Information Technology | Networks » Erik Mitchell - Cloud computing and your library

web.techlib Cloud computing and your library By Erik T. Mitchell Technology proliferation The number of ways in which we access information is making our personal computing infrastructure very complex. My own personal computing infrastructure includes two personal computers, multiple data-subscription services, two external hard drives, a smartphone, and an MP3 player. While my information technology access is fragmenting, my need to access my information at any point or place is growing. This creates tension between how I manage information and its availability in my everyday life. I recently found my library has been dealing with these same issues as we reexamined how well our computing infrastructure supports our library services. As part of this discussion, we talked about cloud computing but were not sure how it fit in our computing infrastructure. Like many organizations, we had multiple local servers, each of which supported different applications. Implementing a new

application meant finding available resources on one of our machines and ensuring this new application would not conflict with our current environment. This was only one of a number of underlying technical issues that added a level of complexity to providing user services. We wanted a platform that offered us easy scalability along with redundancy and security but found that achieving this with traditional in-house servers was cost-prohibitive. We decided we needed an easier way to allocate technology resources, and we were left wondering if cloud computing might help fill that need. Cloud computing refers to a wide field which includes hosted applications, web-based client services, and a distributed, platformindependent server architecture. Our library was already a heavy user of hosted services and wanted to find out how a distributed server architecture would provide us with the infrastructure we needed for some of our core systems. Finding solutions that help you virtualize or

“go cloud” has gotten considerably easier in recent years. For example, implementing cloud-based data storage used to involve custom solutions and an in-depth understanding of computing infrastructure. In recent years, services such as Dropbox (http://dropbox.com) and MobileMe (http://www.mecom) have made this process as simple as saving your data to a local folder on your hard drive. What is particularly amazing about the proliferation of services like this for end-users is that studies are beginning to show how commonplace these sorts of services and computing paradigms are becoming. For example, the 2009 Educause Center for Applied Research ECAR study discussed the growing use of cloudbased services by undergraduate students (http://www.educauseedu/ecar), and the 2009 Horizon report (http://www.educauseedu/ELI/2009HorizonReport/163616) observed that cloud computing solutions are within a one-year implementation horizon. The next obvious question is, What does knowing that our

patrons are more comfortable with network and cloud-based solutions mean for how we approach providing services to these patrons? The valuable data is no longer owned One of the first big shifts in how libraries manage resources was the move from print-journal purchasing models to database-subscription and electronic-journal purchasing models. Libraries found this transition helped them scale their resources and provide better service just by thinking a bit differently about how they provided journals. Likewise, current cloud-computing initiatives allow technologists to think about how we handle our computing resources. Shifting to cloud solutions gives libraries an opportunity to save time and resources and also re-allocate resources to improve service. The kicker is that libraries do not really have an option any more. Distributed digital library models like the HathiTrust (http://www.hathitrustorg) are based on the idea that the really interesting stuff, the connections between

data, cannot be owned or kept in a single place. Likewise, our patrons’ use of computers has changed We have less control over their computing platform and have to provide services in a context in which our patrons have an expectation of seamless interoperability and appropriate contextualization of services. Cloud solutions allow a library to respond more quickly to service needs by allowing a library to scale its technology resources, employ a pay-on-demand resource model, and provide IT infrastructure on a subscription model that could be difficult to acquire and manage otherwise. Some of these services offload technical management responsibilities and even provide a level of data management for libraries. There are a host of management possibilities, from the newly formed DuraCloud project (http://www.duraspaceorg/duracloudphp) to offerings from commercial vendors But using cloud services means more than purchasing access to a specific system. A number of hosting companies

offer dedicated virtual platforms that enable libraries to completely control their computing environments. For more complex needs, companies such as Amazon (http://aws.amazoncom/ec2) provide infrastructure services as opposed to server hosting. Going cloud for your IT infrastructure Cloud computing is based on the simple concept that systems and data can be stored and accessed over the network using large, scalable servers to handle the heavy lifting. Cloud computing uses virtualization to separate software from hardware, provides centralized security along with high-speed internal communications, and serves advanced data management needs where client computing is impractical. This isn’t to say that there aren’t risks and downfalls with cloud computing. For example, network connectivity is not a guarantee, and we cannot always rely on being able to connect to our data. Likewise, agreeing to a cloud service places a lot of trust in a company or service with which you may feel

little affiliation. Finally, for some patrons and institutions, engaging in social activities in the cloud is rife with privacy concerns. Add to this the need for a new set of skills to implement and manage complex cloud services, and the whole thing may seem to be a bit too much. At its core, however, cloud computing allows libraries to re-examine their resource allocations. Finding your place in the cloud Making the decision to use cloud-based services means balancing the elements of cost, risk, and benefit to decide whether those services advance the mission of the library. When done correctly, though, cloud services offer the opportunity to increase functionality and work with our patrons in new and hopefully unexpected ways. When faced with the idea of moving from local solutions for our library data and applications to cloud-based solutions, we sat down and asked what cloud services we needed. Some vitale ones came to mind: file storage, archiving and preservation,

application hosting, and scalable production environments. Rather than coming up with a perfect plan, we decided to just try a few things. We worked with virtualization applications (e.g http://wwwvirtualboxorg) to get an understanding of how virtual computing environments worked. We launched development servers on Amazon’s EC2 service and began setting up new services in the cloud as opposed to on our local servers. Along the way we found many of these enterprise-level systems had individual-scale counterparts. Continuous Data Protection /file co-location services such as Dropbox make it simple to backup and replicate your data across multiple machines. Virtualization services such as VirtualBox, Google’s upcoming Chrome OS, and Web-based solutions like g.host (http://ghost) provide computing environments based on the idea that your computing platform is separate from the machine you work on. While the move required us to reconsider how we approach our data storage and application

management and how we think about technology as a subscription service as opposed to infrastructure, it is also positioning us to be able to respond to patron needs more quickly by scaling and adapting our information systems to meet their needs. What I have taken away from the process so far is that while cloud computing adds a level of complexity to our computing environments, that complexity comes with a new set of tools and services. These tools help us deploy services faster and on a much larger scale than we could do otherwise. Further, community-focused applications allow libraries to build a service together rather than working on projects in isolation. I expect the challenging part of this will be less about technical hurdles and more about figuring out how to re-shape our library and provide services to better fit our patrons needs