At Flybridge, we
periodically dive deep into areas in which we see significant market growth and
investment opportunity. Recently, we did so in the whole field of Cloud
Computing and thought it would be interesting to share some of our thoughts.
I would also like to thank our great summer associate, Ravi Inukonda, who
worked with me over the last few months on this effort.
If you have been
following trends in enterprise IT as of late, you have been unable to miss
Cloud Computing and all the bad puns that normally accompany such commentary (“Forecast
Mostly Sunny for Company Opting for Cloud Computing” or “Future bright for
cloud computing” being two such examples).
While the ideas are
not new – IBM called it Utility Computing back in 2002 – the Cloud Computing
trend really gained steam over the past couple of years. The momentum is
more than marketing, as the advent of Cloud Computing has really been enabled
by advances in server/CPU architectures, storage and networking subsystems,
open source software adoption and most importantly virtualization, all of which
were perhaps too nascent to truly enable the cloud until recently. In
addition to the technical advances, the current economic environment has
furthered these trends, as customers look for ways to reduce capital
expenditures and increase the efficiency of their IT operations. As a
result, we now believe that Cloud Computing is at a tipping point and that it
represents a fundamental architectural shift that will create numerous
opportunities up and down the enterprise IT landscape. Past tipping points
include the shift to client server architectures, enabled by more powerful PCs
and widespread adoption of LANs, and the shift to Internet based three tier
enterprise applications, driven primarily by the adoption of the web.
Given that Cloud
Computing is used so frequently as an adjective, it is worth a quick segue to
define the term and architecture. Simply put, cloud computing is
infrastructure or applications that are delivered as a massively scalable
service that is elastic in its ability to quickly scale up and down, where the
complexity is abstracted from the end user, that are paid for in some form of a
by the drink utility model. Taking this down a level, we think about
clouds along two dimensions.
- From an
architectural approach, cloud computing services tend to be Infrastructure as a
Service, for example Amazon's EC2 compute platform, but also
including other compute, networking and storage infrastructure; Platforms as a
Service, for example Salesforce.com's Force.com platform, that allow
developers to rapidly build and deploy clound-based applications; or
Applications as a Service (SaaS), such as Netsuite's ERP and accounting
- The other dimension
is whether these services are offered to customers as public clouds by third
party organizations, private clouds in an enterprise's own data-centers or as a
hybrid of the two.
From Flybridge’s lens as an early stage venture capital
investor, we believe that starting a company focused on public clouds offering
infrastructure as a service is a fool's errand. In that field, the keys
to success are massive scale, a low cost of capital and operational excellence.
These capabilities are in abundance at places like Amazon, Google and
Microsoft, but not at start-ups. Further, we think that general purpose
Platforms as a Service are also difficult for start-ups given the need to drive
developer adoption and the long gestation cycle as the platform is built, users
recruited, applications deployed and ultimately scaled. The exception to
this rule of thumb will be companies that successfully latch onto existing
developer communities, such as what Engine Yard has done for the Ruby
on Rails community. Finally, we also believe that many of the horizontal
SaaS application opportunities have been played out. Whether it is
Salesforce for CRM, NetSuite for ERP or Workday for HR, most if not all of the
obvious markets have credible, large, well-funded companies with market leading
positions offering solutions.
Does that mean that all the promising
opportunities are already covered or are there big white space opportunities
for new companies? We think there are a few trends and themes that offer
great promise, each one of which can support numerous successful companies:
- Enable enterprise data centers to be more cloud like. CIOs of large
companies that we speak to are generally concerned about control, security and
delivering a good, reliable service to their end-users. The first two
issues lead to a generic concern with seeing any of their mission critical or
sensitive applications hosted in a public cloud. That said, the mere
existence of services such as Amazon's Web Services puts pressure on their
ability to deliver a good service to their end-users as they fundamentally
change expectations. If a user can have a full server OS, storage and
networking provisioned on AWS in minutes, the current enterprise service level
of weeks or months to provision the same looks pretty weak. As a result,
enterprise CIOs that are responsible for data centers will over time look to
manage and operate these data centers much in the same way as a public cloud.
This will, therefore, create opportunities around management,
provisioning, billing, access control and other technologies that allow a
private cloud to be effectively operated. As an aside, there will also be
an opportunity at the networking layer as the deployment of cloud architectures
in enterprise IT data centers will create the need for high speed, low latency,
low power, non-blocking network switches.
- Bridge the cloud
and enterprise boundary. The advent of public cloud infrastructure
will create opportunities for companies that bridge the divide and smooth the
transition between private and public clouds. While there is a risk that
the Platform as a Service vendors will view this as their domain, witness
Amazon's recent VPC announcement there are many services such as backup
and email archiving that benefit from a combination of on-premise and cloud
based software that will be outside of the domain of the platform providers.
The other approach that we see promise with along this front are
applications and platforms that support cloud-bursting, in other words an
application that runs most of the time in an enterprise data center but under
heavy load relies on the public cloud for scalability.
- Provide tools
to better manage cloud environments. Any time there is a fundamental
architectural shift in the enterprise IT landscape, it has created
opportunities for new companies to move more quickly into the space than
incumbent vendors with installed bases to protect and legacy applications to
support can. These companies have generally focused on systems
management, network management, security and the like. This trend is
already underway in the cloud market and can been seen in the success of
companies such as RightScale and New Relic. In addition to
tools for the management of the cloud, we also believe that new infrastructure
services, such as horizontally scalable databases written solely for cloud
applications, will see great adoption as more applications are written from the
outset to be run in the cloud.
- Develop SaaS
applications for vertical markets. While there may not be much whitespace in
major horizontal application markets, there are many vertical market niches
that can support the creation of very large companies. In these markets,
a start-up can build deep domain knowledge around the problem set and offer its
applications in a cost-effective manner (often leveraging public clouds for
their infrastructure), making their solutions easier to adopt, maintain and extend
as compared to traditional enterprise applications in the same field. We
are seeing this with the success of our insurance SaaS company, FirstBest
Systems, but there are numerous other examples in the healthcare IT,
bio-medical research, energy and other fields.
- Create companies
that are enabled by the Cloud. One of the aspects of cloud computing
that excites us the most is that it is a huge enabler of innovation.
Cloud computing makes it so easy to develop and deploy applications that
it levels the playing field for start-ups versus large, well capitalized,
competitors. As an entrepreneur, I can now buy access to world-class data
center infrastructure for dollars an hour that would have previously cost me
millions to purchase and deploy. As a result, savvy start-up executives
can look to create companies that would have previously been too capital
intensive to contemplate. This is especially true if the computing need
of such applications, for example software testing (see SOASTA for one such
provider) or financial modeling are intermittent for any one customer but
relatively smooth across many such customers.
Given the benefits
of a more flexible, dynamic and efficient IT environment, Cloud Computing,
under this name or new ones to come, will see significant adoption across the
enterprise IT market over the next several years. This will create many
opportunities for savvy entrepreneurs, so if you have ideas or thoughts, please
let us know!