Library Hi Tech

Avoiding the death zone: choosing and running a library project in the cloud
Denis Galvin Mang Sun
Article information:
To cite this document:
Denis Galvin Mang Sun, (2012),"Avoiding the death zone: choosing and running a library project in the
cloud", Library Hi Tech, Vol. 30 Iss 3 pp. 418 - 427
Permanent link to this document:
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

http://dx.doi.org/10.1108/07378831211266564
Downloaded on: 07 January 2015, At: 12:23 (PT)
References: this document contains references to 14 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 886 times since 2012*
Users who downloaded this article also downloaded:
Michael Groenendyk, Riel Gallant, (2013),"3D printing and scanning at the Dalhousie University Libraries: a
pilot project", Library Hi Tech, Vol. 31 Iss 1 pp. 34-41 http://dx.doi.org/10.1108/07378831311303912
Judith Mavodza, (2013),"The impact of cloud computing on the future of academic library practices and
services", New Library World, Vol. 114 Iss 3/4 pp. 132-141 http://dx.doi.org/10.1108/03074801311304041
Nuria Lloret Romero, (2012),"“Cloud computing” in library automation: benefits and drawbacks", The
Bottom Line, Vol. 25 Iss 3 pp. 110-114 http://dx.doi.org/10.1108/08880451211276566

Access to this document was granted through an Emerald subscription provided by 191412 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for
Authors service information about how to choose which publication to write for and submission guidelines
are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as well as
providing an extensive range of online products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee
on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation.
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

*Related content and download information correct at time of download.

 The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0737-8831.htm

LHT
30,3 Avoiding the death zone: choosing
and running a library project in
the cloud
418
Denis Galvin and Mang Sun
Library IT, Rice University, Houston, Texas, USA
Received February 2012
Revised March 2012
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

Accepted May 2012 Abstract

Purpose – The purpose of this paper is to explore the evolving field of cloud computing and its
usefulness for library technology departments. It seeks to consider what types of projects are good
candidates for the cloud and which are not.
Design/methodology/approach – The authors spent a year trying out different projects using
Amazon’s Web Service. They discuss what went right and what went wrong. They brought up their own
machine image and tested out web sites and applications in the cloud. They discuss the different types of
cloud services, evaluating and choosing a provider and the types of projects which best fit into this
architecture. Advantages and considerations are highlighted as well as an alternative to the public cloud.
Findings – There are some projects that are better suited to cloud computing than others. Flexibility
and cost savings are the best reason for moving projects to the cloud. There are also good and valid
reasons not to move some projects off into the cloud.
Originality/value – This is an on-the-ground look at running projects in the cloud that used to be
done on back-end servers. The bulk of the paper looks at infrastructure as a service. All of the work
that has been done is in production and has been tested for over a year.
Keywords Cloud computing, Information technology, Omeka, Amazon Web Service, Archivists Toolkit,
Libraries, Private cloud, Linux, MySQL, Library systems, Worldwide web
Paper type Technical paper

Introduction
There are few terms as nebulous as cloud computing. It is defined in many different
ways, in many different papers, by many different people. This is because the term
cloud computing is used to define numerous different types of online environments
where computer operations are happening. Most of what people are referring to when
they say cloud computing is infrastructure as a service (IaaS), software as a service
(SaaS) and platform as a service (PaaS). These three environments make up the bulk of
the cloud computing services available.
Cloud computing can be differentiated from traditional data center or server room
computing because it is typically on demand and off premise. It is elastic in nature and
it is typically, but not always, maintained by a third party. It has been applied to other
services such as monitoring and communications, but most of what most people mean
when they say cloud computing is IaaS, PaaS and SaaS.
Probably the best definition is the most open-ended; cloud computing is a set of
services delivered via the internet (Korzeniowski, 2010). This allows the broadest
Library Hi Tech possible interpretation of what it is, but problematically it suggests the necessity of an
Vol. 30 No. 3, 2012
pp. 418-427 external entity when cloud computing does not need one. There is no reason why the
q Emerald Group Publishing Limited
0737-8831
model can’t work on an internal basis. In fact an ideal scenario might be IaaS delivered
DOI 10.1108/07378831211266564 through central IT to departments on an academic campus.
 Literature review Running a
With the maturation of cloud computing over the past several years, much interest and library project in
attention has been paid to this technology and its services by academic libraries and
other institutions. The authors Yang (2012), Mitchell (2010) and Doelitzscher et al., 2010 the cloud
explicitly adopt the three-tier classification of Software-as-a-Service (SaaS),
Platform-as-a-Service (PaaS), and Infrastructure-as-a-Service (IaaS) to categorize
cloud computing. Lin et al. (2009) also adopted the same SaaS/PaaS/IaaS notation to 419
introduce cloud computing, he further expands this idea by using a layering system.
They use the same three-layer architecture that is most often used to describe cloud
computing. However, the way the model is presented has the potential to mislead new
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

cloud users with the impression that one type of cloud computing is reliant
hierarchically on another type. Although there is some correlation and interdependence
between types of cloud computing (SaaS/PaaS/IaaS), each can be thought of separately
without any obvious relationship. Chorafas(2011) generalizes cloud computing into
four pillars:
(1) applications;
(2) platforms;
(3) infrastructures; and
(4) enabling services.

The first three pillars are the same categories that cloud computing is most often
segmented into:
(1) SaaS;
(2) PaaS; and
(3) IaaS.

The last pillar, enabling, is more similar to IT consultancy/support provided by

vendors like IBM, HP and Cisco to help their clients successfully transition to cloud
computing. The fourth pillar can be removed when defining cloud computing.
Discounting the fourth pillar, Chorafas’ classification system is largely the same as
most other definitions of what constitutes cloud computing.
Han (2010) describes his successful experience migrating and hosting the Koha ILS
system into the cloud for the Rebuilding Higher Education in Afghanistan project led
by the University of Arizona Libraries. The cloud provider selected by him is not as
widely known when compared to other major providers. Yang (2012) introduces
Amazon’s EC2 service with the emphasis on cost analysis and comparison. Little
technical information is given on how to set up a machine instance with Amazon EC2.
Yang also discusses the trend of hosting library applications (ILS, discovery tools etc.)
in the cloud. She acknowledges that data security is a challenge, but offers no
recommendations or solutions.
Mitchell (2010) presents three dimensions to cloud computing:
(1) quality of service;
(2) impact on library services; and
(3) cost comparison.
 LHT These are used to gauge the success a cloud migration project implemented at Z. Smith
30,3 Reynolds Library. Though the project is deemed a success by the author there is some
concern over the reliability of connections for key library applications (the integrated
library systems) off the campus into Amazon EC2 cloud without a full contingency
plan in place. Han (2010) presents an extensive overview of the major types of cloud
computing as well as their providers. The article discuses Google’s AppEngine and
420 provides a rough total cost of ownership analysis of Amazon’s services.
Doelitzscher et al. (2010) introduces an expandable and extensible private cloud project,
developed and implemented at Hochschule Furtwangen University (HFU). The project is
a cutting-edge and powerful cloud computing environment built at an academic
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

institution. The high technical threshold means that most institutions would not be able to
build similar services. Nevertheless, it is an interesting look at private/hybrid cloud
computing which would nicely supplement or replace the public cloud.

Examples of cloud services

The most well-known type of cloud service is SaaS. Examples include free e-mail
services like Yahoo mail and Google docs. SaaS offers the least amount of control for
the customer, but also requires little effort. It is possible to purchase large software
offerings as SaaS and many ILS vendors offer their products in this environment.
There is little or no opportunity to modify code or work through an API using SaaS.
This is a good choice for those who want little responsibility or control over their
software, but bad for anyone who wants a high level of customization.
Software as a service has the most potential implications for security. If users will
be authenticating through a library or campus server such as LDAP that means patron
IDs and passwords will be passing through a third party server. Users will be logging
in through a server on the web and their credentials will be passed back to the campus
or library. In theory this means anyone in control of the server can log user names and
passwords. For many large schools that means security departments need to be
consulted and the vendor needs to have clearly established polices for privacy and
intrusion. If problems arise over user authentication it can become problematic to solve
them, as the vendor points to IT and IT points at the vendor.
Platform as a service is not quite as common as SaaS and is probably a little less
understood. Many of the discovery tools that are on the market today, such as Ebsco
Discovery Service and Serials Solution Summon, are sold as SaaS. The only way to
make modifications to the software is through a web user interface. Typically, there is
little control over the look and feel of the site. There is at least one discovery tool that is
sold in tiers. One tier offers access to a web user interface that allows minor stylistic
changes. This is SaaS. In another tier they allow command-line level access. This
means that code can be modified directly. This is an example of PAAS. There is access
to code but not at the machine level; there is no control over the server (be it virtual or
physical). Control over a software package is given at a fairly low level, but the user is
not responsible for anything that might go on with the operating system.
Infrastructure as a service is probably as common as SaaS, but is less known
outside of IT. It has the greatest potential impact for both IT departments and
managers and it will be the focus of this article. Using IaaS is for the most part like
running a virtual host on a computer. There is access to code and configuration of the
server all the way down to the system level. The user is free to do with it what they
 want and they are responsible for any problems or issues that might crop up. There is Running a
very little or no support through most of the big IaaS providers. Backups are the library project in
responsibility of the user. If the node that the virtual host is running on dies, all the
information on the host is potentially lost. This paper discusses how IaaS cloud the cloud
computing was explored, tested and adopted at an academy library.

Embarking into the cloud 421

It is typical for an academic library to get requests for web sites or access to
applications that groups or departments want to try out. This means either finding
space on existing servers or purchasing machines for the specific task. Fondren
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

Library at Rice University has had a projects sever for a number of years. It hosts
numerous sites and projects. The server was built and deployed with Debian Etch in
mid-2008. At the request of the preservation department Archivist’s Toolkit, an open
source data management tool, was installed on the server. Over the years it
accumulated a number of other web sites and it eventually ran a mobile application.
Sometime in 2010 the preservation department made a request for the installation of
Omeka, an open source web-publishing tool. The server largely had all of the
prerequisites for the installation. It was running MySQL, PHP and Apache, Omeka
installed on the machine easily. The site was turned over to the preservation
department and they did some testing and quickly found some issues. Every time they
tried to delete items or files from Omeka’s web interface they were not able to do so and
the application threw a MySqli error suggesting some MySQL table was not valid, or
the application would crash. After investigation and testing it was found that the
server needed to be at a higher version of MySQL. Though not directly related to the
Omeka issue, it was determined that it was better to also bring PHP and the essential
modules to a higher version to avoid any other bugs. This meant upgrading Debian to
Lenny, which was a newer version of the operating system (OS), which had the version
of MySQL and PHP that was needed.
Most versions of Linux today, including Debian, use a package management
system. This means software is installed out of repositories using utilities built into the
system. This solves a great problem that existed in the early days of Linux called
dependency hell, where users would install a piece of software, only to be informed that
they needed to install some other piece of software to get the current one to work.
Package management is great in that it solves this problem. It also means that users
must use the version of software which is in the repositories, or compile their own.
Compiling software means taking source code and turning it into a usable program for
the computer. This is a workable solution until the operating system needs to be
upgraded. If the operating system is upgraded, and the libraries that the compiled code
depend on change, it breaks the program. The projects server had codes that had been
compiled and shared libraries had been manually linked for other projects, so
upgrading the production system without thorough testing and practicing was going
to be a concern. In theory another version of MySQL could have been compiled to get to
the right version, but why use a package managed system if it means continually
compiling code?
Cloud computing offered a solution to this problem. Why not use cloud
infrastructure to host the applications and projects on the Debian box and then rebuild
it with a newer operating system and then move back to the rebuilt box?
 LHT Choosing a cloud service provider
30,3 While the rest of the market was suffering over the past few years, cloud-computing
providers saw their fortunes rise. Rackspace (NASDAQ ticker symbol: RAX), a hosting
and cloud computing service, saw their stock price rise by over ten times between
March of 2009 and December of 2011. Since cloud computing has become so hot more
providers have come to join in an already crowded market. Comprehensive or quick
422 comparison charts and matrices on cloud computing providers can be found on the
internet to help choose a vendor. Comparisons are made mainly in respect to:
.
types of cloud computing;
.
pricing schedule;
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

.
supported platforms and/or operating systems;
.
add-on features, services and tools;
.
support and service level agreements;
.
administrative interface;
.
data security; and
.
number of years in cloud computing field.

The site http://www.cloudorado.com/ provides has a tool for comparing cloud providers.
Users can calculate price based on ram size, storage and CPU. Some stats may be
misleading however. It only shows the price for a standard large server from Amazon and
this makes the prices look significantly higher than had they chosen a smaller server.
Amazon is the largest supplier of cloud-based solutions. They have been around for
a long time, and they do not appear to be going anywhere. For Library IT department
they were a logical choice. They were easy to set up, and a staff member had already
been using their service outside of work.

An introduction to Amazon’s cloud services

Founded as an internet book seller by Jeff Bezos in 1994, Amazon has grown into a
mighty omnibus e-business platform and e-business ecosystem (Isckia, 2009). Over the
course of its growth Amazon has been innovative in developing and adopting new
technologies to help fulfill its market goals and to capitalize on its technological
innovations. In early 2006, three years later after its launch of Amazon Web Services
(AWS), Amazon released two cloud computing products; Amazon S3 for storage and
Amazon Elastic Computing (EC2). In 2008 Amazon brought out Elastic Block Storage
(EBS), which acts as a persistent storage solution for EC2. These three core products
EC2, EBS and S3 comprise the backbone of Amazon’s cloud computing solutions.
Amazon also offers numerous add-ons features and services such as Auto Scaling and
CloudWatch, which differentiate them from their competitors.
EC2 is the key computing component of Amazon’s IaaS Cloud computing solution.
A user can order, on demand, a virtual computing environment called an Amazon
Machine Instance (AMI) which is typically just referred to as instance. They offer small
32-bit instances (1.7 GB of memory /1 EC2 Compute Unit /160 GB instance store) and
up to 64-bit Cluster Quadruple Extra-large Instances (33.5 EC2 Compute Units/23GB of
memory/1,690 GB of instance storage). An instance is not persistent and data is wiped
out when it is terminated. Conversely, data in EBS is persistent so if for some reason an
 instance is terminated or fails it can be retrieved. A user can create and store up to 1TB Running a
of data in EBS. This is similar to attaching a disk to a computer through a network library project in
share. If the computer fails the information on the disk persists. By combining EC2
with EBS a user can create a virtual computing environment, which has a flexible the cloud
instance with a persistent data store.

Bringing up an Amazon machine instance 423

It is necessary to have an Amazon account in order to create an instance through AWS.
The account must have a phone number associated with it. Additionally, the AWS
account setup process provides users with a PIN number. When the account is
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

launched Amazon calls the phone number and the user is prompted to input a PIN.
From this point forward a user can start using the S3 service, but EC2 where users
create instances, can take a little longer. Once available it is possible to log into the
AWS console and choose EC2, where instances are launched.
There are two interfaces for launching an instance; classic and quick launch. Classic
allows for greater control over the type of instance and its configuration. Quick launch
allows for rapid deployment of an instance, but it gives less control over how it is
configured. Classic lets users choose more distributions and has more options. Classic
also lets a user choose community AMIs. This is helpful for a number of reasons.
Community AMIs allows users to choose preconfigured systems that others have
created. If someone wants to bring up an AMI that already has Drupal installed, and
they do not want to do it from scratch, they can do so through the Classic interface.
Amazon allows user to choose from Windows or a number of different versions of
Linux for the operating system of the AMI. This article focuses on the creation of an
AMI based on Linux. Costs vary, depending on the operating system chosen. Amazon
has their own Linux distribution and it is based on Red Hat so anyone familiar with
that flavor of Linux should feel right at home. It uses the same package management
system and has roughly the same layout and settings as Red Hat does. If it is necessary
to use a distribution that Amazon does not offer, they can be built, or often they are
available through community AMIs.
An instance has two requirements; a name and a key pair. The name is arbitrary, but
it is necessary to understand how a key pair works in order to be able to log into a
system. When an instance is created a user is prompted to either pick a key or to create
one. If it is the first time any instance has been brought up the system offers a prompt to
create one. Amazon then creates an access key, which is downloaded. A private key will
be placed in the home directory of the default user, which matches up with the access
key. The home user is likely to be named “root” or ec2-user. The two keys are linked
through a mathematical algorithm and allow for encrypted communication between a
computer and the AMI. Once the key pair has been created the user downloads the access
key. This key will need to have permissions set so that it is read-only by a single user.
The whole process of launching the instance takes only a few minutes and is almost
deceptively easy. Once it is up and running it is important to remember that all of the
same rules that apply to physical hardware apply to a cloud server. Anything
worthwhile is worth backing up. Amazon could have a failure, and human error is
always a factor. Once an instance is up and running the user logs in through the key
that was created. On a machine running Linux this is fairly easy. A user need only type
the correct command and they authenticate with their key.
 LHT Even though Amazon and many other third parties provide a large number of public
30,3 AMIs ready for immediate use there may be reasons to create an AMI from scratch.
Among them is the desire to have clean installation of a Linux distribution, which
Amazon does not offer. It is not unusual for an IT department to make the decision to
standardize on one distribution of Linux. If that distribution is not supported by
Amazon, it can be built, or it can be installed from a community AMI. Community AMIs
424 often has configuration changes and other tweaks to them. To get a clean install of a
distribution not offered by Amazon it is necessary to build it from scratch.

Into the cloud

Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

A project like Omeka does not need use a lot of computational resources. Amazon
offers different sizes of instances each with its own tier of pricing. A small project like
Omeka running Linux, Apache and MySQL can be run on a small instance. Amazon’s
small instance (m1.small) is roughly equivalent to a 32 bit computer with 1 1.0-1,2Ghz
2007 Opteron or Xeon CPU, 1.7 GB memory and 160GB disk space including an
optional partition. This needs to be paired with a subscription to Amazon’s Elastic IP
service, which allows users to assign static IPs to the instance. Once the IP has been
attached a hostname can be assigned through DNS.
Once the instance is up and running a user would then sign in using SSH. From
there installing Omeka is simple. Prerequisite software packages would need to be
installed including the right version of MySQL and the correct PHP module. Also, any
packages that the Omeka software requires would need to be installed. These can all be
pulled in using a package management system like Yum.
Omeka runs over HTTP and is an excellent candidate for the cloud. The library IT
department also tried to port Archivists Toolkit to the cloud. Testing showed this
application to suffer from a noticeable performance hit. It runs in as a traditional
client/server application that runs at the socket level. For reasons that could not be
explained – possibly transmission of larger chunk of data, bottlenecks on the network
path between Rice and Amazon – there was an unacceptable amount of slowness with
this application. The decision was made to move this server back to campus. This also
led to the conclusion that nothing would be ported to the cloud but web sites and web
applications. If an application doesn’t run over HTTP or HTTPs it may not be a good
candidate for the cloud.

Picking a project for the cloud

There are good and bad projects and services to move to the cloud. It’s important to
keep in mind that anyone can get an account with Amazon and set up a machine
image. There is no vetting process that determines who gets to run an instance. It
might be best not to pass patron information across the web. It might not be a good
idea to run a file server with sensitive data on it up in the cloud. Services typically cost,
based on bandwidth and storage. This probably also rules out a file server. A library
web site may be too big and the bandwidth may be too intensive. Many integrated
library systems do run over SaaS. However, many academic libraries have hooked this
system into other local services like billing and acquisitions, which may make this
impractical.
What is a good candidate for the cloud? Particularly if it will be running as IaaS?
Something that runs over the web. Something that has limited logins. Something that
 is at least not initially system critical. A project like Omeka is a good choice because it Running a
is small. A project that is new because it is probably not yet critical yet. library project in
Important considerations
the cloud
Cloud computing is not fool proof. It is not redundant. Amazon freely admits that if a
node goes down the instances running on that node are likely to have lost all their data.
Human error is as great a problem as it is when physical hardware is used. It is a good 425
practice to backup both data and AMIs. There are as many schemes for doing backups
in Amazon’s IaaS service as there are for physical hardware. There are numerous
commands that can be used to copy data from an AMI to a more permanent storage
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

location such as S3 or EBS. These commands can be used to create whole new AMIs,
which can then be started from the AWS console. If an instance were to be lost for some
reason it can instantly be restarted from the point of its last creation. Important data
may be better off being stored on EBS, which is persistent. That way if the AMI goes
down a new one can be started which can then mount EBS. This way no data is lost.
Cloud computing has security issues and like all forms of outsourcing, cloud
computing raises serious concerns about the security of the data assets that are
outsourced to providers of cloud services ( Julisch and Hall, 2010). Security departments
should be notified if sensitive information is passed between campus and a cloud
service provider. Academic universities have security policies that may apply to cloud
computing. SaaS solutions that require user names and passwords which authenticate
through a campus LDAP server, are vulnerable if they are compromised. The
information passed through a cloud provider is potentially susceptible to harvest and
snooping. There are steps that can be taken to mitigate these vulnerabilities and they
include using secure connections. Cloud providers should have clear and obvious
security policies, which they can provide to campus security personnel.
Commercial cloud providers are normally expected to have better more reliable data
centers as well as more advanced and efficient disaster prevention measures. However,
Amazon did experience a major outage in April of 2011 (Pepitone, 2011). The incident
left users without services for nearly 24 hours and took down some of the largest sites
on the internet. It is tempting to look at this issue as proof of the dangers of cloud
computing, but what it should signal is that the cloud, like any new technology, should
be approached with caution but approached nonetheless (Preimesberger, 2011).

Advantages of using Amazon’s cloud services

By adopting cloud computing, users can buy computing literally by the drink recording
the purchase as an operating expense rather than a capital investment (Kambil, 2009).
One of the greatest advantages to the cloud is the flexibility it affords. A production
system can be backed up on demand and in literally minutes a mirror of the same system
can be restarted. This allows for easy testing and refining of a system without worry of
taking it down. It’s a great way to try out new plug-ins for projects like Omeka or even
Drupal. It doesn’t require buying new servers and it can be done rapidly.
The flexibility of cloud computing is one of its most compelling selling points. The
Library IT at department was asked to prepare a second Omeka server for an
upcoming workshop. The server did not need to be permanent and it needed to be up in
two days. There was already an Omeka server running as an EC2 instance. A backup
script was ran which created a new AMI. The new instance was started and this meant
 LHT that there were now two identical versions of the same server running. In a traditional
30,3 computing environment this would have taken longer and would have had a number of
complications. It also would have meant working on a production server, which is a
potentially riskier activity. The total cost of running the new instance was somewhere
around four dollars.
A small EC2 instance does not offer a lot of computational resources. There is not a
426 lot of RAM, disk space or processing power. If demand rises, it is possible to scale this
server up without interrupting the service for longer than a few minutes. Amazon
offers Auto Scaling and Elastic Load Balancing services, which allow computing
capacity to be grown or shrunk in a way that is transparent to users. With the
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

scalability and expandability of cloud computing, we can prevent server sprawl

(Castro-Leon and He, 2009) and over-provisioning IT infrastructure (Doelitzscher et al.,
2010) which are issues frequently seen in traditional way of deploying and scaling
computing servers.

Private and hybrid clouds

There are many benefits to adopting cloud-computing technologies, but should all IT
departments replace data centers and trust everything to an external cloud-computing
provider? Assuming an organization has a decently sized technology infrastructure, the
answer to this question is probably no. If organizations fall blindly into the arms of
vendors because cloud computing becomes a fashion then they will be in a trap and they
will deserve it (Chorafas, 2011). So what’s the best approach and strategy for an
established institution in dealing with cloud computing? The answer is to build a private
cloud or hybrid cloud. Private cloud deployments’ primary aim isn’t to sell capacity over
the internet through publicly accessible interfaces but to give local users a flexible and
agile private infrastructure to run service workloads within their administrative domains
(Sotomayor et al., 2009). Hybrid clouds are designed to supplement local infrastructure
with computing capacity from an external public cloud (Sotomayor et al., 2009). By
building a private cloud, an institution can take advantage of their existing computing
capacity at the least risk and with the most control. There are already universities
running private clouds. One example is the Cloud IA (Cloud Infrastructure and
Application) developed and implemented at Hochschule Furtwangen University (HFU)
(Doelitzscher et al., 2010). It is an example of a private cloud, which can grow its
computing capacity beyond its underlying local resources by seamlessly adding in
computing resources from public clouds like Amazon’s EC2 cloud.

Conclusion
Rice’s Library IT began using cloud computing due to a routine administrative task. A
project was picked which could be gently ported to the cloud as a safe test. If it didn’t
work it could be brought back to the data center. Some conclusions were drawn about
would could and could not be ported to the cloud. After a year’s worth of time a number
of other projects and services are being considered for the cloud. A number of older yet
popular web sites are good candidates. Windows AMIs is also being considered for
some sites which were going to be ported to Linux. By moving these types of projects
to the cloud it is thought that both time and money can be saved. The cloud is not right
for every kind of service, but it is certainly correct for some. The preliminary
experience in the cloud is encouraging enough that more applications will be ported
 and it is thought this can be done so in an efficient manner. There is advantage in Running a
server deployment and support, as well as some advantages to budgeting.
library project in
References the cloud
Castro-Leon, E. and He, J. (2009), “Virtual service grids: integrating IT with business processes”,
IT Pro, Vol. 11 No. 3, pp. 7-11.
Chorafas, D.N. (2011), Cloud Computing Strategies, CRC Press, Boca Raton, FL. 427
Doelitzscher, F., Sulistio, A., Reich, C., Kuijs, H. and Wolf, D. (2010), “Private cloud for
collaboration and e-learning services: from IaaS to SaaS”, Computing, Vol. 91 No. 1,
pp. 23-42.
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)

Han, Y. (2010), “On the clouds: a new way of computing”, Information Technology and Libraries,
Vol. 29 No. 2, pp. 87-92.
Isckia, T. (2009), “Amazon’s evolving ecosystem: a cyber-bookstore and application service
provider”, Canadian Journal of Administrative Sciences, Vol. 26 No. 4, pp. 332-43.
Julisch, K. and Hall, M. (2010), “Security and control in the cloud”, Information Security Journal:
A Global Perspective, Vol. 19 No. 6, pp. 299-309.
Kambil, A. (2009), “A head in the clouds”, Journal of Business Strategy, Vol. 30 No. 4, pp. 58-69.
Korzeniowski, P. (2010), “SaaS, IaaS, PaaS? You must be talking ‘cloud’; it’s no passing fancy;
cloud computing keeps growing, as do the tech definitions of the term”, Vol. 3, November,
p. 3.
Lin, G., Fu, D., Zhu, J. and Dasmalchi, G. (2009), “Cloud computing: IT as a service”, IT Pro,
March/April, pp. 10-13.
Mitchell, E. (2010), “Using cloud services for library IT infrastructure”, Code4lib, No. 9.
Pepitone, J. (2011), “Amazon Ec2 outage downs Reddit”, Quora, available at: http://money.cnn.com/
2011/04/21/technology/amazon_server_outage/index.htm (accessed December 2, 2011).
Preimesberger, C. (2011), “Thoughts on AWS outage”, eWeek, Vol. 28 No. 9, p. 14.
Sotomayor, B., Montero, R.S., Llorente, I.M. and Foster, I. (2009), “Virtual infrastructure
management in private and hybrid clouds”, Internet Computing, Vol. 13 No. 5, pp. 14-22.
Yang, S.Q. (2012), “Move into the cloud, shall we?”, Library Hi Tech News, Vol. 29 No. 1, pp. 4-7.

About the authors

Denis Galvin is currently employed by Rice University. He has been working in academic
libraries for almost 12 years. His primary responsibilities have been with the integrated library
system but necessity, curiosity and flexibility have allowed him to work on many other projects.
Prior to working at Rice University he worked at Carnegie Mellon University in Pittsburgh.
Before that he worked for Northland Public Library in Pittsburgh. He is three classes shy of
finishing his MLIS from the University of North Texas. Denis Galvin is the corresponding author
and can be contacted at: dgalvin@rice.edu
Mang Sun works for Fondren library at Rice University where he is the systems librarian. His
main job duties include administrating several library systems including the ILS system and
developing in house applications and solutions when needed. Prior to Rice he worked for
Saskatchewan Provincial Library in Canada as a systems librarian (information resources
specialist). His research interests cover a wide spectrum of library and IT technologies. He got his
MLIS degree from Dalhousie University in 2006.

To purchase reprints of this article please e-mail: reprints@emeraldinsight.com

Or visit our web site for further details: www.emeraldinsight.com/reprints
 This article has been cited by:

1. Mu‐Yen Chen, Edwin David Lughofer, A.N. Zainab, C.Y. Chong, L.T. Chaw. 2013. Moving a repository
of scholarly content to a cloud. Library Hi Tech 31:2, 201-215. [Abstract] [Full Text] [PDF]
2. Yan Han. 2013. IaaS cloud computing services for libraries: cloud storage and virtual machines. OCLC
Systems & Services: International digital library perspectives 29:2, 87-100. [Abstract] [Full Text] [PDF]
Downloaded by University of Maryland College Park UMCP At 12:23 07 January 2015 (PT)