project. In some cases, they are bringing in
at Accenture or Deloitte or Slalom to actually modernize
a particular application, where a large global integrator
may have deep expertise in modernizing
a consumer application, in modernizing and rewriting
a supply-chain application or something to do with data science
and data analytics. On the other hand, some customers want to point
a partner at the operating model, at the infrastructure operations
cost, and as Sallie Mae used AWS Managed Services, so you might see
other customers like Aon turn to somebody
like Rackspace Onica to take over the operational
cost of a piece of the infrastructure freeing up staff to focus
on development and innovation. We've actually also got
a set of customers that want to bring in domain
expertise over security and networking and so in some cases
we would see customers contracting with people like Presidio
in North America and Octave or a computer center
in EMEA, to bring in domain such over networking with Cisco
or over identity with Okta. So, we very much
encourage and observe that where customers
use multiple consulting partners perhaps of
different parts of the problem set, migrations can go 25% faster and in some cases they have
better business outcomes and greater
cost savings. It also brings skills from those
partners to train your own staff because your staff are going to be
upskilling and capturing new skills. Now, you may want to use partners
in different parts of the world and so we also work hard
to certify partners and I am pleased to say
that we now have over 130 partners in the APN network that are certified
migration competency partners. This is over 50%
from the end of 2019 and we’re actually
observing migrations going on right now in 42 countries
around the world where partners are engaged
all the way from large global integrators down
to local experts in a given country and we continue to add
those partners and we help them get trained
on particular domain areas that they can bring
to accelerate your migration. Now AWS Professional Services
has also built out skills and a common pattern
is that we work with a partner, so very often we will find
we’re engaged with a client where it’s AWS Professional Services
working in conjunction with one
or more multiple partners that you have already got
that you prefer, or that we will introduce
as a trusted partner, and we’re seeing that pattern
play out around the world and you may engage AWS
ourselves on particular domains. The first area is very much
around training and certification and we have a broad array of
training and certification services for you to upskill your teams
in particular domain areas that you want your teams
to become expert. You may actually ask us to come in
and bring in AWS Managed Services as was done
at Sallie Mae, or as another customer will be
talking in a minute, Thomson Reuters, and Managed Services teams
will take over control of some part of your infrastructure.
You might actually want to use us to project manage
the migration itself but not the application
modernization, or you will prefer
to just exclusively work with one of our partners. Now, interestingly, we’ve also put
Professional Services into Marketplace, as I mentioned earlier today,
and I am pleased to say that there are over a hundred
Professional Services listings now in Marketplace very much associated
with the implementation of software, where you can contract for services
right out of the Marketplace, and just over a year ago
we launched a service called ‘AWS IQ’ where you’re going to find
task-based contractors that will do small projects
very quickly over a particular domain area, and of course we continue
to expand our AWS support with our premium offer
being AWS Enterprise Support and we have thousands
of large entities around the world contracting for AWS
Enterprise Support. These are all mechanisms to augment
the skills of your staff and to help your staff
get up to speed so that you become autonomous, but contract out to us when we help
you modernize and transform faster. Let’s talk about Thomson Reuters,
Thomson Reuters is a major information
provider around the world. Justin Wright is a guy
that I have been working with for several years,
he’s the VP of Cloud Operations working in a very large
Thomson Reuters fleet where they are already migrating
hundreds of applications and over 12,000 servers
onto AWS. They have refreshed
their security posture, they have worked closely
with one of our top partners, Trend Micro, and where
they have contracted in AWS Managed Services to manage
a number of those applications freeing up
the Thomson Reuters Team to concentrate on the innovation
of other applications that they want to deliver and the delivery of
their information services out to their thousands
of customers around the world. Justin continues to look at ways
to optimize his supply chain and his use of various
service providers in order to innovate on behalf of
his customers at Thomson Reuters. This is a good example
of a large global information provider modernizing their
application portfolio on AWS. So, in summary, taking you across
the idea of an application portfolio, the idea that you want
to have a very efficient and cost-efficient operating model, and the idea that you
only accomplish it really effectively using
professional services providers, let’s think about three things
I would love you to walk away with. Number one is prioritize
the applications that you’re going to transform
and differentiate your company. Out of your portfolio
what would we observe is that there may be five
or ten applications either in customer experience
in supply chain or in research
or in engineering, that really differentiate you
as a company, or in government it could be
over taxpayer experience and a particular service
being given to the consumer. Secondly, set up an operating model,
a service management framework, that will really allow you
to yield cost efficiency and increased levels
of automation on how you manage
your distributed teams, your distributed services, and so we can put in that right level
of observability and automation to reduce labor costs
and increase efficiency and resiliency of everything that
is running at your company. And finally, pick the best partner
to augment your skills, teach your people, and to take on some attribute
of your project to compress the time
in both migration and modernization so that you sooner realize
cost saving as you migrate your portfolio
to the AWS Cloud. Those are the key learnings
we give you from studying
over a thousand migrations and on executing migrations with thousands of companies
around the world. I hope you found this very useful
and I would like to finish by just pointing out that
in the next few days there are going to be
multiple sessions where I think you will find
additional follow-on knowledge. These are just some of
those sessions on this slide and I encourage you
to do follow ups and also to reach out
to the appropriate AWS teams that can follow up on this talk.
Thank you. [music playing]
