Showing posts with label Cloud Computing. Show all posts
Showing posts with label Cloud Computing. Show all posts

Tuesday, October 4, 2011

Infrastructure as a Service (Cloud Computing)

IAAS stands for “infrastructure as a service” which offers customers “on-demand” access to computational, storage and networking services. The setup, payment and use of these resources are easily scaled up and down as the consumer’s needs change. The boundaries between IAAS and PAAS are fluid, and will likely continue to change as the technology develops. IAAS is enabled by the technology of hardware virtualization, which allows computer servers to host multiple virtual machines (VMs for short) which each behave like an individual computer, capable of running their own operating systems. Customers begin by renting time from a IAAS provider such as Amazon EC2 or IBM Smart Cloud Enterprise. The customer specifies what type and how many VMs that they intend to lease and the lease period.

Upon payment, customers can immediately access their computing resources from an Internet connected web browser. If the customer needs more computing power, they can instruct their programs to allocate these resources automatically, allowing online companies to dramatically increase and reduce their IT infrastructure and expenditure. This ability to “auto-allocate” computing resources represents a revolutionary tool for online companies. From the onset, a firm can access exactly the IT resources it needs with negligible setup time.

Another contributor to IAAS is the lower cost of “commodity hardware” which can be installed in centralized data centers. This increase in computing capacity allows firms with large IT departments (such as Google or Amazon) to offer their excess capacity to customers as IAAS. Moore’s law, a statement on the ability of scientists and engineers to double the capacity of transistors in computer microprocessors every two years, has allowed the capability of hardware to increase and its cost to decrease.

The final keystone in the successful implementation of IAAS is the growth of the Internet itself, which bridges the gap between different hardware and software systems. It is the “central hub” bridging gaps between existing hardware infrastructure. Access to high speed Internet connections is a prerequisite to IAAS because it enables customers to remotely access their leased resources.

Customer access to IAAS is becoming increasingly easy. Hardware and software flexibility is a crucial part of IAAS. It allows customers to free themselves of technical concerns and focus on using their leased computing resources. As new computers become available, they can be installed in state of the art facilities of the IAAS providers. In this way, a crucial division of labor function has been performed. For limited projects where a quick calculation is required, a customer can lease a single hour of CPU time on the most sophisticated hardware available. With minimal difficulty customers can setup and provision their own customized “on-demand” IT department.

Saturday, September 24, 2011

Cloud Computing (Class Project)

  1. Halli's Thoughts:  Google Doc
  2. Final Document Link: https://www.wiki.ed.ac.uk/x/vootC

Cloud Computing (Operation Management Class Project)


Definition

    1. Cloud Computing is the term used to describe an internet based, distributed computing system where data is stored off site "in the cloud" rather than locally.  This means that data is not stored physically on computers which are owned by one's organization, but on other 3rd party server computers (known simply as "servers") connected to the internet.  In the previous IT model before cloud computing, firms managed their own IT infrastructure as a combination of hardware and software physically residing inside the organization, or in a data center owned or leased by the organization.  An example of some typical IT components are higher performance server computers running database, computational, and email hosting programs , and smaller desktops computers running word processing, spreadsheet and email client software.  This model, called the Client-Server model, was prevalent from the 1990's until the 2000's.  Cloud Computing, in contrast, is similar to treating computer resources as a "utility" such as electrical power that can be distributed to homes and firms, freeing them of the need to worry about managing the source themselves.  In the case of electricity, the source is AC generators, whereas with computing, the source is a series of servers which can reside in almost any location on the globe, limited only by network connectivity.  The internet bridges the link between these remote resources and the end user, who is not required to know the inner workings of the machines his programs are running on.  Further, the end user is freed from having to manually install or configure programs.  Because of the popularity of the term "Cloud Computing", considerable confusion can result as individual firms adapt their existing IT infrastructure to meet growing demand for such services.  However, the NIST definition of cloud computing involves four types of services, which are listed below.  Additionally, KPMG defined the cloud simply as:The Cloud =Internet-based data access & exchange + Internet-based access to low cost computing and applications".  Whichever definition is used, Cloud Computing is a trend in the development of Information Technology which simplifies access to computer resources and empowers users to access greater amounts of technology in a flexible way.
    2.  According to NIST, the four services which are considered Cloud Computing are:
      1. hardware (PIC=Andrea)
      2. infrastructure (infrastructure as a service, IAAS) (PICs=Eric, Petal)
        1. Functions of IAAS
          1. software development and testing.  Access to different operating systems, hardware
          2. computational computing:  Access to CPU time for making calculations.  Pay as you go system allows this to be very cost effective for SMEs
        2. Benefits of IAAS over alternatives
          1. "pay as you go" model. Just like cell phone service in the UK, you pay for what you need (hours of CPU time) number of "instances"
          2. scalable (de-scalable) as needed
          3. very attactive for software development
        3. Drawbacks of IAAS over alternatives
          1. limited customization, locked in to a certain provider's resources
          2. limited to options available at a given provider
        4. Providers of IAAS (Public and Private)
          1. Amazon EC2
          2. IBM Smart Cloud Enterprise
            1. data centers physically located in Toronto, Ehningen, Tokyo and Singapore
            2. support is offered along with IAAS
          3. University of Edinburgh ECDF- Educational Computing and Data Facility (both a provider and a consumer)
          4.  
        5. Customers of IAAS (Public and Private)
          1. Obama Cloud First Policy
          2.  
        6.  
        7.  
        8.  
      3. platform  (platform as a service) (PIC=Halli)
      4. software (software as a service) (PIC=Frank)
        1.  
        2.  
  1. Case Study:  Animoto 
    1. Company had a compelling product: a way to create user slideshows based on uploaded images and music.  The online slideshow combines images with the rhythm.
    2. Company experienced dramatic growth and needed its IT infrastructure to scale accordingly.
    3. Company was hosted on AWS (Amazon Web services) and this allowed the company to quickly double its IT capacity without investment in hardware, and without the need to configure software.
    4. Amazon had realized that it had excess server capacity and decided to offered the hosting service. 
  2. Case Study:  Cloudsoft
    1. Startup Company housed within the Informatics building of the University of Edinburgh
    2. Delivers tools such as Monterey to platform developers that allow applications to be run in parallel in a scalable fashion across multiple physical or virtual PCs
    3.  
  3. Virtualization has increased the ability of server farms to allow multiple paused machine instances to reside on one PC, thereby freeing up the resources (and allowing either cpu time sale or power off) 
  4. Security Concerns:  Because data is not stored securely at your own location, there is a greater risk of data loss or politically motivated espionage.  
    1. many companies are cautiously approaching "cloud computing" (especially in Japan) because of the challenges that it brings, particularly the lack of total control over an organization's IT infrastructure
    2. certain types of work is conducive to cloud computing, such as software development, some design work