"Ultimately, we believe that advancement in cloud computing technology will be driven by open source initiatives where large communities of engineers can collaborate and develop new code for the new applications and demands posed by the cloud model," says Shelton Shugar, SVP Cloud Computing at Yahoo! - and upcoming Keynote Speaker at SYS-CON's 4th International Cloud Computing Expo, in this Exclusive Q&A with Cloud Computing Expo Conference Chair Jeremy Geelan.

Jeremy Geelan: What are the chief advantages of Cloud Computing from the point of view of Yahoo!’s customers worldwide?

Shelton Shugar: Yahoo! has more than 500 million unique users per month across the world. Yahoo! Cloud services enable us to provide superior user experiences and deliver targeted content to our enormous audience. Examples include faster content access around the globe, real-time sports updates, a personalized homepage experience, targeted news feeds, geo-specific ads and many more.

In addition, Yahoo! Cloud technologies enable us to innovate faster based on common, global and scalable platforms, thus enabling consumers to gain access to innovative features and products faster than ever before.

As one of the largest providers of consumer Internet services in the world, Yahoo’s cloud operates at virtually unprecedented scale, making it a unique environment and testing ground for cloud computing technologies.

Geelan: You have been quoted as saying that “Cloud is pushing up the Operational Excellence curve” – what exactly do you mean by that?

Shugar: Almost no other company can boast of having to tune its infrastructure to deal with the technical requirements and high standards of performance that are involved in serving more than 500 million unique users per month across the world. To meet this challenge, Yahoo!’s Cloud includes a collection of infrastructure and functional services targeted at dramatically improving the company’s efficiency throughout the entire product development cycle, from gathering user feedback and insight, to feature testing and iteration to ongoing product operations.

Cloud technologies allow us to achieve higher agility and quality while maintaining scale to meet the needs of our users. When I say “Cloud is pushing up the Operational Excellence (OE) Curve,” I mean simply that cloud enable our developers to focus more on creating great products for our users and less on the “heavy-lifting” of building complex infrastructure.

Geelan: As we approach the zettabyte (and perhaps even yottabyte) age, can value really be extracted from the voluminous data that is now in existence? How is that possible?

Shugar: Yes, it’s possible using technologies such as Hadoop. Hadoop is an open source distributed file system and parallel execution environment. Yahoo! runs Hadoop on tens of thousands of servers, enabling us to process and extract value from massive amounts of data. Apache Hadoop is an open source project of the Apache Software Foundation. Yahoo! is the largest contributor to Hadoop technology as well as its largest user.

Yahoo! recognized that the next-generation Web-scale services demand large distributed systems, and a growing number of other companies and organizations are likely to need similar capabilities. In addition to contributing most of the Hadoop code base, Yahoo! provides open source resources to the academic research community enabling them to access Internet-scale supercomputers for conducting systems and applications research.

Geelan: So exploring data-intensive computing in industry is clearly gaining momentum; what specific initiatives has Yahoo! been taking to encourage expertise in Hadoop and progress toward faster supercomputers? (Just a few, maybe – I know there are many!!)

Shugar: Earlier this year, Yahoo! announced the Yahoo! Distribution of Hadoop, in response to frequent requests from the community. Yahoo! is opening up its investment in Hadoop quality engineering via the Yahoo! Distribution of Hadoop, which has been tested and deployed at Yahoo! on the largest Hadoop clusters in the world and is based entirely on code available from Apache Hadoop. By making the Yahoo! Distribution of Hadoop generally available, Yahoo! is contributing back to the Apache Hadoop community so that the ecosystem can benefit from Yahoo!’s quality and scale investments.

In addition to Hadoop, Yahoo! is heavily investing in other cloud-related technologies such as storage, distributed caching, and serving to solve massive data-intensive computing challenges with serving our consumers.

Yahoo! also has very strong expertise in next generation cloud computing and data management technologies, and is leveraging its leadership in open source software, including Hadoop and Pig, to contribute to global, collaborative efforts around Internet-scale computing.

Over the past few years, we have made significant partnerships and have contributed technology to some of the leading research and development entities worldwide:

• November 2007: Deployment of a supercomputing-class data center, called M45, for cloud computing research (first deployed at Carnegie Mellon University).
• July 2008: Open Cirrus™ testbed formed with Hewlett Packard, Intel, the University of Illinois at Urbana-Champaign, the Infocomm Development Authority (IDA) in Singapore, and the Karlsruhe Institute of Technology (KIT) in Germany
• April 2009: University of California at Berkeley, Cornell University and the University of Massachusetts at Amherst join Carnegie Mellon University to take advantage of Yahoo!’s cloud computing resources
• June 2009: Participation at Open Cirrus Summit, three new sites to join Open Cirrus: the Russian Academy of Sciences, Korean Electronics and Telecommunications Research Institute (ETRI), and Malaysian Institute of Microelectronic Systems (MIMOS).

Geelan: When Yahoo! contributes to open source software such as Pig and Hadoop, what does the company get back?

Shugar: Yahoo! is a long-time supporter and contributor to open source software. Yahoo! is a platinum sponsor of The Apache Software Foundation, and the leading contributor to Hadoop to date, for example. Several members of Yahoo!'s development teams are active, long-term code contributors to Apache Hadoop, and we are committed to advancing the state-of-the-art in distributed computing through the incubation of new Apache projects.

We currently, and will continue to, actively collaborate with the industry, academia and the open source community, including through our Open Cirrus consortium, involvement with Hadoop, Pig and gradually other cloud-related technologies, and support of Apache. We are contributing back to the community so that the ecosystem can benefit from Yahoo!’s quality and scale investments, renowned technologists and innovation in next-generation Web technologies.

Yahoo! benefits from the contributions of others to open source projects and is able to collaborate with and sometimes hire the most talented researchers and engineers in the world, based on their interest in large scale application of open source technologies.

Geelan: And how does helping developers help Yahoo! in bottom line terms?

Shugar: We believe that the developer community is a key component in making Yahoo! a success. The challenges the industry is facing today in terms of large-scale, global cloud solutions are bigger than any one company (big or small) is able to solve on its own. As we contribute to the community, we also learn from the community, and third party developers are a valuable resource helping to speed innovation

Yahoo! is gradually open sourcing its cloud technology for this exact reason. We want to provide developers out there with an open source framework of scalable cloud technology that will enable them to build tools and solutions that in return will help Yahoo! (and the entire industry) to address these complex challenges.

In addition we believe this ecosystem of solutions “powered by” the core cloud technology will directly benefit Yahoo! consumers. This will provide consumers with customized web-applications (targeted at solving specific user needs) faster and better than ever before.

Geelan: What is it technically that makes Hadoop so powerful for large scale data processing? Are there alternatives?

Shugar: Hadoop provides the software infrastructure used across Yahoo! for large scale distributed computing and back-end data analytics, including fighting spam in Yahoo! Mail, content optimization for the Yahoo! homepage, and better ad targeting based on data analysis.

Hadoop provides a framework that distributes data and processing across thousands of computers, which working in parallel allows us to process and analyze enormous amounts of data in a very efficient manner.

We find Hadoop the ideal solution for Y! based on its scalability and flexible programming environment.

Geelan: How about enterprises, are they going to be adopters of Hadoop too?

Shugar: Yahoo! was the first major technology company to adopt Hadoop. Since then, many other companies, large and small, have begun using Hadoop. You can find more information about how others using Hadoop here: http://wiki.apache.org/hadoop/PoweredBy.

Geelan: What happens to Yahoo!’s Distribution of Hadoop and growth of the Apache Hadoop project overall in light of the deal with Microsoft?

Shugar: Hadoop is our de facto standard scalable data processing platform at Yahoo!. In addition to Search, Hadoop is used to support display advertizing, content platforms, personalization, machine learning for filtering email spam, research, and all other large scale data analysis and mining.

We remain as committed as ever to developing and using Hadoop, and in contributing our code to the open source community. The Yahoo! Hadoop development team will be busy delivering on short and long term roadmaps.

We look forward to continuing to work with the wider Hadoop community to build an increasingly better Hadoop that will support Yahoo! and the industry’s needs in this area. We have by far the largest team of developers and testers working on the project and hundreds of internal customers who use tens of thousands of computers in large Hadoop clusters. We plan to continue to work on improving the Hadoop core to make it faster, more scalable, reliable and secure.

Geelan: What exactly is the Yahoo! Cloud Serving Platform?

Shugar: For the time being, Yahoo!’s is focused on developing a “private cloud”, focused on making the Yahoo! experience as extraordinary, effective and productive as possible for consumers and advertisers across the world. We see this as a multi-year effort that will provide significant advantages for Yahoo! now and in the future.

The Yahoo! Cloud Serving Platform provides Yahoo! developers, engineering and operations with a technology to build, test, deploy, and manage application in an elastic cloud infrastructure that can grow and shrink based on changing workloads.

The cloud serving architecture is based on the following key concepts:

• Abstracts concerns of the underlying infrastructure and network communication from developers and operations

• Uses virtualization and commodity hardware

• Employs a declarative programming language for defining services and applications • Automates the deployment process

• Standardizes software stacks and packaging

The Yahoo! cloud serving architecture represents a shift in the programming paradigm and, as such, can require a fundamental shift in application-design thought processes.

Geelan: Are interoperability and openness important in Cloud Computing, as far as Yahoo! is concerned?

Shugar: Ultimately, we believe that advancement in cloud computing technology will be driven by open source initiatives where large communities of engineers can collaborate and develop new code for the new applications and demands posed by the cloud model. Yahoo! is a leader in supporting open and collaborative research and development as a competitive advantage, enabling the open source community to drive forward the pace of innovation. We believe open source and collaborative innovation is the way to address the complex challenges that large-scale cloud solution presents. Yahoo! will continue to participate and contribute to such efforts wherever appropriate.

Geelan: What other plans does Yahoo! have, moving forward, for helping deliver a cloud at the scale of the Internet?

Shugar: We are investing internally in further building out and deploying cloud computing technologies and services across the global Yahoo! operation so as to help our product teams innovate faster and deliver high-quality experiences to our customers across the globe.

Over time, we may consider exposing our cloud services in a more comprehensive manner through the Yahoo! Developer Network, which serves as Yahoo!’s front door for third parties seeking to engage with our developer tools and web services. However, we have nothing specific to share at this time.

Geelan: Lastly, 2009 has so far been a year of obvious challenges, from both a CapEx and an OpEx perspective, for anyone involved with Enterprise IT. So, what’s your top tip, as a seasoned software executive, to those other IT execs out there right now – especially CTOs of embattled start-ups who may be looking for some magic bullet to ensure they’re alive (and well) as a company in 2010?

Shugar: For startups and small companies, there are public cloud offerings coupled with other commercial cloud vendors that can provide a pay as you go infrastructure minimizing CapEx investments and reducing the risk in deploying new products. For larger or enterprise companies, a hybrid model maintaining existing compute infrastructure while leveraging public cloud offerings for less critical or experimental projects may provide more flexibility at reduced CapEx.

• Yahoo! to Keynote 4th Cloud Expo: Accelerating Innovation with Cloud Computing
• Live Demo of Yahoo! Query Language to be Given at 4th Cloud Computing Expo
• Yahoo! to Present at 4th International Cloud Computing Expo
• An Exclusive Interview with Oracle, Cloud Expo 2010 Diamond Sponsor