Towards understanding uncertainty in cloud computing with risks of confidentiality, integrity, and availability

https://doi.org/10.1016/j.jocs.2016.11.011Get rights and content

Highlights

  • Study the role of uncertainty in the cloud computing resource and service provisioning.

  • Review sources of uncertainty and approaches for scheduling under uncertainty.

  • Consider privacy in the presence of the risks of confidentiality, integrity, and availability.

  • Discuss mitigating the risks of the loss of information, denial of access, interruptions in connections, and information leakage.

  • Discuss the challenge of defining a multi-criteria version of the problems.

Abstract

An extensive research has led to a general understanding of uncertainty issues in different fields ranging from computational biology to decision making in economics. However, a study of uncertainty on large scale computing systems and cloud computing systems is limited. Most of works examine uncertainty phenomena in users’ perceptions of the qualities, intentions and actions of cloud providers. In this paper, we discuss the role of uncertainty in the resource and service provisioning, privacy, etc. especially, in the presence of the risks of confidentiality, integrity, and availability. We review sources of uncertainty, and fundamental approaches for scheduling under uncertainty. We also discuss potentials of these approaches, and address methods for mitigating the risks of confidentiality, integrity, and availability associated with the loss of information, denial of access for a long time, and information leakage.

Introduction

Cloud technologies are widely used in the construction of IT infrastructure of business, academic, government, and people as a valid solution for data storage and processing. While having many advantages they still have many drawbacks, especially in the areas of security, reliability, performance of both computing and communication, to list just a few. The transition to big data and exascale also pose numerous unavoidable scientific and technological challenges.

In the cloud computing, services and resources are subject to considerable uncertainty during provisioning. Uncertainty may be presented in different components of the computational, communication, and storing process. It requires waiving habitual computing paradigms, adapting current computing models to this evolution, and designing novel resource management strategies to handle uncertainty in an effective way.

The management of cloud infrastructure is a challenging task. Reliability, security, Quality of Service (QoS), performance stability, and cost-efficiency are important issues in these systems. Available cloud models do not adequately capture uncertainty, inhomogeneity and dynamic performance changes inherent to non-uniform and shared infrastructures. To gain better understanding of the consequences of a cloud computing uncertainty, we study resource and service provisioning problems related with existing cloud infrastructures such as hybrid federation of public, private and community ones.

Extensive research examines the uncertainty phenomena in users’ perceptions of the qualities, intentions and actions of cloud providers, etc. among other aspects of cloud computing (Trenz et al.) [1]. But still, the role of uncertainty in the resource and service provisioning, provider investment, operational cost, programming models, mitigating risks of confidentiality, integrity, and availability etc. have not yet been adequately addressed in the scientific literature.

In this paper, we discuss two main topics: how to provide reliability, safety and privacy of information, and how to deliver scalable and robust cloud behavior under uncertainties and specific constraints, such as budgets, QoS, SLA (Service-Level Agreement), energy costs, availability, etc.

Reliability, safety and privacy. As more users use cloud technologies for building IT-infrastructure, reliability, safety and privacy become crucial for both providers and consumers.

Preservation of confidentiality interpreted as a limited access to information, integrity as the assurance that the information is trustworthy and accurate, and availability as a guarantee of reliable access to the information by authorized people are three most crucial components of cloud computing.

Cloud-based services can crash just like any other type of technology. For example, access to information Amazon users has been limited for a long time due to distributed denial-of-service (DDoS) attacks in 2009. In 2013, a series of cloud outages are reported for Amazon, Microsoft and Google. Technical failures and data loss due to power outages are reported by Amazon, Dropbox, Microsoft, Google, and Yandex Disk. In the first quarter of 2014, Dropbox has experienced service outages twice. Bankruptcy is imposed for cloud storage company Nirvanix in 2013.

Common methods of ensuring confidentiality are data encryption, user identity documents (IDs), passwords, and other ways of authentication through cards, retina scans, voice recognition, and fingerprints. Other options include security tokens, key fobs tokens, etc.

Integrity involves maintaining the consistency, accuracy, and trustworthiness of information, so that they are not changed and altered by unauthorized people.

Service availability depends on the robustness of the hardware, hardware repairs and maintaining a correctly functioning operating system environment, system upgrades, preventing the occurrence of bottlenecks, etc. Redundancy, failover, redundant array of independent disks (RAID) etc. can mitigate consequences when hardware failures occur.

An uncertainty in these areas demands the companies be smart providing solutions preventing losing any data in the long run even if the service is offline during the time.

Performance, scalability and robustness. The vast majority of the research efforts in scheduling assumes complete information about the scheduling problem and a static deterministic reliable execution and storage environments.

In (Tchernykh et al.) [2], we show a variety of types and sources of uncertainty: dynamic elasticity, dynamic performance changing, virtualization with loosely coupling applications to the infrastructure, resource provisioning time variation, inaccuracy of application runtimes estimation, variation of processing times and data transmission, workload uncertainty, processing time constraints (deadline, due date), effective bandwidth variation, and other phenomena (Table 1).

The performance can be changed due to sharing of common resources with other virtual machines (VMs). It is impossible to get exact knowledge about the computer system. Parameters like an effective processor speed, number of available processors, or actual bandwidth are changing over time. Elasticity has a higher repercussion on the QoS, but adds a new factor of uncertainty. It is difficult to estimate runtime of jobs accurately (Ramírez et al.) [3].

In most existing solutions, it is assumed that behavior of VMs and services is predictable and stable in performance. On actual cloud infrastructures, these assumptions do not hold. While most providers guarantee a certain processor speed, memory capacity, and local storage for each provisioned VM, the actual performance is subject to the underlying physical hardware as well as the usage of shared resources by other VMs assigned to the same host machine. It is also true for communication infrastructure, where actual bandwidth is very dynamic and difficult to guarantee.

A pool of virtualized, dynamically scalable computing resources, storages, software, and services of cloud computing add a new dimension to the service delivering problem. The manner in which the service provisioning can be done depends not only on the service property and resources it requires, but also users who share resources at the same time.

The growing number of scientific scalable applications require resources at the exascale. This demands increased scope for optimization and uncertainty quantification. Uncertainty analysis is an important tool for tuning application parameters making them adaptive to configuration and environment changes to take advantage of the vast amount of computational resources and extreme available concurrency.

Section snippets

Uncertainty

In spite of extensive research of uncertainty issues in different fields in the past decades ranging from physics, computational biology to decision making in economics and social sciences, a study of uncertainty for cloud computing systems is still not available. There are numerous types of uncertainties associated with cloud computing, and one ought to account for aspects of uncertainty in assessing the efficient service provisioning. Mitigating impact of uncertainty on the performance,

Towards secure cloud computing

Information security assumes defending information from unauthorized access, use, disclosure, disruption, modification, etc. Important research and industrial streams of cloud computing are to design a secure and fault tolerant multi-cloud environment, where confidentiality, integrity, and availability are not violated in the presence of the deliberate threats, accidental threats, and failures (Srisakthi and Shanthi) [7].

Confidentiality considers a set of rules and restrictions that limits

Reliability and privacy under uncertainty

In order to increase reliability and confidentiality of the data processing and storing, six basic approaches are applied: data replication, secret sharing schemes, redundant residue number system, erasure code, regenerating code, and homomorphic encryption.

Conclusions

The uncertainty is an important issue that affects computing efficiency bringing additional challenges to scheduling problems. It requires designing novel resource management strategies to handle uncertainty in an effective way. We address areas such as resource provisioning, application execution, and communication provisioning. They have to provide the capability to dynamically allocate, manage resources in response to changing demand patterns in real-time, and dynamically adapt them to cope

Acknowledgement

Part of the work was supported by CONACYT, México, grant no.178415 and a grant of the President of the Russian Federation for the young scientists SP-1215.2016.5.

Andrei Tchernykh received the Ph.D. degree from Institute of Precise Mechanics and Computer Technology of the Russian Academy of Sciences, Russia in 1986. He is currently a full professor in Computer Science Department at CICESE Research Center, Ensenada, Baja California, Mexico, and a head of Parallel Computing Laboratory. He is a member of the National System of Researchers of Mexico (SNI), Level II. He leads a number of national and international research projects. He delivered more than 50

References (47)

  • Top Threats to Cloud Computing v1.0. Cloud Security Alliance

    (2010)
  • D. Kliazovich et al.

    CA-DAG: modeling communication-aware applications for scheduling in cloud computing data centers

    IEEE 6th International Conference on Cloud Computing

    (2013)
  • A. Rodriguez et al.

    Algorithms for dynamic scheduling of unit execution time tasks

    Eur. J. Oper. Res.

    (2003)
  • U. Schwiegelshohn et al.

    Online scheduling for cloud computing and different service levels

    IEEE 26th International Parallel and Distributed Processing Symposium Workshops & PhD Forum

    (2012)
  • A. Tchernykh et al.

    Bi-objective online scheduling with quality of service for IaaS clouds

    3rd IEEE International Conference on Cloud Networking

    (2014)
  • A. Lezama et al.

    Performance evaluation of infrastructure as a service clouds with SLA constraints

    Computacion y Sistemas

    (2013)
  • J. Gonzalez et al.

    Load balancing for parallel computations with the finite element method

    Computacion y Sistemas

    (2013)
  • A. Tchernykh et al.

    Online hierarchical job scheduling on grids with admissible allocation

    J. Sched.

    (2010)
  • A. Tchernykh et al.

    Adaptive energy efficient scheduling in peer-to-Peer desktop grids

    Future Gener. Comp. Syst.

    (2013)
  • I.N. Sotskov et al.

    Sequencing and scheduling with inaccurate data

    Applied Statistica Science

    (2014)
  • N. Megow et al.

    Models and algorithms for stochastic online scheduling

    Math Oper. Res.

    (2005)
  • X. Cai et al.

    Scheduling with stochastic approaches

  • N. Megow et al.

    Approximation in preemptive stochastic online scheduling

    LNCS

    (2006)
  • Cited by (88)

    • On the improvement of uncertain cloud service capacity

      2023, International Journal of Production Economics
    • A systematic literature Review: Risk analysis in cloud migration

      2022, Journal of King Saud University - Computer and Information Sciences
    View all citing articles on Scopus

    Andrei Tchernykh received the Ph.D. degree from Institute of Precise Mechanics and Computer Technology of the Russian Academy of Sciences, Russia in 1986. He is currently a full professor in Computer Science Department at CICESE Research Center, Ensenada, Baja California, Mexico, and a head of Parallel Computing Laboratory. He is a member of the National System of Researchers of Mexico (SNI), Level II. He leads a number of national and international research projects. He delivered more than 50 keynote speeches and invited lectures, served as a program committee member and general co-chair of more than 100 professional peer reviewed professional conferences. His main interests include resource optimization technique, adaptive resource provisioning, multi-objective optimization, computational intelligence, and incomplete information processing.

    Uwe Schwiegelshohn received the Diploma and the Ph.D. degrees in Electrical Engineering from the TU Munich in 1984 and 1988, respectively. He was with the Computer Science department of the IBM T.J. Watson Research Center from 1988 to 1994 before becoming full Professor at TU Dortmund University where he heads the Robotics Research Lab. In 2008 he was appointed vice president of this university. Also in 2008 he became managing director of the Government sponsored D-Grid corporation to coordinate the Grid projects in Germany. His main research interest are scheduling problems and Grid computing.

    El-Ghazali Talbi received the Master and Ph.D. degrees in Computer Science from the Institut National Polytechnique de Grenoble in France. He is a full Professor at the University of Lille and the head of DOLPHIN research group from both the Lille's Computer Science laboratory (LIFL, Universite Lille 1, CNRS) and INRIA Lille Nord Europe. His current research interests are in the field of multi-objective optimization, parallel algorithms, metaheuristics, combinatorial optimization, cluster and cloud computing, hybrid and cooperative optimization, and applications to logistics/transportation, bioinformatics and networks. Professor Talbi has to his credit more than 150 international publications including journal papers, book chapters and conferences proceedings

    MikhailBabenko graduated from Stavropol State University (SSU) in 2007 with degree in mathematics. Received Ph.D. degree in mathematics from SSU in 2011. He works as assistant professor in Department of Applied Mathematics and Mathematical Modeling since 2012. He is an author of over 63 publications and 5 patents. His research interests include cloud computing, high-performance computing, residue number systems, neural networks, cryptography.

    ☆A preliminary reduced version of this article appeared in Proceedings of the SPU’2015 - Solving Problems with Uncertainties, in conjunction with The 15th International Conference on Computational Science (ICCS 2015), Reykjavík, Iceland, June 1–3, 2015. Procedia Computer Science, Elsevier, Vol. 51, Pages 1772–1781, 2015, DOI: doi:10.1016/j.procs.2015.05.387.

    View full text