1 Overview of Cloud Computing Journal Articles Brian Davis VCDE WS March 15, 2012.

1

Overview of Cloud Computing Journal ArticlesBrian Davis

VCDE WS March 15, 2012

2

Journal Articles(posted on Wiki: https://wiki.nci.nih.gov/x/IQIpAw )1. Buyya, R., Yeo, C., Venugopal, S., Broberg and J. Brandic, I. (2009) “Cloud computing and

emerging IT platforms: Vision, hype, and reality for delivering computing as the 5 th utility” Future Generation Computer Systems 25: 599-616.

• http://www.few.vu.nl/~kgr700/cloud%20computing%20and%20emerging%20it%20platforms.pdf

2. Marston, S., Li, Z., Bandyopadhyay, S., Zhang, J., and Ghalsasi, A. (2011) “Cloud computing – the business perspective” Decision Support Systems 51: 176-189

• http://faculty.winthrop.edu/domanm/csci566/StudyGuide/Cloud_Business.pdf

3. Rosenthal, A., Mork, P., Li, M., Stanford, J., Koester, D. and Reynolds, P. (2010) “Cloud computing: A new business paradigm for biomedical information sharing” Journal of Biomedical Informatics 43: 342-353.

• http://www.mitre.org/work/tech_papers/tech_papers_09/09_3026/

4. Sultan, N. “Reaching for the “cloud”: How SMEs can manage.” (2011) International Journal of Information Management 31: 272-278

• http://www.sciencedirect.com/science/article/pii/S0268401210001143

5. Chahal, S, Hahn-Steichen, J., Kamhout, D., Kraemer, R., Li, H. and Peters, C. (2010) “An Enterprise Private Cloud Architecture and Implementation Roadmap” an Intel IT White paper

• http://download.intel.com/it/pdf/Entrprse_Priv_Cloud_Arch_final.pdf

6. Leong, L. and Chamberlin, T. 2011 “ Magic Quadrant for Public Cloud Infrastructure as a Service” Gartner Report.

• http://www.gartner.com/technology/reprints.do?id=1-18BC06X&ct=111213&st=sb

33

Definitions (1)

“Cloud computing represents a convergence of two major trends in information technology —

(a) IT efficiency, whereby the power of modern computers is utilized more efficiently through highly scalable hardware and software resources and

(b) business agility, whereby IT can be used as a competitive tool through rapid deployment, parallel batch processing, use of compute-intensive business analytics and mobile interactive applications that respond in real time to user requirements.”- Marston et al. 2011

44

Definitions (2)

“It is an information technology service model where computing services (both hardware and software) are delivered on-demand to customers over a network in a self-service fashion, independent of device and location. The resources required to provide the requisite quality-of service levels are shared, dynamically scalable, rapidly provisioned, virtualized and released with minimal service provider interaction. Users pay for the service as an operating expense without incurring any significant initial capital expenditure, with the cloud services employing a metering system that divides the computing resource in appropriate blocks.”

-Marston et al. 2011

55

Aspects

We note that our definition does not explicitly require that the

services be provided by a third-party, but emphasizes more on the

aspects of

(1) resource utilization,

(2) virtualized physical resources,

(3) architecture abstraction,

(4) dynamic scalability of resources,

(5) elastic and automated self-provisioning of resources,

(6) ubiquity (i.e. device and location independence) and

(7) the operational expense model.

Cloud computing can be provisioned using an organization's own servers, or it can be rented from a cloud provider that takes all the capital risk of owning the infrastructure.

-Marston et al. 2011

66

Security in the Cloud

• Big concern, all articles talk about it (different pov)• “We find that some risks decrease and some

increase, with neither side of the argument overwhelming the other. Thus, each laboratory or consortium will need to assess security for its environment, while also considering tradeoffs...”

• – Rosenthal et al. 2009• Bottom line:• You need to address security no matter what• Cloud may or may not be suitable for your particular

needs (“it depends”)

77

Cloud Architecture

• Fig. 1 shows a schematic of the cloud computing model. It shows how the computing resources in the cloud can be accessed from a variety of platforms through the Internet.

• Marston et al. 2011

88

Virtual Machines

• High level view of cloud architecture, showing Virtual Machines (VMs).

• (Buyya et al. 2009)

99

More Definitions

The cloud computing industry often speaks about different delivery models of cloud computing, all of which refer to the different layers of the cloud computing architecture.• SaaS (Software as a Service)• PaaS (Platform as a Service• IaaS (Infrastructure as a Service)

-Marston et al. 2011

1010

SaaS, PaaS, IaaS (wikipedia)

• http://en.wikipedia.org/wiki/Infrastructure_as_a_service#Infrastructure

1111

Comparisons to other paradigms (Buyya et al, 2009)

See Table 1 of Buyya et al. 2009 for comparison of clusters, grids and clouds

1212

1313

Key Players (Table 1: Marston et al 2011)

• Established players• IBM• Google• Microsoft• AT&T

• Key Technology Providers• Apache (Hadoop)• EMC (VMWare-

Virtualization)• Cisco (standards for

portability)

• Innovators• Amazon (Elastic Compute Cloud

2; EC2)• Sales Force.com• Enomaly

• Enabler• CapGemini• RightScale• Vordel

1414

Key Players (Table 1: Marston et al 2011)

1515

Comparison of Some Key Players (Buyya, et al. 2009)

1616

Consortium Computing Roles and Examples (Rosenthal et al. 2010)

• Compute services• Eg, protein folding and simulations

• Extensive Storage • Eg, Gene and protein data

• Biomedical consortia • That facilitate the exchange of data and applications

among participants

1717

Cloud Possibilities for Biomedical labs (Rosenthal, et al, 2010)

1818

Good targets for near-term cloud initiatives (Rosenthal, et al. 2010)• “Cloud tends to be preferable when service demands are variable or

demand is unknown in advance and where the cloud vendor passes on large economies of scale in procuring servers, power, space and in supplying specialized staff and tools.”• The project has high costs for computing, admin, space, and

electrical power.• Members wish to share with outsiders• Project requires highly variable amounts of processing and

storage resources• System requires off-site backups• Applications have easily parallelized code• Want long term repositories to outlive the lab that hosts the data

• Examples:• Archiving, backup and fault tolerance, Sharing data and tools across

a consortium. High Performance computing (eg, BLAST)

1919

Less suitable targets (Rosenthal, et al. 2010)

• High performance computing applications where physical characteristics of hardware need to be known (for performance) and/or substantial data movement is needed among processors

• Institution already has decided against clouds for various reasons (eg, security)

• For applications that need to be highly available and communications systems are variable

• Need only local data• Existing projects may have sunk costs (that cannot

easily be recovered)

2020

Finis

2121

Extra slides

2222

SWOT Analysis (Strength, Weakness, Opportunities, Threats) (Marston et al. 2011)

Strengths

1. Lowers cost for entry

2. Immediate access to hardware resources

1. (shift Cap ex to Op ex)

3. Lower Information Technology Barriers to Innovation

4. Enterprises can Scale services easier

5. Makes possible new classes of applications and services

2323

SWOT Analysis (Strength, Weakness, Opportunities, Threats)

• Weaknesses• Loss of physical control of data• Trusting mission critical applications • Providers cannot commit to High QoS and availability guarantees

• (AWS SLA promises uptime 99.95% over 356 days)

• Opportunities• By definition, a “disruptive technology”

• Innovations that upset the existing order of things in a particular industry

• Developing countries (benefits without upfront costs)• China (74% of Chinese firms interested, 29% planning)• Ethiopia (education: common syllabi)

• Small Business• Eg Enterprise Resource Planning (ERP) software available.• Mashups (eg, Amazon’s GrepTheWeb)

• GreenIT - Reduce Energy consumption

2424

SWOT Analysis (Strength, Weakness, Opportunities, Threats)• Threats

• Backlash from entrenched incumbents• Threat to corporate IT culture• data security (75%)• Performance• Reliability• IT audit policies• job security, etc.

• Lack of Standards• Lock in to proprietary systems (not open source)• Tho ISO has started • And Open Web Foundation• Google’s “Data Liberation Front”

• Regulation at Local, national and International levels• Data privacy• Data access• Audit requirements (• Data location requirements• Eg, Sarbanes-Oxley and HIPPA

2525

IaaS players

• Leong and Chamberlin 2011

2626

Benefits (Rosenthal, et al 2010)

• Capacity at lower cost• Lower sys admin• Lower idle capacity• Lower power usage• Lower facility usage• Example given:

• Conventional System yearly cost= $128K (plus additional $60K energy costs)

• Cloud = $61K per year

• Quality Benefits• Less to Manage• Superior resiliency• Homogeneity ((eg, interoperability with in consortium)• Fewer issues to negotiate with institutional authorities (eg, firewall

protection, load issues, SLAs are more formal)

2727

Transition Obstacles (Rosenthal, et al 2010)

• Software portability (for critical applications)• Cloud unfamiliarity• Cloud immaturity• New management practices

• Transitioning to a cloud will change the ways in which biomedical systems are built, managed and funded. (eg, PIs need to expand skills in contracting, including service level guarantees and help facilities for developers)

• Models used for costing computational acquisitions need to be changed, to better reflect true costs (institutional level accounting need to change, too)