S.S. Manvi, G.

Krishna Shyam / Journal of Network and Computer Applications ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

related to the problem of resource allocation in VNs. It provides a (v) Data management is an essential aspect in particular, for
concise overview of various existing techniques for resource storage clouds, where data is flexibly distributed across multiple
allocation in VNs. These can be helpful for developing detailed resources. Implicitly, data consistency needs to be maintained over
designs, specifications and performance evaluation techniques for a wide distribution of replicated data sources. At the same time,
VNs. However, this paper does not attempt to provide an exhaus- the system always needs to be aware of the data location (when
tive survey on the resource allocation/management techniques in replicating across data centres) taking latencies and workload into
VNs. Multi-tenancy is a highly essential issue in cloud systems, consideration.
where the location of code and/or data is principally unknown and (vi) APIs and/or programming enhancements are essential to
the same resource may be assigned to multiple users. This affects exploit the cloud features. Common programming models require
infrastructure resources as well as data/applications/services that that the developer takes care of the scalability and autonomic
are hosted on shared resources but need to be made available in capabilities, whilst a cloud environment provides the features in a
multiple isolated instances. Multi-tenancy implies a lot of poten- fashion that allows the user to leave such management to the
tial issues, ranging from data protection to legislator issues. system.
While hardware based virtualization has many benefits, it lacks (vii) Tools are generally necessary to support development,
from a high level of scalability required to offer cost effective cloud adaptation and usage of cloud services. Getting a final product in
computing for masses. Multi-tenant virtualization remedies this cloud can be a significant challenge. Cloud computing reduces
bottleneck by focusing on software based virtualization. Siddhisena some of these problems by introducing tools and processes that
et al. (2011) present an approach to application virtualization using provide a complete server and storage environment without the
multi-tenant concept. need to interact with technical specialists.
(ii) Resource management: At any time instant, resources are to Today, resources are pooled on storage and virtualization plat-
be allocated to effectively handle workload fluctuations, while forms, a degree of elasticity is present, and services are more “on
providing QoS guarantees to the end users. The computing and demand” than ever before. Behind the scenes, there is some level
network resources are limited and have to be efficiently shared of automation present, with many of the tools and processes in
among the users in virtual manner. In order to perform effective place to build a self-service portal. There are tools that monitor,
resource management, we need to consider the issues such as tools that provision, and tools that cross the divide between both.
resource mapping, resource provisioning, resource allocation and There are a few vendors that offer pervasive approaches in
resource adaptation. The lack of mature virtualization tools and handling provisioning and managing metrics in hybrid environ-
powerful processor's have prevented growth of cloud computing. ments as in http://searchcloudcomputing.techtarget.com: Right-
Although relatively new, a fair amount of work by Urgaonkar et al. Scale, Kaavo etc.
(2010a) and Vaquero et al. (2009) has been done to examine RightScale's management environment provides interface to
current and future challenges for both users and providers of cloud users for managing the resources. It is designed to walk a user
computing. However, little has been done to understand the range through the initial process of migrating to the cloud using their
of operational challenge faced by users as they attempt to run templates and library. The core of Kaavo's product is called IMOD.
applications in the cloud. Chase et al. (2010) have considered the IMOD handles configuration, provisioning and changes to the
problem of energy-efficient resource management of homoge- cloud environment across multiple vendors in a hybrid model.
neous resources in Internet hosting centres. The main challenge is Like all major CIM (common information model) players, Kaavo's
to determine the resource demand of each application at its IMOD sits at the “top” of the stack, managing the infrastructure
current request load level and to allocate resources in most and application layers. One great feature in IMOD is its multi-
efficient way. cloud, single system tool. For instance, one can create a database
Metering of any kind of resource and service consumption is backend in Rackspace while putting presentation servers on
essential in order to offer elastic pricing, charging and billing. It is Amazon.
therefore a pre-condition for the elasticity of clouds. The issue here Both Kaavo and RightScale offer scheduled dynamic allocation
is to see that the users are charged only for the services that they based on demand and monitoring tools to ensure that information
use for the specific period of time. Cloud computing alone will not and internal metrics (like SLAs) are transparently available. The
help an organization to determine who will pay for what resource, dynamic allocation even helps meet the demands of those SLAs.
but it can help provide a platform for an infrastructure design that Both offer the ability to keep templates as well to ease the
establishes a charge-back model for metering and billing. deployment of multi-tier systems. There are several tools that
(iii) Network infrastructure management: Managing millions of are available to do a specific task. The issue is to see that we
network components (hubs, bridges, switches etc.) leads to develop tools which are effective and gives accurate results.
unsustainable administrator costs, requiring automated methods The non-technological issues also play a major role in realizing
for typical system management tasks. These automated methods these technological aspects and in ensuring viability of the infra-
needs to deal with increased monitoring data size of several orders structures in the first instance. Non-technological issues are as
of magnitudes higher than current systems. Gupta and Singh follows. (1) Economic aspects which cover knowledge about when,
(2009) suggest putting network interfaces, links, switches and why, how to use which cloud system, how this impacts on the
routers into sleep modes when they are idle in order to save the original infrastructure (providers long-term experience is lacking
energy consumed by the Internet backbone and consumers. in all these areas). (2) Legalistic issues which come as a conse-
Chiaraviglio and Matta (2010a) have proposed cooperation quence from the dynamic handling of the clouds, their scalability
between ISPs (Internet Service Providers) and content providers and the partially unclear legislative issues in the Internet.
that allow the achievement of an efficient simultaneous allocation
of compute resources and network paths that minimize energy
consumption.
(iv) Security, privacy and compliance are obviously essential in 3. Resource management problems in IaaS
all systems dealing with potentially sensitive data and code. To
ensure adequate security in cloud computing, various security There are several problems to be considered while managing
issues, such as authentication, data confidentiality, integrity, and resources, such as, type of resource required (physical/logical),
non-repudiation need to be considered. allocation, brokering, provisioning, mapping, adaptation and

Please cite this article as: Manvi SS, Krishna Shyam G. Resource management for Infrastructure as a Service (IaaS) in cloud computing:
A survey. Journal of Network and Computer Applications (2013), http://dx.doi.org/10.1016/j.jnca.2013.10.004i