Cloud Computing Evolution: Disrupting the IT Supply Chain
Chris Weitz, Director, Technology Strategy and Architecture, Deloitte Consulting LLP
John La Bouff, Senior Manager, Deloitte Consulting LLP
It's the dirty little secret few want to talk about in the information
technology (IT) industry. As cloud computing scales upward
across consumer and enterprise systems, it is inevitably raising
disruptive implications for the future of industry incumbents,
specifically those in the semiconductor industry.
Cloud computing essentially is a paradigm or model of
computing for delivering on-demand, location-independent,
dynamically scalable and often virtualized computing resources as
a service over the Internet. Network access with cloud computing
is ubiquitous, rapidly elastic and generally performed on a pay-per-use
model. Whether this occurs through internal private clouds or
external vendor clouds, it represents a huge architectural shift—and
it's going gangbusters. A 2009 Gartner report estimated the U.S.
cloud computing market to reach $3.5 billion in 2008 and forecasted
its compound annual growth rate (CAGR) to be 40 percent between
2008 and 2013, generating an estimated $18.6 billion in 2013.
Truly, this is where a lot of new growth is occurring in the IT
marketplace, and IT suppliers are pushing related innovation efforts.
Cloud computing is dramatically changing the way IT is consumed,
packaged and sold. For IT vendors and their suppliers, cloud
computing is the new basis of competition, placing tremendous
pressure on incumbents and traditional relationships with suppliers,
including the semiconductor industry. Within a few years, the
semiconductor industry structure, including the concentration of
sectors and its leaders, may be significantly transformed. With new
buyers emerging and long-standing demand sources disappearing,
many traditional suppliers risk displacement.
Figure 1. Infrastructure IT Spending on Cloud Utility Models
Will
Increase Exponentially over the Next Decade
Source: Deloitte Consulting LLP
Semiconductor industry incumbents that cannot establish a
position in the cloud computing supply model risk being pushed into
shrinking "pre-cloud" sectors and rapid obsolescence as enterprises
shift from purchasing computer equipment—the "undifferentiated
heavy lifting" of traditional IT architectures—and begin using
commodity services provided by new cloud service providers. And
there are other implications. Because the cloud's emergence is
propelled by demonstrated efficiencies and economies of scale in
various computing service deliveries, it represents a net compression
of the amount of hardware required to deliver a given level of service.
All things being equal, that means fewer enterprise buyers, reduced
demand for silicon and deflationary price pressure.
For the semiconductor industry, uncertainty lies in the emerging
range of new cloud computing services and the level of demand
that will result. New possibilities—provisioned supply of enterprise
software, outsourced enterprise architecture, computing modes
such as real-time rendering of games and video in the cloud and
the convergence of higher speed broadband with high-definition
television (HDTV), home telepresence and 3-D—raise questions
about both total market demand and what they portend in terms of
semiconductor needs. The cloud and its efficiencies could stimulate
such new demand for computing services that overall markets and
demand for processing power expand. Or it could put a lid on
that demand by sharply raising the efficiency gains of computing
infrastructure. The future is unknown, but whatever the scenario, the
landscape will be profoundly changed.
Cloud Computing 101
Cloud computing is a significant change from traditional in-house
IT systems. Where currently the typical application system must
use a massive dedicated infrastructure that includes everything
from software to security, cloud computing essentially liberates
applications from a fixed infrastructure. Right now it's settling into
three distinct service types:
- Software-as-a-Service (SaaS): On-demand use of software over
the Internet or private networks. To date, this is the path of
greatest adoption, which makes sense since it's been on the
market for several years. Businesses are flocking to this service,
transitioning from licensing software to buying it on a per-use
basis.
- Platform-as-a-Service (PaaS): Tools and environments to build
and operate cloud applications and services. PaaS is a young
player in the marketplace and has more operational challenges,
so there will be a longer path to adoption. But adoption will
potentially grow over the next few years as the time to scope,
develop or procure, test and deploy internal applications is
reduced. The need to converge applications on a single platform
will further drive adoption.
- Infrastructure-as-a-Service (IaaS): Storage, network and
operations resources as a service from the cloud. The driving
forces to adoption of IaaS are lower IT costs, increased flexibility
and the ability to rapidly start up and shut down services. It also
helps moderate carbon footprints since IaaS reduces the need to
set up an on-premise data center infrastructure.
Regardless of the service type, cloud computing tends to possess
five major qualities: on-demand service, ubiquitous network access,
location-independent resource pooling, rapid elasticity and pay-per-use. None of the service types previously detailed can deliver
all these qualities yet, so it's a mixed bag in the marketplace right
now. However, given today's economic pressures, cloud computing is
clearly becoming attractive to the enterprise for several reasons.
First, thanks to pay-as-you-go models, costs being treated
as operating expenses and reduced hardware purchases, cloud
computing can help reduce IT capital equipment spending. Second,
companies using clouds can complete projects in less time and
achieve accelerated time to benefits, enabling greater flexibility and
agility. Third, the use of cloud technologies allows businesses to
reallocate staff resources from IT to other core activities.
Today, cloud computing is being delivered in the marketplace through
either a public cloud, in which multiple customers can
get access through the Internet or a private network; a private
cloud used internally by an enterprise and controlled within an
enterprise; a hybrid cloud of both public and private cloud services,
including "virtual private clouds"; or a community cloud used across
organizations for collaboration.
A Disruptive Force for the Enterprise Market
With all these benefit opportunities and a clear path to adoption,
pressure is being placed on incumbents and IT industry leaders that
may put their existing sales channels at risk. The reality is that the
products currently implemented in the enterprise as well as within
service providers are no longer as appropriate to a cloud delivery model.
For instance, new products from many equipment suppliers are
intended to be cloud-enabled or cloud-specific. For engineers, this
means that the primary functions previously assigned to a device—whether it's a network switch or a server that focuses on processing
and interfaces with storage—are now being shared among device
classes. And the distinctions are no longer as relevant as they've been.
This will upset the product design, marketing and selling logic for
many industry incumbents.
Figure 2. Accelerating Adoption of Cloud Utility Models
Will Flatten Conventional Infrastructure Concepts
- Virtualized cloud architectures will exploit automated provisioning infrastructure
sharing ("multi-tenancy") and capacity elasticity.
- Reduced provisioning lead times—weeks/months to days/hours.
- Simplified virtualization permits much simpler planning (e.g., large-grained
APIs for application architects to access infrastructure).
- Potential order of magnitude efficiency improvements.
Source: Deloitte Consulting LLP
Enterprise buyers are also changing because they are modifying the
way they consume services, as previously mentioned. Buyers are no longer
as concerned with the long-term viability of specific devices as they are
with the service itself. Basically, they're reallocating their resources from a
hardware acquisition model to a service acquisition model.
For these reasons, industry analysts looking at the overall IT
landscape are anticipating deflationary pressures on enterprise IT
hardware suppliers. They see buying power moving from relatively
disempowered enterprises with little leverage to very powerful,
concentrated service providers with leverage.
The Consumer Side: Semiconductor Dislocation or Opportunity?
The disruption takes a slightly different form in the consumer market,
particularly where semiconductors are concerned. Consumers are
still going to buy devices, but some of what makes up the "guts" of
the devices today will become virtualized and move into the cloud,
reducing the required computing power of traditional PCs and other
devices. So central processing units (CPUs) and graphic processing
units (GPUs) will differentiate based on which master they serve. The
portion going to the cloud will have a different yardstick for
ideal features, packaging and configuration compared to today.
Consumers will derive benefits from devices such as cell phones and
laptops that won't need as much horsepower. And battery life will last
longer if video is rendered on a connection in a server farm rather
than on the device.
On the other hand, inside the cloud, densely packed clusters of
servers and storage may turn out to be optimized by a different power/performance/form factor model. The high virtualization of cloud
computing opens the door for new packaging and more integrated
server architecture as it scales. Someday this may be the predominant
footprint for chips, rather than having it go into standalone servers,
desktops, laptops and various mobile devices.
An example of one company addressing this is Advanced Micro
Devices (AMD), which announced with partners OTOY and Super
Micro Computer in March 2010 that it would bring its AMD Fusion
Render Cloud products to market in Q2 2010. Designed to deliver
thousands of concurrent HD games, remote desktops and live HD
video streams to any Internet-enabled device, the new product enables
server-side rendering of fully interactive HD content, taking the power,
processing and storage pressure off the actual consumer device.
The Future of the Industry
Clearly, the cloud is not only going to change a semiconductor
company's customer base, but also how they satisfy customer demand.
The race to create the fastest processor may be rendered moot with
cloud computing. Instead, the victory may be creating a processor
that conserves power and enables the integrated architecture features
required for cloud fabrics at a low cost.
Those businesses that will be best positioned to adapt will likely be
those that have adjusted their portfolios and provided cloud computing
service providers with the products they need. They'll understand what
the cloud looks like, enabling them to better understand their future
customers. And they'll shape go-to-market strategies around what will
be a shifting set of assumptions about the future marketplace.
Intel is moving in this direction. In a story published by Forbes in
January 2010, Intel said that they see the move to cloud computing
as "one of the biggest opportunities in computing history." They
view the transition to cloud computing as a boon to potential new
sales. But their future customer isn't necessarily the traditional
end-user enterprise, but instead cloud computing service providers
with enormous data centers filled with homogenous technology
architectures. Of course, the downside is that there will be less
demand for their long-time bread-and-butter products—servers and
server chips. As demand drops, prices drop and, consequently, lower
the market profit levels for server chips.
Intel is by no means the only one eying the cloud computing
market's potential. With a level playing field offered by a new
computing model, it's almost a throwback to the early days of
technology competition. Chip companies can sell directly to cloud
service providers, cutting out the "middleman" of server companies.
This offers chip companies the opportunity to make specialized cloud
chips and move the point of differentiation to the cloud instead of
enterprise servers.
For those in the consumer market, the opportunities and
challenges are as great. As with those who have long served the
enterprise market, those chipmakers who have supplied consumer
devices have long focused on speed as their point of differentiation.
But if much of the capacity and processing issues are transferred to
the cloud, chipmakers will need to create products that better suit the
new needs of consumer devices.
In anticipation of this major shift, there are a few key strategies
that semiconductor manufacturers and suppliers should consider
implementing:
- Recognize and anticipate significant changes in demand levels
for enterprise computing, with cloud service providers moving
into position as the dominant buyers in the marketplace for
computing infrastructure.
- Work with the growing needs of cloud computing service
providers to adapt product requirements, and focus, in
particular, on the density and high virtualization of cloud
computing architectures.
- Shape the product set designed for end-users and consumers
around the needs for remote accessing and processing from
cloud computing service providers, not intensive "end-node"
computing.
With the increasing adoption of cloud computing, those
semiconductor businesses that have the foresight to identify and act
on the inevitable changes coming to their industry will be better
positioned to profitably meet customers' needs in the future. The
cloud doesn't have to signal a threatening storm, but it is gaining
momentum as a force to be reckoned with—and it's headed in your
direction.
About the Authors
Chris Weitz is a director of technology strategy and architecture at Deloitte
Consulting LLP, based in San Jose, California. He has over 25 years of
experience and works with leading technology companies around the world in
the areas of IT strategy, architecture, integration and operations, with a focus
on IT infrastructure. Chris currently leads Deloitte's cloud computing practice,
working with innovative technology vendors and service providers to bring value
to clients through new cloud computing architectures and services models. Chris
is a frequent speaker and panelist at technology industry events, focusing on
innovation in technology services. He can be reached at cweitz@deloitte.com or
408-704-2825.
John La Bouff is a 30-year veteran of the semiconductor industry, and was among
the first to develop supply chain practices and systems designed around the fabless
model. John is now a senior manager in Deloitte Consulting LLP's Silicon Valley
practice, where he focuses on operational performance and product development
improvement for semiconductor companies. You can reach John La Bouff at
jlabouff@deloitte.com or 650-450-6056.
This article contains general information only and is based on the experiences
and research of Deloitte practitioners. Deloitte is not, by means of this article,
rendering business, financial, investment, or other professional advice or services.
This article is not a substitute for such professional advice or services, nor should it
be used as a basis for any decision or action that may affect your business. Before
making any decision or taking any action that may affect your business, you
should consult a qualified professional advisor. Deloitte, its affiliates and related
entities shall not be responsible for any loss sustained by any person who relies on
this article.
As used in this article, "Deloitte" means Deloitte Consulting LLP, a subsidiary of
Deloitte LLP. Please see www.deloitte.com/us/about for a detailed description of
the legal structure of Deloitte LLP and its subsidiaries.
Back to Articles Home
