Resource Allocation in a Network-Based Cloud Computing Environment: Design Challenges (1309.1208v1)

Abstract: Cloud computing is an increasingly popular computing paradigm, now proving a necessity for utility computing services. Each provider offers a unique service portfolio with a range of resource configurations. Resource provisioning for cloud services in a comprehensive way is crucial to any resource allocation model. Any model should consider both computational resources and network resources to accurately represent and serve practical needs. Another aspect that should be considered while provisioning resources is energy consumption. This aspect is getting more attention from industry and governments parties. Calls of support for the green clouds are gaining momentum. With that in mind, resource allocation algorithms aim to accomplish the task of scheduling virtual machines on data center servers and then scheduling connection requests on the network paths available while complying with the problem constraints. Several external and internal factors that affect the performance of resource allocation models are introduced in this paper. These factors are discussed in detail and research gaps are pointed out. Design challenges are discussed with the aim of providing a reference to be used when designing a comprehensive energy aware resource allocation model for cloud computing data centers.

• The paper analyzes major design challenges for resource allocation in network-based cloud computing, emphasizing the integration of computational, network, and energy efficiency considerations.
• It classifies existing resource allocation models into categories focusing on data center processing, network resources, and energy efficiency strategies.
• The paper details external and internal design challenges including regulative constraints, charging models, data locality, network reliability, SDN design, and fault tolerance issues.

Resource Allocation in Network-Based Cloud Computing Environment: An Analysis of Design Challenges

The paper "Resource Allocation in a Network-Based Cloud Computing Environment: Design Challenges" offers a detailed examination of the multifaceted challenges related to resource allocation in cloud computing environments. The authors, M. Abu Sharkh, M. Jammal, A. Shami, and A. Ouda, assert the critical importance of achieving a balance between computational and network resource provisioning, while considering energy efficiency. The paper provides a framework for understanding these challenges, vital for developing comprehensive resource allocation models in cloud data centers.

Overview of Resource Allocation Models

The paper classifies existing Resource Allocation (RA) models into three distinct categories:

1. Data Center Processing Resources: Models in this category prioritize computational resource scheduling based on user requests. Emphasis is placed on minimizing the distance between VMs and optimizing communication costs within data centers.
2. Data Center Network Resources: These models address network provisioning problems, treating RA as an optimization issue aimed at maximizing revenue. The focus is on routing data effectively among virtual networks without addressing connection start times or durations.
3. Energy-Efficient Resource Allocation: Efforts here include minimizing energy use by consolidating processing tasks onto fewer servers. Such models often face challenges concerning network component power consumption, particularly when optimizing for minimal server utilization.

Design Challenges in Network-Aware Resource Allocation

The paper delineates several design challenges that must be overcome to establish a robust network-aware RA system. These challenges are categorized into external and internal factors:

• External Challenges:
  • Regulative and Geographical Constraints: These arise primarily from laws like HIPAA, which dictate physical data locations and security practices.
  • Charging Model Issues: Cloud services need to integrate client billing models with efficient RA, considering varied pricing structures for virtual network usage and resource excess.
• Internal Challenges:
  • Data Locality: A pertinent issue for enhancing scalability and performance, the paper stresses the need for data-aware schedulers and an understanding of data movement costs.
  • Reliability of Network Resources: Ensuring reliability requires decisions on network topology, virtualization, and bandwidth allocation.
  • Software-Defined Networking (SDN) Design: Concerns here involve SDN controller reliability, scalability, visibility, and placement problems.
  • Fault Tolerance vs. Performance: There's a need to balance fault tolerance strategies with system performance, especially concerning VM distribution across fault domains.
  • Portability and Vendor Lock-in: Agility in moving applications across different cloud providers remains a critical concern for clients.

Energy-Efficient Resource Allocation

Energy efficiency is a predominant concern for cloud data center operators, given the significant operational costs associated with power consumption. The paper illustrates common solutions, such as application consolidation on fewer servers, which although energy-saving, can lead to I/O and network bottlenecks. Moreover, considerations related to VM migration and its associated power trade-offs are pivotal for optimizing RA models.

Implications and Future Directions

The insights from this paper underscore the necessity of designing RA models that integrate computational, network, and energy considerations cohesively. While the paper highlights current practices and associated trade-offs, these models are crucial for achieving sustainable and optimized cloud service offerings. The implications are extensive, affecting both economic factors and broader regulatory compliance, thus shaping future research and technological advancements in the cloud computing domain.