Information Technology Reference
In-Depth Information
REST. The need for inter-cloud interoperability was highlighted by Vint Cerf, a co-designer
of the Internet's TCP/IP, who likened the current lack of cloud communication standards to
that of computer networks in the early 1970s (Krill, 2010). However, there are currently
efforts by some organizations such as the Cloud Computing Interoperability Forum to address
this issue (Grossman, 2009, p26).
Furthermore, failure of a cloud provider which owns data centers can have serious
repercussions for end users who trusted their data with such provider. This issue may force
potential cloud users to go for well established and large companies that are likely to be
around for many years to come.
Lastly, reliability can also be a serious problem for cloud users. Salesforce.com, for
example, left customers without service for six hours in February 2008 while Amazon's S3
and EC2 suffered a three-hour outage in the same month a few days later and an eight-hour
outage in July of the same year by S3 (Leavitt, 2009). In early 2009, Google's Gmail (its
Webmail service) went down for three hours, thus preventing its 113 million users from
accessing their emails or the documents which they store online as “Google Docs”
(Naughton, 2009). Vendors often provide service credits for outages. However, those credits,
according to a director of a US market research firm, are “cold comfort for sales opportunities
missed and executives cut off from business information” (Leavitt, 2009).
Cloud computing may not be suitable for all organizations. For example, for large
companies, the loss of service as a result of cloud glitches would be a major concern,
particularly if it impacts on their customers and results in substantial loss of sale opportunities
and customer dissatisfaction. The issue of reliability with relation to cloud services will
continue to be a problem. Similar glitches that befell the cloud services of Amazon and
Google are likely to surface again as the number of cloud providers and users increase.
However, for small companies struggling to survive the current global economic downturn
and cash-strapped educational establishments, often used to similar glitches caused by their
old in-house systems, cloud computing is likely to remain an attractive option due to its cost
structure and flexibility.
Furthermore, evidence is also emerging to suggest that even large companies, contrary to
conventional wisdom, are actually embracing cloud services. A recent report by Forrester (the
independent technology and market research company), based upon a survey of small and
large enterprises located in North America and Europe, revealed that large firms were more
interested than small firms in leveraging IaaS (Infrastructure as a Service) external cloud
capability (Golden, 2009).
For organizations involved in scientific research, most of the above concerns may not be
as important to them as they might be to those who provide services to consumers e.g., e-
commerce companies. For example, the loss of a few hours of services may not be as
dramatic for an organization conducting a research experiment as it would be for an online
auction or an online chain of retailers. Furthermore, issues of privacy and data protection are
likely to be of less concern or indeed relevance, especially if research organizations only use
an IaaS cloud for high speed compute operations. So, while the aforementioned cloud
drawbacks will remain of concern to many organizations contemplating using cloud
computing, their current advantages are likely to outweigh their potential disadvantages for
the scientific community. On that basis one could deduce that cloud computing's market
share of HPC is more likely to continue rising in the years to come. Indeed, there are
speculations that cloud computing's share of the computing service industry as a whole will
