Recently, I got a chance to read the Bernstein research report titled “The
Long View: Netbooks, Wireless and Cloud Computing – Client Software’s Imperfect
Storm?”. Larry Dignan of ZDNet has already written about it along with his
take. I was somewhat skeptical about the report and its conclusion when I read
Larry’s post and it was confirmed further when I actually read the report. The
authors of the report have done extensive research compiling quite a bit of
data. But, I have my own reservations on their assumptions and analysis. Let me
take the relevant portions of their report and offer my own opinions on them.
Before I pick on the report findings, let me offer some perspective on Cloud
Computing. Unlike some of the other technological evolutions, Cloud Computing is
actually a paradigm shift from the way we used software in the traditional
desktop and on-premise world. It needs an entirely new thinking and mindset.
Whenever such paradigm shift occurs in the technology world, we could see fear,
skepticism, reluctance, etc. towards its adaption. The shift from status quo to
new technology also happens suddenly because once the critical mass is reached,
the adaption becomes more widespread and it doesn’t follow any linear trend.
Many times, the growth in the market becomes almost exponential leaving the
pundits wondering how they went wrong.
This report is trying to make predictions about a paradigm shift based on the
current market trends and, as Ed Cone correctly
pointed out, the future doesn’t end with 2012. Since the enterprises are
always slow to adapt a paradigm shift, expecting a complete embrace in such a
short time (compared to the eon-scale timeline of enterprise technological
adoptions) is too much to ask. Let me now pick from their report and offer my
take on them.
- The report says they still see the future of data processing as
heterogeneous both in the clouds and in the device. Well, they are not
alone in such predictions about hybrid nature of enterprise computing in the
future. People assume that the hybrid computing is the end of the evolution in
enterprise computing. Rather, it is an intermediate step in the evolution of the
enterprise from the traditional world to a Clouds based world. An intermediate
point in the evolutionary process cannot be construed as the end point of the
process. Given enough time and with maturation of cloud based technologies, the
dinosaur-like enterprises will eventually move to Cloud. I would prefer the
analysts to see a long term picture than making predictions based on short term
- They quote the lack of complete internet coverage (whether wired or
wireless) in many geographies and, also, talk about users need to have offline
apps and data for travel in the airplanes, etc.. This, again, fits my argument
against short term view while making predictions. It is just a matter of time
before we get the ubiquitous internet. It may or may not happen in the time
period they are considering but it is definitely going to happen. While making a
case against the netbooks and its emergence along with Cloud Computing, they
also point out that most of the applications benefit greatly from local caching,
processing and rendering, especially if the host side has less than total
availability. Then they quote that a Blackberry device is still valuable when
disconnected from a wireless network because a user can still read cached emails
and generate responses offline that will be sent and synchronized when the
device reconnects with an implication that Cloud Computing lacks any such
features and hence of lesser utility value. They are wrong again. Google has
released Gears technology as open source. It is now used in many SaaS apps
including Google Apps, Zoho Suite and, also, by smaller players like
Mindmeister. In fact, they do mention the presence of Gears technology at the
later part of their report but they never connected it with the above argument.
If they had actually made the connection, they wouldn’t have made this argument
in the first place.
- They are using the results of Pew Internet and American Life project. They
argue that consumer usage of online apps is technically greater than enterprise
usage and is an important mechanism by which users become familiar with online
applications in their workplace. There is no problem with this argument but the
Pew research they quote includes people with the age of 65+ in the calculation
of the national average in SaaS usage. If this average was taken as a data
point in their analysis, I would say that their analysis has some flawed
assumptions. People over 65 rarely play a major role in shaping the enterprise
IT usage. Under such circumstances, including them while calculating the average
and then using it to extrapolate the trends is not a correct approach.
- Their report terms the impact of Cloud Computing on Software and OS
development as moderately disruptive in the next 5 years. They quote the annual
revenues of Amazon in 2012, through their Cloud services, to be 50 Million and
the annual revenue of Google, through its PaaS and SaaS offerings, to be a very
moderate 1.5 Billion, as one of their arguments to minimize the disruptive
potential of Cloud Computing. Such smaller revenues aren’t surprising at all
because it is the main selling point of SaaS, in particular, and Cloud
Computing, in general. Instead of paying bloated amounts for software like
Microsoft Office, users can save tons of cash by using SaaS based productivity
apps. Also, unlike the case of Microsoft Office, the consumer side of SaaS
productivity apps like Google Docs or Zoho Apps are free. They only charge for
the premier edition, usually used by medium sized companies and enterprises.
This, along with the fact the SaaS applications are designed to be cheaper, will
contribute to the heavily reduced revenues. This reduction is an unavoidable
aspect of the software market in the future and it is, in fact, an advantage for
SaaS. While making any comparison between Cloud based companies with the
traditional software companies having comparable products, it is not right to
make an argument just by comparing their respective revenues. The competitive
advantage of the cloud over the traditional software is, actually, the huge cost
savings it offers. It is just not possible any cloud vendor to earn comparable
earnings while offering such heavy cost savings.
There are some glaring mistakes too.
- Though I wouldn’t keep this against them, I want to point out a trivial
mistake in their report. While discussing about Netbooks, they talk about the
cost of Linux license being $5. This is plain wrong. Linux licenses doesn’t cost
a dime but support services might cost money. Even though this doesn’t go
against their analysis, it will be better to avoid such glaring mistakes.
- They also claim that Google locks up users into their platform by using
Python. First, Python is an open source language. Second, Google App Engine SDK
is also open source. Finally, Google App Engine is already
ported to Amazon EC2. Their argument about vendor lock-in is plain
But I am neutral about their quote on the lack of flexibility and
customization and being locked into an ongoing cost with a single cloud
provider, as a disadvantage against cloud proliferation. This is true right now.
But it is changing fast with more and more vendors opening up and adapting data
portability and interoperability. Such moves will ensure that clients are not
struck with a single vendor.
However, the report also correctly points out the trends we see in the
- They do not deny the increased use of clouds in the consumer segment, small
and medium business segments and, even, enterprise segments.
- They are right about their prediction that there will be increase in the
SaaS revenues and enterprises moving to clouds, even if it is on a hybrid level.
They also mention that SaaS will have substantial savings on the software costs.
- They also correctly identify that SaaS is the fastest
growing sector in the software market.
- They are right on target when they identify Microsoft as the biggest loser
due to the current cloud push by various vendors.
I am disappointed with the short sightedness of this report. I wish the
authors had spoken more with the Cloud vendors and analysts before coming to
these conclusions. As I pointed out in the beginning of this post, paradigm
shifts cannot be predicted by the current market trends alone.
Related articles by
a Service: The many faces of the cloud
Computing” buzz — what is it ?
computing to the max
Networks Launches: Google App Engine For Java
cloud computing: The big picture
Computing Challengers: Pundits Cheer for Their Clients
computing not the answer for every busines