If you haven't had the chance to read The 451 Group's recently released
report on Grid adoption in the pharmaceutical industry, don't panic.
GRIDtoday editor Derrick Harris spoke with Steve Wallage, director of
research at The 451 Group, about it. Although he doesn't give away all
the information contained within the report's pages, Wallage does offer
great insights into who's using Grids, why some are not and what
vendors can do to get a foothold in this potentially large Grid market.
What is the overall conclusion of the report? What does the adoption of Grid in the pharmaceutical market look like?
After speaking with several pharmaceutical and
biotech firms about their current and future strategies for deploying
Grid into the enterprise, The 451 Group found that some of the leading
pharmaceutical companies have established significant Grid deployments
for drug discovery, but for a majority of them the jury is still out
regarding the broader value of Grid computing. A need for
high-performance computing makes the pharmaceutical sector a natural
adopter of Grids, but we found that compared with other industries,
such as financial services, the pharmaceutical sector is behind when it
comes to broader deployments.
Some pharmaceutical firms have made great strides in Grid adoption,
moving quickly from small Grids, usually based around cycle scavenging,
to thousands of devices. The leaders -- like Johnson & Johnson and
Novartis, for example -- are now looking at sharing infrastructure
across their companies and putting enterprise IT applications on Grids.
But the majority of pharmaceutical firms hear the hype around the early
adopters and have yet to identify a compelling reason to adopt Grid
Bottom line: Despite all the claims made around the benefits of Grid deployment, there has yet to be a drug discovered this way.
Clearly, some pharmaceutical companies are unconvinced of
the benefits Grid computing can bring. What are some of the main
reasons for this skepticism? What are some of the major obstacles to
For those pharmaceutical firms that have not yet
invested in Grid technology there are a number of reasons given. The
top reason is that there are no compelling applications, meaning the
company is not convinced of the benefits of running virtual docking
algorithms on a Grid. Another reason is that the current distributed
computing architectures work fine. For example, AstraZeneca was an
early user of clusters, and it sees no reason to move to Grid
technology. In addition, many pharma companies are reluctant to run
applications across the PCs of researchers and the greater penetration
of laptops was a particular problem.
The biggest obstacle to the pharmaceutical firms has been software
licensing. The pharmaceutical firms have looked at various options to
address this challenge, from license management to legal action. The
ultimate issue is how much pressure the pharmaceutical firms can put on
the ISVs in terms of licensing. The situation will improve, but it will
probably take another 12 months before more Grid-friendly pricing is
Another major obstacle to deployment is cultural issues, since those
who make decisions about Grids are often not high up enough in the
enterprise decision-making hierarchy, or they run autonomous R&D
units. Other obstacles include the challenge of federated data
applications and semantic integration. Integrating data that resides in
multiple database sources as a result of M&A or the nature of
research is a huge headache in this sector. Linking disparate offices
is another challenge.
What are some of the main drivers of adoption?
Although some pharmaceutical firms had invested in
high-performance computing technology for target identification, others
have relied on the PCs of researchers. Most companies have used
clusters to increase performance, and linking PCs and servers in a Grid
was a natural next step for some firms.
Overall, we've identified five main reasons for deploying a Grid:
- For some early Grid adopters the hardware has come from dedicated
server farms. For pharmaceutical firms Grid usage has mostly been about
cycle scavenging. It has been about increasing utilization on existing
- Pharmaceutical companies see Grid technology as an opportunity to reduce hardware costs.
- A particular weakness for some pharmaceutical firms' R&D
operations has been the lack of sharing of information and analysis.
Deploying Grid technology across the enterprise would increase
cooperation and communication.
- As with other early Grid users, many of the pharmaceutical
companies claim the ability to do things quicker, and also the ability
to do more complex analyses.
- Pharmaceutical companies have been overwhelmed with Grid success
stories and ISVs looking for them to try out the technology. In such a
competitive and global industry as the pharmaceutical market, none of
the major companies wants to risk missing out on a potentially
transformational technology. This means even those companies that have
yet to invest in a Grid deployment have at least carefully considered
such an investment.
What companies are currently the "shining stars" of Grid
adoption within the pharmaceutical market? Are there any companies with
unique deployments that might be making news in the future?
Early adopters such as Johnson & Johnson and
Novartis are well ahead of other companies, and are likely to remain so
as they seek to adopt Grid technology across all parts of their
business. Johnson & Johnson stands out in terms of its vision of
Grid technology and the breadth of usage. Meanwhile, Novartis is
leading in terms of its PC Grid.
GlaxoSmithKline (GSK) is currently doing some interesting work on data
analysis and clinical modeling. Pfizer is as well. Sanofi-Aventis is
leading in possible usage of semantic Web work.
If pharmaceutical companies are all aware of Grid computing,
and still many aren't convinced of its value, what can vendors do to
get them on board?
Grid computing vendors face two challenges within the
pharmaceutical sector: supporting those users which have established a
beachhead for achieving enterprise-wide adoption, and making a better
business case for Grid computing technology to the mainstream users.
Instead of educating users, vendors need to have strong business
arguments to convince them to move forward.
Vendors also need to be aware of some of the obstacles that
pharmaceutical companies are facing with deploying Grid technology.
Vendors need to be aware of the pain caused by drug discovery software
licensing practices, and they need to work with the pharmaceutical ISVs
and users to find ways around this issue.
With cultural issues being very complex and far-reaching in the
pharmaceutical world, it would benefit vendors to be sympathetic to the
needs of individual researchers and to look at best practices on
gaining widespread employee support. Many vendors have done very well
by selling to certain individuals -- typically the head of IT for an
R&D unit. This approach will only have limited success in the long
term, though, and it has clear operational risks. Vendors need to reach
out to central enterprise IT in order to gain wide visibility and
support across other departments.
What Grid vendors are doing the best in selling their wares
to in the pharmaceutical industry? How are smaller companies like
United Devices and Platform performing versus giants like IBM and HP?
Overall, the smaller Grid middleware players are the
most associated with Grid deployments in the pharmaceutical sector.
United Devices is particularly strong and has done a good job of
securing sales and awareness among pharmaceutical firms. The company is
the main provider to Johnson & Johnson and Novartis, two of the
leading pharmaceutical companies in the market and the leaders in
deploying Grids across their enterprise. Vendors need to be clear on
how to work with, or against, United Devices.
Some other key players include IBM, who is often involved as a systems
integrator and is also a close partner to United Devices. Platform
Computing also has a number of deployments, while vendors such as
DataSynapse and Altair are not yet on the radar.
It's my understanding that companies in all vertical markets
often cite the difficulty of Grid-enabling applications as a major
detractor in making the switch to Grid infrastructures. If more ISVs
start Grid-enabling popular pharmaceutical apps, will we see a
significant rise in adoption?
It's true that the lack of Grid-enabled applications
remains a problem not only in the pharmaceutical sector, but also for
Grid computing more generally. In all, about 30 or so ISV applications
in this sector are Grid-enabled. The tipping point to a procurement
decision is how easily an application can be Grid-enabled, how many
applications can be Grid-enabled, and what is the corresponding cost of
Grid-enablement versus buying more hardware. Keep in mind that there is
also a disconnect between what vendors claim and what users actually
see as Grid-enabled.
As far as it causing a significant rise in adoption -- it would
certainly help. But software licensing issues also need to be addressed
at the same time.
I find it peculiar that Globus is not very popular among the
companies profiled, especially considering how popular it is in
academic/research settings. Why do you think this is?
We've found that there are some enterprise IT end users
who have looked at the Globus Toolkit, but we have yet to come across
any end users who have implemented it in a production environment. From
users, we hear that they find Globus complex. The iteration is beyond
them, and they don't know what to do with it. The immediate opportunity
is for someone to define a stack. Enterprise IT end users say that
there are too many Grid middleware stacks and that they would be happy
to work with just one moving forward.
To some extent, pharmaceutical firms have chosen Linux and Oracle as de
facto standards, but they have not settled on a standard Grid "stack"
as such -- so there is an opportunity for vendors and open source
projects such as Globus to drive the development of this stack.
If pharmaceutical companies know that the major industry
players like Pfizer, GSK and J&J are strong proponents of Grid
computing, why haven't the others invested more heavily in order to
reduce the competitive advantage these companies are likely
Most pharmaceutical firms have at least assessed Grid
computing. All the major pharmaceutical firms continually evaluate
Grids but the obstacles we mentioned continue to deter users.
Something very relevant to the potential use of Grid technology is the
setup of many pharmaceutical firms' R&D operations. They tend to be
autonomous units for two main reasons. One is that multiple units often
exist as the result of M&A activity or to harness the strengths of
a particular location, such as proximity to a university research
department. The pharmaceutical firms want to maintain these units'
autonomy in order to keep their effectiveness. The second reason is
that the current thinking in the pharmaceutical world is that small
R&D facilities are good. For example, GSK is setting up small
Centers of Excellence for Drug Discovery, since it believes these
groups, which can operate independently of central control, are far
more likely to be successful.
It is also relevant to look at the research staff. The lack of IT
skills among researchers can be seen in the fact that many get along
very nicely in target identification and other drug discovery work
without using any sort of computer model, let alone a Grid. For
example, Eli Lilly reports that many of its researchers prefer to use
tried-and-tested procedures -- "wet research" -- that don't involve any
Where does the pharmaceutical industry stack up against other vertical markets in terms of Grid adoption?
We found that compared with other vertical markets,
such as financial services, the pharmaceutical sector is behind in
terms of Grid adoption, especially in the depth and scope of adoption.
In the financial services market, investment banks often have a clear
roadmap with implementing service-oriented architectures as the
ultimate goal. In the pharmaceutical market there are limited
deployments, and typically R&D department-led Grid deployments,
which limit Grid usage to a few drug discovery applications. Other
verticals, with a broader and enterprise-focused view of Grids, are
expected to expand their usage over the pharmaceutical users.
Finally, I'm wondering how what you foresee for the future
of Grids in the pharmaceutical industry assuming there are no drug
discoveries directly linked to Grid. How would a concrete example of a
Grid-based drug discovery change things?
Pharmaceutical companies are intensely competitive, but
at this point, there is still a perception that it is safe to hold off
on Grids or run very limited deployments. A Grid-based drug discovery
or other compelling evidence of the benefits of Grid would certainly
change this perception.
Is there anything else you would like to add?
The main issue for many users is whether to spend on
software like United Devices, or to get the software "free" and spend
on hardware. Life sciences companies will increasingly be turning to
open source technology as discovery pipeline budgets are throttled
back. Most smaller pharmaceutical firms are already using open source
technology in some fashion -- either running Linux or running open
For more information about this report, please visit www.the451group.com/intake/gridtoday-jun05
About Steve Wallage
Steve Wallage is the director of research at New York-based The 451
Group -- an independent technology industry analyst company focused on
the business of enterprise IT innovation. For more information about
the company, please visit its Web site at www.the451group.com