An e-book that explores factors beyond basic price that impact the true cost of ownership for cloud implementations by comparing and contrasting prevalent approaches, using common enterprise workloads, from three prominent cloud service providers.
2. Cloud is evolving, becoming the key platform for
innovation and business value. Yet, as enterprises look
to the cloud to accelerate growth, many still struggle
to understand the true cost of ownership.
Enterprises turn to the cloud to…
What do they hope to gain?
Achieve deployment flexibility...
...to gain interoperability across
hybrid environments
...to leverage emerging ecosystems
and communities
Improve quality and speed
time-to-market...
...to roll out innovative business
models quickly
...to deliver compelling
customer experiences
Leverage existing investments...
...to integrate cloud with current
IT infrastructure and applications
...to deliver an integrated user experience
Quickly gain deeper insight from data...
...to drive real-time decision making using
data from internal and external sources
...to increase productivity, collaboration
and quality of customer relationships
Achieve greater understanding
and learning...
...to identify new business value from
“dark” data
...to improve customer targeting
and personalization
Innovation to maximize opportunity and
business flexibility
Speed to get to value faster
Insight to get the best business outcomes
Security to minimize risk
Flexibility to best deliver on business
objectives
Access to technology without major
capital investment
Transformation to deploy new
business models
Choice to get the best price-performance
Cost to optimize IT spending
Visibility and control to manage the cloud
more efficiently
To seize the advantage that cloud promises, forward-thinking
business leaders are looking for solutions that enable strategic
capabilities at the lowest cost. And providers are responding,
with aggressive pricing that makes their cloud offerings look very
affordable. But are they?
To get an accurate understanding of cloud costs and their
business impact, it’s necessary to look at real-world costs
and performance.
This guide can help you in your comparison and evaluation
of competitive cloud offerings. Using what’s here, you can look
beyond initial pricing to focus on real-world performance, price-
performance and the many considerations that drive total cost
of ownership.
3. 3Cloud IT economics What you don’t know about TCO can hurt you
In the real world,
cost of ownership
is not that simple
“ Some businesses that chose a cloud service provider
based on low published rates are experiencing higher
costs than anticipated. Others are fielding complaints
about unacceptable application performance. Still others
are realizing they have no idea whether they are paying
too much or too little for what they are getting.”
– Frost & Sullivan, The Truth About Cloud Price-Performance1
The total cost is not always evident
Determining the true cost of cloud is more complex than it might seem. By
focusing on initial pricing, enterprises may miss important performance factors
and service add-ons that increase overall cost. An attractive base price is far
less appealing when additional investment is needed to accelerate business
results. The real cost of a cloud that speeds innovation, enables compelling
customer experiences and supports business flexibility is the more meaning-
ful measure.
There’s an important difference between total cost of acquisition (TCA) and
total cost of ownership (TCO). The first is what the enterprise actually pays for
cloud services. TCO is harder to pin down because it also includes intangible
costs like complexity that drives up IT staffing costs, or poor performance that
impacts customer experiences.
Answering the total cost question takes a deep understanding of how the cloud
affects the enterprise, and exactly what’s being purchased. That’s why there’s
more to making a smart cloud decision than choosing the lowest entry price.
When looking at cloud you might ask yourself:
What capabilities will accelerate speed, innovation and
business results?
__ What are the real requirements for applications, workloads,
and service levels?
__ Can the cloud integrate well with conventional IT, today and
in the future?
__ Can the provider meet your requirements for security
and compliance?
What level of performance is needed to drive compelling
customer experiences?
__ Can the provider’s cloud deliver the speed and throughput
that individual workloads require?
__ Are choices available that deliver higher levels of
performance and service?
What are the real economics of the solution?
__ How much will it cost to achieve the needed performance–
initially, and in the future?
__ If upgrades are needed, what will they cost?
__ How much does it cost to migrate workloads and data?
__ Are there hidden costs?
4. What are some of the factors that drive total costs?
Technology
__ Processor and RAM costs
__ Storage and memory costs
__ Connectivity and bandwidth costs
__ Upgrades and scalability to meet high
workload demands
Data centers
__ Number and geographic location
__ Transfer charges between data centers
Software and operating environment
__ Operating system costs and
associated overhead
__ Software licensing
Technical support
__ Cost of internal resources to support
the environment
__ Provider technical support
Management
__ Visibility, control and transparency
__ Complexity
Development
__ Workload deployment efficiency
__ Platform complexity
Integration and workload portability
__ Ability to integrate current and
future applications
__ Interoperability among systems
5. 5
It’s time to think differently
about cost
“ Selecting a cloud service requires an in-depth
understanding of workload performance characteristics
and needs, and the ability to match those needs to the
actual offerings of multiple cloud vendors.”
– Frost & Sullivan, The Truth About Cloud Price-Performance2
In cost of ownership, details make the difference
What might seem the least expensive cloud offering could cost more in the end
because it doesn’t deliver the performance or capabilities that the enterprise
actually needs “out of the box.” The key cost-related question is how well the
cloud performs in the context of real workloads and business requirements. It’s
not just price, but price-performance that matters–and comparisons should
take every cost driver into account.
The challenge with assessing the true cost of cloud is that providers don’t
structure, price, or measure performance consistently. The full picture involves
much more than selecting CPU, storage and networking. Similar specifications
from different providers can result in very different levels of performance and
cost, and it can be next to impossible to make direct comparisons on paper.
To make an accurate comparison it’s necessary to look at key performance
metrics, infrastructure configurations, contractual agreements and the impact
of workloads on cost.
Why similar offerings
don’t cost the same
Environment–Not all clouds are equal,
even if they look alike up front. Multi-tenant,
single-tenant and dedicated server clouds
all perform differently.
Pricing–Not all providers offer pay-for-
what-you-need pricing, and upgrades
like current-generation processors and
permanent storage might cost extra.
Hidden costs–There may be unexpected
charges for intra-cloud network traffic, or
service and support.
Contract conditions–Volume or term com-
mitments can make it costly to move off of
the provider’s cloud.
6. 6Cloud IT economics What you don’t know about TCO can hurt you
Looking at workloads
realistically
“The ‘right’ cloud service isn’t always public cloud,
or even virtualized cloud; the optimal solution may
involve bare metal or dedicated servers, or even a
private or hybrid environment.”
– Frost & Sullivan, The Truth About Cloud Price-Performance3
Leveling the playing field: cost in context
To make cost comparisons more meaningful, IBM has conducted extensive
analyses of cloud configurations from different providers based on the real-
world scenarios and business conditions enterprises are likely to encounter.
The findings, verified by leading analyst firm Frost & Sullivan, are a comparison
of price-performance in context rather than straight pricing.
What’s revealing about these tests is the effect of factors like choice of cloud
environment and workload requirements on cost. The results show that “apples
to apples” comparisons don’t necessarily yield an accurate view of cost, high-
lighting the need to assess the cost of ownership for each workload individually.
Compare the most meaningful measures
Analyzing typical scenarios can be a useful starting point, but to get a solid grasp of
TCO it’s important to understand the exact characteristics of your specific workload.
Web applications Compute-intensive, customer-facing applications that require rapid response, throughput
and scalability.
What to look at
• How many user requests are processed per second on average?
• Do alternative environments deliver better price-performance?
Analytics Storage-intensive, traditional business analytics and innovative cognitive applications.
What to look at
• How many input-output queries per hour can the cloud handle?
• How costly is storage?
Network-intensive workloads Inter-application messaging and data localization that require data to be moved
cloud-to-cloud, cloud to on-premises systems and from one data center to another.
What to look at
• How much cost does a messaging-intensive workload add?
• How cost-efficient is the cloud at moving data and workloads?
Hosted private cloud Moving workloads from on-premises to consolidated cloud environments with speed
and efficiency.
What to compare
How much does it cost to migrate a virtual machine to the cloud?
7. 7Cloud IT economics What you don’t know about TCO can hurt you
TCO in the real world:
Compute-intensive cloud
application workloads
Cloud-based applications like those for the Web,
mobile and the Internet of Things are compute-
intensive, demanding a cloud with flexibility and raw
performance to deliver the experiences custom-
ers expect. It’s about being able to quickly deliver
and manage applications that meet the needs of
the business as enterprises continuously develop
applications in a mixed environment.
Some application workloads require the rapid scal-
ability and provisioning of virtualized infrastructure,
while others will achieve better price-performance
in a non-virtualized bare metal environment where
the enterprise has greater control. This places a
premium on the ability to choose the environment
best suited to each, and easily shift among them.
The comparison:
Environment makes a difference
IBM benchmarked cloud-based application
workloads against a pair of leading providers
in two ways. One set of tests was run in a fully
virtualized environment, using the most nearly
comparable configuration from each provider.
Another was run using the IBM SoftLayer®
bare
metal (non-virtualized) cloud.
One competitor’s virtualized offering was very
close to IBM SoftLayer: a single-tenant virtualized
cloud using the same generation of processors.
However, the other provider offers only multi-tenant
cloud, and neither competitor offers the additional
performance and control of bare metal servers.
For compared workload and cloud environments, IBM SoftLayer
is faster, lower cost, and has better price-performance
IBM SoftLayer virtualized vs. Provider A
23%faster performance with IBM
25%lower price with IBM (3 yr. total)
1.6xbetter price-performance with IBM
IBM SoftLayer, virtualized vs. Provider B
32%faster performance with IBM
1.2xbetter price-performance with IBM
IBM SoftLayer bare metal vs. Provider A
35%faster performance with IBM
47%lower price with IBM (3 yr. total)
2.5xbetter price-performance with IBM
IBM SoftLayer bare metal vs. Provider B
44%faster performance with IBM
20%lower price with IBM (3 yr. total)
1.8xbetter price-performance with IBM
Based on an IBM internal study of a basic web application workload. Customer applications, differences in the stack deployed, data
center locations and options, and other systems variations or testing conditions may produce different results. Pricing based on
published US list prices available for IBM and competitor data center location specified as of 3/21/2016 and calculated by projecting
costs measured over a 3 year TCA period. Costs include compute, network, storage, software, data transfer, and support charges
(and excludes labor). Available options and pricing vary by data center location. Option for non-virtualized servers not available in
competitor environment.
8. 8Cloud IT economics What you don’t know about TCO can hurt you
The implications for TCO
The analysis results show the real benefit of the choices that only IBM SoftLayer
offers. The processing overhead of virtualization and proprietary platforms has
a marked impact on performance. That can make it more difficult to deliver the
compelling experiences that customers demand.
There were also additional charges that added to the total cost of ownership
for competing cloud providers. Higher-cost, production-capable VMs were
required to match IBM SoftLayer, and there were additional costs such
as software licensing, data transfer and technical support. All of these added
dramatically to the total cost–and none are reflected in the competitors’ basic
price quotes.
The IBM advantage for cloud application workloads
Business need What should the cloud provide? How IBM delivers Key benefits can include
Stable, consistent high
performance with lower
operating costs
Choice of environment, with
pricing by usage or by the hour
and no upcharges
No fixed, predefined
configurations in IBM SoftLayer;
options like IBM POWER®
and
IBM z SystemsTM
Match cloud capabilities with
workload requirements for
optimal price-performance
Continuous development
and delivery
Flexible tools that allow the
enterprise to choose its
own course
Automated, orchestrated
deployment with
IBM UrbanCodeTM
Deploy
Reduced time to market,
increased productivity that
lowers cost
Rapid deployment and scale Easy workload and
data portability
Seamless integration of
development across public,
private and dedicated clouds
with IBM Bluemix®
Local
Business agility,
responsiveness and
optimized cost
Innovation Integration with legacy
and external systems, data
and services
Robust IBM Bluemix service
catalog and open APIs, plus
IBM Cloud Managed Services
that bring new services to
market up to 75% faster6
Business agility and improved
customer experiences
Saving time and money
by deploying in the cloud
Amica reduces software deployment times
by 95-98 percent. IBM UrbanCode Deploy
helps automate software release processes.4
Learn more >
Coca-Cola Enterprises Inc.: Moving to the
cloud lowers IT costs and brings customer
engagement programs to market faster.5
Learn more >
9. 9Cloud IT economics What you don’t know about TCO can hurt you
120,000
100,000
60,000
80,000
40,000
20,000
0
16KIOPS
Provider A, block storage,
40K committed IOPS
1.6TB (10 x 160GB x 4000 IOPS RAID0)
SoftLayer, bare metal
internal SSD
1.6TB (8 x 400GB SSD RAID10)
16K_SeqRead
112,419
29,572
IBM SoftLayer
3.8xthe storage performance
3.0xlower price
vs. Provider A for compared workload
and environment.
Increasing the value of data:
Storage-intensive
analytics workloads
The revolution in analytics and cognitive business solutions is driving growth
in cloud adoption, as enterprises seek to leverage data beyond traditional busi-
ness intelligence. Data has become a key resource, enabling richer customer
experiences and powering new, innovative business models through new tech-
nologies like IBM Watson. Not only can analytics help to create better customer
experiences, they can optimize the use of the cloud itself by monitoring usage
patterns and enabling a flexible infrastructure that self-adjusts as needed.
Analytic and cognitive workloads require speed and low latency while support-
ing many concurrent users. This places the emphasis on storage performance
and low processing overhead, so that rated speeds can be achieved in the face
of tens or even hundreds of thousands of requests per second.
The comparison: Analytics relies on storage throughput
In the analytics realm, provisioning for high input-output performance is the
dominant cost driver. IBM tested a mix of complex and intermediate business
intelligence reporting workloads against two industry leading cloud providers,
using configurations optimized for reporting throughput.
The IBM SoftLayer test was run in a bare-metal environment for best
price-performance. Since neither competitor offers bare metal, the closest con-
figuration was used with upgrades as needed to deliver maximum throughput.Test results obtained using an industry standard file system benchmarking tool. Competitor
cloud configuration provisioned for I/O bandwidth of 40,000 IOPS. Results obtained under
controlled conditions testing resources of a publicly available cloud infrastructure. Customer
applications, differences in the stack deployed, differences in data center locations and
options, and other systems variations or testing conditions may produce different results.
For compared workload and cloud environments, IBM SoftLayer bare metal delivers higher throughput with less cost
IBM SoftLayer bare metal vs. Provider A
1.9xmore throughput per processor with IBM
3.0xlower price with IBM (3 yr. total)
3.2xbetter price-performance with IBM
US list prices available for IBM and competitor data center location
specified as of 3/21/2016 and calculated by projecting costs measured
over a 3 year TCA period, and include compute, network, storage, and
support charges (excludes labor, DB2 software license costs (identical
in both environments), and cloud provider data transfer charges.)
Available options and pricing vary by data center location. Option for
non-virtualized servers not available in competitor environment.
IBM internal study of an analytics workload consisting of reports
scanning all or portions of a 1TB denormalized star-schema database.
Competitor Cloud configuration provisioned for I/O bandwidth of 40,000
IOPS. Results obtained under controlled conditions using publicly
available cloud infrastructure. Customer applications, differences in the
stack deployed, differences in data center locations and options, and
other variations may produce different results. Pricing based on published
10. 10Cloud IT economics What you don’t know about TCO can hurt you
The implications for TCO
With some cloud providers, provisioning for high IOPS performance can
make costs skyrocket. Because of virtualization overhead and throttling by the
provider, standard storage offerings may exhibit high latency and be unable to
deliver their rated speed. The alternative is to opt for upgraded storage at a far
higher cost.
Even basic storage costs are not always readily apparent. A provider’s virtual
machines may include local storage–a seeming cost advantage. However,
that storage is temporary. Those “ephemeral” disks are erased each time the
virtual machine is shut down, restarted, moved or resized. With some providers,
permanent block storage must be purchased separately, and is available only
in fixed disk sizes–leading the enterprise to potentially buy more than it needs.
The IBM advantage for analytics workloads
Business need What should the cloud provide? How IBM delivers Key benefits can include
Enhanced customer experiences
and business model innovation
Access to open APIs that build
predictive analytics, business
intelligence and cognition into
the cloud
Extensive service catalog
including IBM WatsonTM
technology and IBM Bluemix-
integrated services
Differentiated offerings that
increase competitiveness and
business agility
Actionable business and
customer insights
Ready access to data
and analytics
Robust analytic solutions
for marketing, finance, risk
management and sales
performance, fully integrated
with the cloud
Accelerated business results
and optimized operations
Reduced costs for
storage-intensive workloads
Consistent, high-performance,
cost-effective storage
Massively scalable
IBM CleverSafe®
object-oriented
storage and IBM SoftLayer bare
metal deployments that optimize
infrastructure resources
High throughput and
data portability with lower
storage costs
Bringing cognition
to the cloud
To see how Whirlpool Corporation is
applying IBM Watson Services, including
cognitive analytics.7
Learn more >
Alpha Modus, in conjunction with
IBM Bluemix, is succeeding in making
sense of–and shaking up–the entire
investment management industry.8
Learn more >
11. 11Cloud IT economics What you don’t know about TCO can hurt you
Moving data:
Network-intensive
workloads
By definition, cloud strategies depend heavily on network performance. Data
is in constant motion, whether as part of application messaging traffic, work-
load migration, or data transfer from one data center location to another. Fast,
reliable network connections unlock value by enabling different systems and
applications to operate seamlessly, and enterprise data to be extended to the
cloud. The priority with these network-centric workloads should be to keep the
cost of moving data as low as possible.
As cloud strategies evolve, the need for high network throughput is increasing.
Cloud-based applications rely on network speed to perform well. Shortened
development cycles require workloads to be shifted rapidly, and often. There
is also the need to localize data, as security concerns and data sovereignty
laws–regulations mandating where data and workloads are physically housed–
become more stringent.
Connecting data and applications
to transform the business
For compared workload and cloud environments, IBM SoftLayer costs less, yields faster messaging
IBM SoftLayer bare metal vs. Provider B
Up to 7xfaster performance with IBM
Up to 11xbetter price-performance with IBM (3 yr. total)
IBM SoftLayer bare metal vs. Provider A
36%faster performance with IBM
2.4xbetter price-performance with IBM (3 yr. total)
Based on an IBM internal study of a message passing workload. Results
obtained under controlled conditions using a publicly available cloud
infrastructure. Customer applications, differences in the stack deployed,
differences in data center locations and options, and other variations
may produce different results. Pricing based on published US list prices
available for SoftLayer Washington DC and competitors’ US East data
center locations as of 3/21/2016 and calculated by projecting costs
measured over a 3 year TCA period that includes compute, network,
storage, and support charges (and excludes client labor costs and cloud
provider data transfer charges.) Available options and pricing vary by
data center location. Option for non-virtualized servers not available in
competitor environment.
IBM Cloud and consulting services help
Shop Direct react to shifting customer
demands in a mobile connected, cloud
enabled world.9
Learn more >
Apptimate migrated its solution to
the SoftLayer platform, capitalizing on the
global network of data centers to facilitate
growth and adhere to national data security
laws.10
Learn more >
12. 12Cloud IT economics What you don’t know about TCO can hurt you
80,000
60,000
40,000
20,000
0
Messagespersecond
Provider B
Haswell vCPU
Provider B
Sandy Bridge vCPU
9,514
12,632
25
20
15
10
5
0
Dollarspermessagepersecond
Provider B
Haswell vCPU
$20.87
Performance for the workload and
environments compared
Higher is better
Provider A
Ivy Bridge vCPU
12kB non-persistent messaging tests
51,995
Cost for the workload and
environments compared
Lower is better
Provider A Ivy Bridge: 32 vCPU, 10 Gbps Provider B Sandy Bridge: 16 vCPU Provider B Haswell: 16 vCPUIBM SoftLayer Ivy Bridge: Bare metal, 12 cores, 10 Gbps
SoftLayer
Ivy Bridge bare metal
Provider B
Sandy Bridge vCPU
Provider A
Ivy Bridge vCPU
$1.85
$19.39
$4.44
SoftLayer
Ivy Bridge bare metal
70,925
The comparison: Virtualization has a significant impact
IBM analyzed messaging performance against two leading cloud providers by
generating message threads with a “sender” application, passing them through
a messaging server, and consuming them with a “receiver” application. Traffic
was increased until the message rate couldn’t go any higher.
The processing overhead caused by virtualization and proprietary software
stacks played an important role, with IBM SoftLayer bare metal delivering far
higher performance at a small fraction of the cost relative to the competitors’
virtualized clouds. The choice of processor in the competitor’s clouds also
had an impact, with upgraded hardware delivering better performance but
at a much higher cost per message.
Based on an IBM internal study of a message passing workload. Results obtained under controlled conditions using a publicly available cloud infrastructure. Customer
applications, differences in the stack deployed, differences in data center locations and options, and other variations may produce different results. Pricing based
on published US list prices available for SoftLayer Washington DC and competitors’ US East data center locations as of 3/21/2016 and calculated by projecting costs
measured over a 3 year TCA period that includes compute, network, storage, and support charges (and excludes client labor costs and cloud provider data transfer
charges.) Available options and pricing vary by data center location. Option for non-virtualized servers not available in competitor environment.
13. 13Cloud IT economics What you don’t know about TCO can hurt you
The implications for TCO
For applications that generate heavy network traffic, messaging server
throughput can be the single most significant limiting factor on performance.
As network traffic backs up, application servers may sit idle and costs rise
because the enterprise must invest in additional capacity to compensate.
Network capacity is not the only cost driver associated with moving data. Most
cloud providers encourage their clients to keep data and workloads in the pro-
vider’s cloud through policies that make it easy and inexpensive to move data
to the cloud, but costly to shift it once it’s there.
The IBM advantage for network-intensive workloads
Business need What should the cloud provide? How IBM delivers Key benefits can include
Application performance High sustained
messaging throughput
IBM SoftLayer bare metal servers
and high-speed networking, with
pricing that allows the enterprise
to buy only what’s needed
Improved customer
experiences and productivity
Data localization High-speed, low-cost, secure
data transfer, large number of
data center locations
Zero transfer costs between
IBM SoftLayer data centers,
worldwide data center presence
with IBM SoftLayer and
IBM Cloud Managed ServicesTM
Customer satisfaction,
enhanced compliance,
reduced business risk
Data portability Network throughput, simplicity,
freedom from barriers
to migration
IBM Aspera®
’s dedicated
protocol that delivers transfer
speeds up to hundreds of
times faster than FTP or HTTP,
regardless of file size
Faster speed to market,
increased business agility,
enhanced productivity
Creation of innovative services Ability to connect applications
and data in a secure,
scalable way
IBM Message Connect
seamlessly links applications
and services, while
IBM Integration Bus connects
processes across systems
Better customer experiences,
business agility and speed
to market
Providers may also charge extra to move data to a different data center,
especially one in another region. These transfer charges can become signifi-
cant if the enterprise needs to move large amounts of data to specific locations
in order to meet regulatory compliance or customer mandates.
14. 14Cloud IT economics What you don’t know about TCO can hurt you
Gaining business agility:
Moving workloads
to the cloud
“When server resources are shared, there is increased
opportunity for data to become contaminated or lost.”
– Frost & Sullivan, All Clouds Are Not Created Equal11
Private clouds have an important place in today’s IT landscape because they
offer greater visibility, control and security. Some enterprises elect to keep
sensitive workloads such as development on-premises to ensure all three.
Today, however, providers like IBM offer hosted private clouds that combine
these advantages with the cost benefits of infrastructure-as-a-service. That
can enhance business agility, because the private cloud can be fully inte-
grated with existing IT infrastructure, on-premises and off-premises clouds in
a hybrid environment.
An increasing number of enterprises are deploying hybrid cloud environments,
enabling them to allocate workloads according to the characteristics of each.
Often, an enterprise will develop on-premises or in a private cloud, and then
migrate the workload to the public cloud for deployment. This makes workload
portability, complexity and the cost of migration key considerations.
The comparison: Simplicity and control tip the balance
IBM assessed the cost of ownership by consolidating 38 virtual machines
running Web applications, migrating them from an on-premises environ-
ment to the cloud. IBM SoftLayer costs were based on a hosted, bare-metal
private cloud. Neither of the two competitors offers a hosted private cloud
option, so pricing for those offerings was based on a fully virtualized public
cloud infrastructure.
For compared workload and cloud environments, IBM SoftLayer lowers migration costs
IBM SoftLayer bare metal vs. Provider B
40%lower cost with IBM (3 yr. total)
IBM SoftLayer bare metal vs. Provider A
44%lower cost with IBM (3 yr. total)
Based on an IBM internal study of a web application consolidation
scenario. Results obtained under controlled conditions using publicly
available cloud infrastructure. Customer applications, differences in the
stack deployed, differences in data center locations and options, and
other systems variations or testing conditions may produce different
results. Pricing based on published US list prices available for IBM and
competitor data center location specified as of 3/21/2016 and calculated
by projecting costs measured over a fixed testing period to a 3 year TCA
period. Costs include compute, storage, network, software, data transfer,
and support charges (and excludes labor). Available options and pricing
vary by data center location. Option for non-virtualized servers not
available in competitor environment.
15. 15Cloud IT economics What you don’t know about TCO can hurt you
The indirect costs of
moving virtual machines
Consider other elements of cost beyond compute and storage
Because bare metal offers higher performance then vCPUs, less software must be licensed. Less software, less cost.
Comparable 8 processor
single-tenant Linux systems
3 year total cost line item
IBM Softlayer, bare metal
(two servers)
IBM SoftLayer, private VM
(two single-tenant VMs)
Provider A, 3-year prepay
(two single-tenant VMs)
Compute + storage $51,840 $34,746 $89,300
Technical support $0 $0 $8,995 (Business)
Data transfer (500GB per month
per server)
$0 $1,620 $3,237
Software cost (including S&S) $49,686 $141,960 $141,960
Total 3 year cost with software $101,526 $178,326 $243,492
Total 3 year cost excluding software $51,840 $36,366 $101,532
3 year total cost projection includes comparable compute, storage, network, software, data transfer, and support charges; pricing based on SoftLayer monthly and
competitor’s 3 year all upfront reserved instance published US list prices as of 3/21/2016. Options and pricing vary by data center location. Competitor single-tenant
pricing includes dedicated per region fees. Options and pricing based on configurations in SoftLayer Washington DC and competitor’s US-East data center locations.
Data transfer charges assume 500GB transfer to internet per month per server for two servers for 3 years.
Migrating to a global cloud to
expand the reach of the enterprise
To offer more agility and speed,
Kasbah System Software relies on high
value IBM SoftLayer cloud capabilities to
host its Genius University Management
System.12
Learn more >
__ Hypervisor conversion
__ Complexity due to
separate manage-
ment interfaces
__ Need to learn new tools
__ Reduced flexibility
and control
__ Speed of provisioning
__ Risk exposure in the
public cloud
16. 16Cloud IT economics What you don’t know about TCO can hurt you
The implications for TCO
The way cloud providers price their cloud infrastructure has a major impact
on the cost to consolidate workloads. Charges are typically for fixed vir-
tual machine (VM) configurations rather than by utilization of the underlying
hardware, and the enterprise pays for capacity whether it is used or not.
IBM SoftLayer bare metal allows enterprises to achieve higher VM density
on a given physical server by “stacking” workloads, opening the door to
significant savings.
Indirect costs also come into play. Most cloud providers limit the choice of
computing architecture and hypervisor, and in some cases the management
platform. Each virtual machine that gets migrated from an on-premises data
center to the cloud may require conversion to a different hypervisor, and might
need different tools and skills for management. This adds complexity and with
it, cost.
The IBM advantage for consolidating workloads
Business need What should the cloud provide? How IBM delivers Key benefits can include
Security, visibility and control Physical, platform and
network security, consistent
management tools
Dedicated private cloud
infrastructure with choice of
hardware including IBM POWER,
IBM z Systems, with visibility
down to the machine
serial number
Reduced risk, enhanced
productivity, improved
business agility, lower cost
through higher VM density
Efficient workload migration Ability to extend existing tools,
processes and skills to the cloud
Open standards-based
IBM Bluemix platform, with
IBM SoftLayer support for
all hypervisors
Consistent management
interfaces, lower conversion
costs, less complexity, no
provider lock-in
Availability and data protection Ability to back up critical data
and recover workloads in a
timeframe that meets business
requirements
Disaster Recovery as a Service
and Backup as a Service
optimized for cloud workloads
Reduces risk of downtime
and data loss
17. 17Cloud IT economics What you don’t know about TCO can hurt you
To choose wisely,
look at the big picture
“ When the enterprise finds a provider that has
expertise across technology, infrastructure, professional
and managed services, it can make a selection that
is more likely to deliver maximum business value with
optimal ROI.”
– Frost & Sullivan, Using Hybrid Cloud Strategy to Drive Business Value13
On the surface, cloud seems so simple and affordable. But as has been
demonstrated in this guide, there’s more to it than low-cost virtual infrastruc-
ture. The initial price is only the starting point. The total cost depends on
everything from hardware price-performance to cloud architecture, pricing
models, software licensing, technical support, contracts and “soft” cost drivers
like platform complexity and provider lock-in.
The IBM approach is markedly different because it takes these additional
factors into account. While other cloud providers are working to limit and com-
moditize cloud services in pursuit of low initial prices, IBM adds real-world value
through openness and choice. When total cost of ownership is viewed through
that lens, the IBM advantage becomes clear.
For today’s enterprises, that distinct approach matters because business priori-
ties are fluid. As cloud strategies driven by cost reduction give way to the use of
cloud for strategic advantage, enterprises are adopting integrated hybrid envi-
ronments that combine public and private clouds with existing IT. Productivity,
business agility, flexibility and ease of integration are as important as low cost.
These are what IBM helps to achieve, through a combination of expertise, tech-
nology and services unmatched by any other cloud provider.
Base your business case on the real economics of IT
Which cloud is best? The answer depends
on your workloads, IT requirements and
business strategy. There are questions of
which workloads to place in the cloud, and
which environment best suits them. It’s a
choice that balances IT economics against
business agility and workload optimization.
The IBM Eagle Team IT Economics Practice
can run a no-cost analysis using your own
data and business assumptions to help you
make the best decision.
Learn more >
IBM Global Technology Services can
help you architect an infrastructure tuned
for cloud workloads, to optimize your
cloud ROI.
Learn more >
18. 18Cloud IT economics What you don’t know about TCO can hurt you
The IBM cloud aligns
with what enterprises
are looking for
Assessing cloud value is a balancing act, pitting what the solution enables the
enterprise to accomplish against the required investment. The IBM cloud can
help achieve desired business outcomes quickly, at low cost–the right combi-
nation of capabilities, performance and economics.
Capabilities
Innovation–Leverage open communities and
cognitive computing to maximize opportunity
Speed–Take advantage of hybrid environments to
change only what’s needed and get to value faster
Insight–Combine enterprise data with new
sources of data to get the best outcomes
Security–Minimize exposure of sensitive data
Performance
Flexibility–Position workloads to best deliver
on business objectives
Powerful computing–Access leading-edge
technology without major capital investment
Transformation–Build a fast, scalable foundation
for new, cognitive business models
Economics
Choice–Choose the mix of technologies
and deployment models that deliver the
best price-performance
Cost–Optimize IT spending for maximum ROI
Visibility and control–Manage infrastructure
and IT services more efficiently
19. The future of cloud is hybrid.
The future of business is cognitive.
IBM is making it affordable.
Through ongoing investments in open technologies and infrastructure,
IBM is laying the foundation of a new kind of cloud that delivers better capa-
bilities and higher performance along with improved economics. Reducing
total cost of ownership is essential–but so is enabling competitive advan-
tage. The IBM cloud is designed to do both.
IBM is enabling business innovation and the cognitive enterprise because
these are what will differentiate tomorrow’s businesses. Along the way,
there’s an understanding of business reality, and that there’s a balance to be
struck between investment and outcome. That’s why IBM focuses on the
total picture: those costs and capabilities that matter most to the enterprise.
The IBM cloud: Greater business agility.
Increased competitiveness. Less cost.
Choice with consistency Put the right workload in the right place, at the right time–because where and how you develop and
deploy workloads determines price-performance.
Hybrid integration Unlock existing data and applications to build on what you have today, changing only what you need to–
because less complexity means lower costs.
DevOps productivity Develop, experiment and iterate at speed–because business agility enhances cloud ROI.
Powerful, accessible data and analytics Extract deeper insight to understand your business and get closer to the customer–because smarter
decisions lead to greater efficiency.
Cognitive solutions Build understanding and learning into decisions and interactions–because that’s what’s going to drive
competitiveness in the age of digital business.