Chapter 4
Examining Cloud Computing
From the Perspective of Grid
And Computer-Supported
Cooperative Work
Jinlei Jiang and Guangwen Yang
4.2.4 A Summary
Table 4.1 summarizes the differences between grid
and cloud computing. The key points are highlighted
below.
Grid computing adopts a resource-centric design and
tries to meet various needs by a unified resource pool.
As a result, many unnecessary details of the
infrastructure are brought to both end-users and
developers, making it difficult to use and hard to
program. In addition, it also raises a heavy burden of
system administration due to many administrative
domains being involved.
4.2.4 A Summary
4.2.4 A Summary
Cloud computing, on the contrary, adopts a
user- and task-centric design and it meets
diverse needs by different kinds of services, for
example, infrastructure services, platform
services, and software services. In cloud
computing, the complexity is shielded from
users. As a result, it is easy to use and program.
In addition, managing a cloud is also easy since
for most of the time there is only one
administrative domain involved in it.
4.3 Examining Cloud Computing from
the CSCW Perspective
• In the previous section, we examined
the differences between grid and
cloud computing. In this section, we
present some findings in CSCW
(Computer-Supported Cooperative Work)
research and utilize them to analyze
the cloud.
4.3.1 CSCW Findings
CSCW is a research field dealing with the issue of how to
use computers, or more broadly information and
communication technology (ICT), to facilitate a group of
people to fulfill a common task [6]. Over 20 years of
development since the term was first coined in 1984,
people have gained much knowledge about this field.
Cloud computing holds the promise to deliver computing
as a utility, so it is a socio-technical system (社會的技術系
統). It is in this sense that we think that the following
findings of CSCW (Computer-Supported Cooperative Work) can also
be used to answer the questions raised in Section 1 – why
cloud computing is so attractive and how to make the
vision of cloud computing really come true?
4.3.1 CSCW Findings
(Computer-Supported Cooperative Work)
Finding 1: To derive the greatest benefit from CSCW,
the supporting technology must infiltrate (滲透) as
widely as possible throughout the populace (全體居
民)[16]. In this way, cooperative system designers
can relieve themselves from hard work on such
issues as heterogeneous resource management and
interoperability and focus their efforts on more
essential issues such as understanding and
accounting for the characteristics of cooperative
work and then devising (設計) proper mechanisms
to support them.
4.3.1 CSCW Findings
(Computer-Supported Cooperative Work)
• Finding 2: Besides technical factors such as
usability and functionality, the deployment of
CSCW (Computer-Supported Cooperative
Work) is affected by social factors such as
various administrative and policy decisions
[16]. Sometimes, the social factors function
dominantly in making the decision on whether
to adopt a certain technology or not.
4.3.1 CSCW Findings
(Computer-Supported Cooperative Work)
Finding 3: A successful collaborative (合作的)
system must provide enough respect for the
social habits of end-users [9]. Human is the most
active and dynamic element in a collaborative
environment, and providing respect for his/her
habits means users can get better experiences
during collaboration. This in turn implies that the
collaborative system will be adopted by more
and more users, and therefore, the critical mass
(眾多)problem [14] will easily be met.
4.3.1 CSCW Findings
(Computer-Supported Cooperative
Work)
Finding 4: Incentives (動機) are critical [1]. CSCW
suffers from Grudin’s inequality [9], which says that
those who do additional work (capture and record
the articulation (接合) work associated with
collaboration) to make collaboration succeed may
not be the ones who benefit most from the results,
and thus it is necessary to provide some incentives
or reward to those persons. In this way, more
people will join the collaboration process and the
contribution of people will also increase.
4.3.2 The Anatomy of Cloud
Computing
The attractiveness of cloud computing can be
accounted for using Finding 1, Finding 3, and Finding
4 as follows.
First, cloud computing, in general, presents no new
technology. Virtualization technology, which is at the
core of cloud computing, was first developed in the
1960s. Other technologies such as web services and
Rich Internet Applications (RIA) also have a history of
no less than 5 years. Today, all these technologies are
well supported and popular.
4.3.2 The Anatomy of Cloud
Computing
For example, Intel and AMD have released several processors
with support for virtualization technology and there are many
virtual machine monitors (VMMs) available on the market (e.g.,
ESXi and vSphere from VMware, Hyper-V from Microsoft, XEN
and KM from the open-source community). In addition, web
services and RIA (Rich Internet Applications) have become a must for
the development of web applications. Thus, adopting these
technologies sets up a good basis for the prosperity (繁榮) of
this market because it makes entrance low. As a contrast,
there was no prevalent (流行的) integration technology at the
beginning of grid computing and designers have to develop
their own ways to integrate various resources.
4.3.2 The Anatomy of Cloud
Computing
Though the convergence of grid computing and SOA (Service-oriented
architecture) provides new opportunities for resource integration,
much work is still needed to reconcile various ways of information
representation. Put simply, the entrance for grid computing is high. As
a result, even after 10 years of development, grid computing today is
still in its infancy in many aspects [12].
Next, cloud computing provides enough respect for the social habits
of users because using a machine in the cloud is no different to using a
local machine. As mentioned above, users of cloud computing can
always work with their familiar tools and settings. In other words,
users of cloud computing adapt the running environment to their
applications rather than adapt the applications to the environment.
Therefore, they do not need to change habits developed over years,
which have a solid base.
4.3.2 The Anatomy of Cloud
Computing
On the contrary, as we discussed in Section 2, users of grid
computing, whether end-users or application developers,
have limited, if any, control over the running environment,
and have to bear many constraints being put on them. For
developers, to take the full advantage of grid computing,
they have to learn much for developing new applications or
adapting the existing ones to the grid, which is a heavy
burden to them. In addition, system administrators also
face many new challenges in coordinating resource sharing
and in guaranteeing the reliability, availability, and security
of the running environment due to the involvement of
multiple autonomous (自主的) domains.
4.3.2 The Anatomy of Cloud
Computing
The last but the most important point, users of cloud computing
need not do much, if any, additional work to use the services
provided by the cloud. All the work they do is necessary and the
same as what they do every day without cloud computing. For
example, reserving a VM in clouds is an analog of buying a physical
machine, but with much greater convenience. Installing software
in a VM is no different to that in a physical machine. What’s more,
users can benefit from the advanced features of cloud computing
such as unlimited resource being available on demand, no upfront
(預付的) commitment (承諾) and pay-as-you-go usage of
resources [2], and the great potentials for group collaboration as
well as the universal access to information and services [15].
4.3.2 The Anatomy of Cloud
Computing
These features are especially attractive to small- and mediumsized businesses (SMBs) or start-ups that do not have enough
resources for buying and maintaining servers and developing
applications from scratch (起跑線), for they imply a lot of
savings of running costs. In contrast, things are quite different
with grid computing. To use grid computing, much more
should be paid on application development, system
management, and so on. Particularly, since resource providers
in grid computing receive no reward for sharing their
resources, they are reluctant (勉強的) to help to solve various
problems encountered.
4.3.2 The Anatomy of Cloud
Computing
In summary, compared with grid computing, cloud computing
provides more benefits and rewards without changing the
working way that people are familiar with. Therefore, it is not
strange at all that cloud computing is attractive. Indeed, cloud
computing has accumulated a huge (potential) base of both
service providers and consumers, and many market-research
firms (e.g., IDC, Forrester, and Merrill Lynch) believe that
cloud computing has enormous growth potential.
In spite of the facts above, cloud computing is still in its
infancy and only has a limited adoption to now [11]. To make
the vision of cloud computing really come true, we examine
the obstacles (障礙物) to (rapid) growth of cloud computing.
4.3.2 The Anatomy of Cloud
Computing
As with any other new paradigm (範例), there are fears and
concerns about cloud computing related to technology, social
factors, or both. For example, Armbrust M et al. [2] listed the
top ten obstacles as availability of service, data lock-in, data
confidentiality (機密) and auditability (可聽見的), data
transfer bottlenecks, performance unpredictability, scalable
storage, bugs in large-scale distributed systems, scaling
quickly, reputation (名聲) fate (命運) sharing and software
licensing; Leavitt N [11] identified the challenges facing cloud
computing as control, performance, latency, security and
privacy, related bandwidth costs, vendor lock-in (關在裡面)
and standards, transparency, reliability, and others.
4.3.2 The Anatomy of Cloud
Computing
In our opinion, issues such as performance,
latency, scalability, and data transfer
bottlenecks are related to technology and
have been suffered for a long time before the
emergence of cloud computing. Though they
have some impact on the adoption of cloud
computing, the impact is limited. It is the
following issues that hinder the wide adoption
of cloud computing.
4.3.2.1 Security and Privacy
According to a survey by IDC, security and privacy is the main
concern of chief information officers and IT executives [11]. To us,
such a concern arises from the violation or change of users＇
social habits-data and applications in cloud computing are usually
stored or running on an external infrastructure outside a
company’s firewall, and users have to rely on service providers
NOT themselves, to protect their data and applications. Obviously,
this may be quite different from what users are used to when
using local machines. Since change of habit is a slow process, it is a
natural result that only cloud computing has a very limited
adoption just now according to Finding 2 and Finding 3. However
security and privacy might be only a perceived risk as asserted by
Armbrust M et al. [2]:
4.3.2.1 Security and Privacy
We believe that there are no fundamental
obstacles to making a cloud-computing
environment as secure as the vast majority of inhouse IT environments, and that many of the
obstacles can be overcome immediately with wellunderstood technologies such as encrypted
storage, Virtual Local Area Networks, and network
middleboxes (e.g., firewalls, packet filters).
4.3.2.2 Data and/or Vendor Lock-In
This concern arises from the fact that there are currently no
standards for IaaS, PaaS and SaaS interfaces, and as a result,
much work is needed for customers to port their data or
programs from one cloud to another. While we admit that
such concerns really matter it will become increasingly
important as more and more cloud providers emerge, we also
argue that its impact may not be as great as people think. On
the one hand, people undergo such constraints in the real
world. For example, designers have to make a choice between
J2EE and .NET platform when developing new applications.
When new hardware is bought, people have to install system
and application software again to deliver their services.
4.3.2.2 Data and/or Vendor Lock-In
If we treat migrating an application from one infrastructure
to another as the process of buying new hardware, the
inconvenience caused by incompatible virtual image
formats – a major problem with IaaS – would no more be a
problem. On the other hand, people have recognized this
problem and as a result, many standardization activities are in
progress. Some of them are listed below. For more
information, please refer to http://cloud-standards.org
Cloud security Alliance² was set up recently “to promote the
use of best practices for providing security assurance within
Cloud Computing, and provide education on the uses of Cloud
Computing to help secure all other forms of computing. ”
4.3.2.2 Data and/or Vendor Lock-In
The Open Grid Forum (OGF) established the Open Cloud Computing
Interface Working Group (OCCI-WG)³ in March 2009 to develop a clean,
open API for infrastructure clouds. The Storage Networking Industry
Association (SNIA) has created a technical workgroup to develop the new
Cloud Data Management Interface (CDMI).⁴ The Open Cloud Consortium
(OCC),⁵ another newly established organization, aims to “support the
development of standards for cloud computing and frameworks for
interoperating between clouds.” Finally, the Distributed Management
Task Force (DMTF) has released the Open Virtualization Format (OVF)
Specification [5] that “ describes an open, secure, portable, efficient, and
extensible format for the packaging and distribution of software to be run
in virtual machines.” Therefore, it is reasonable to believe that
interoperation between clouds will get easier, making the concern about
data/vendor lock-in less important.
4.3.2.3 Service Availability/Reliability
This concern ranked first in the list given by Armbrust M et al. [2].
It is a radical requirement of business continuity – users will not
adopt a system that is unreliable and often unavailable to run
their business. The reason why such a concern becomes so
important lies in the change of operating mode – services are
running outside a company’s firewall and the quality of services
relies not only on software vendors who develop services but
also on providers who host services. No doubt, the well-known
outages of Amazon S3, Google App Engine, and Salesforce.com
make the worry even much severe. However, as pointed out by
Armbrust M et al. [2], the IT infrastructures of Amazon, Google,
and Salesforce are better than their peers.
4.3.2.3 Service Availability/Reliability
In summary, as a new paradigm, cloud computing does bring
changes to business operation; that is, the operation is done
remotely, out of the users reach and full control. Since this differs
from what users are used to, it is natural to see that much
concern is raised according to the Findings 2 and 3. To address
this, time matters. We need time to tackle technical challenges;
we need time to cultivate (培養) application developers; we need
time to build trust between customers and service providers; we
need time to develop use cases to demonstrate the benefits of
cloud computing. Once people get to know the reward of cloud
computing over its risks, the wide adoption of cloud computing
will come true as implied by Finding 4.
4.4 Conclusions
In this chapter, we first examined the differences
between cloud and grid computing from their
development and the viewpoint of system and
users, respectively. Then, we analyzed the reasons
why cloud computing is so attractive and some
related concerns using the findings in CSCW
research. Since cloud computing adopts a user- and
task-centric design philosophy and shows enough
respect for the social habits of users in using
computers, its popularity is a natural result.
4.4 Conclusions
At the same time, like any other new thing, cloud
computing faces some challenges that slow its wide
adoption. As time goes on and more and more
experience is gained, cloud computing will
eventually become an effective and efficient way to
deliver computing as a utility. During this course, we
researchers should address how to overcome the
obstacles and demonstrate the real benefits and/or
advantages of cloud computing.