CLOUD COMPUTING

[seminar class]

presented by:
Norman Wilde
Thomas Huber

---

[attachment=10282]
Virtualization and Cloud Computing
An opening caveat ...
 This talk is based on speeches at conferences, discussions with people in industry, and some experimentation.
 A lot of people think they will make a lot of money – so there is lots of hype!
 But there seems to be something fundamental going on.
 Two Technologies for Agility
Virtualization:
The ability to run multiple operating systems on a single physical system and share the underlying hardware resources*
Cloud Computing:
“The provisioning of services in a timely (near on instant), on-demand manner, to allow the scaling up and down of resources”**
 The Traditional Server Concept
 And if something goes wrong ...
 The Traditional Server Concept
 System Administrators often talk about servers as a whole unit that includes the hardware, the OS, the storage, and the applications.
 Servers are often referred to by their function i.e. the Exchange server, the SQL server, the File server, etc.
 If the File server fills up, or the Exchange server becomes overtaxed, then the System Administrators must add in a new server.
The Traditional Server Concept
 Unless there are multiple servers, if a service experiences a hardware failure, then the service is down.
 System Admins can implement clusters of servers to make them more fault tolerant. However, even clusters have limits on their scalability, and not all applications work in a clustered environment.
The Traditional Server Concept
 Pros
 Easy to conceptualize
 Fairly easy to deploy
 Easy to backup
 Virtually any application/service can be run from this type of setup
 Cons
 Expensive to acquire and maintain hardware
 Not very scalable
 Difficult to replicate
 Redundancy is difficult to implement
 Vulnerable to hardware outages
 In many cases, processor is under-utilized
 The Virtual Server Concept
 Close-up*
The Virtual Server Concept
 Virtual servers seek to encapsulate the server software away from the hardware
 This includes the OS, the applications, and the storage for that server.
 Servers end up as mere files stored on a physical box, or in enterprise storage.
 A virtual server can be serviced by one or more hosts, and one host may house more than one virtual server.
 The Virtual Server Concept
 Virtual servers can still be referred to by their function i.e. email server, database server, etc.
 If the environment is built correctly, virtual servers will not be affected by the loss of a host.
 Hosts may be removed and introduced almost at will to accommodate maintenance.
The Virtual Server Concept
 Virtual servers can be scaled out easily.
 If the administrators find that the resources supporting a virtual server are being taxed too much, they can adjust the amount of resources allocated to that virtual server
 Server templates can be created in a virtual environment to be used to create multiple, identical virtual servers
 Virtual servers themselves can be migrated from host to host almost at will.
The Virtual Server Concept
 Pros
 Resource pooling
 Highly redundant
 Highly available
 Rapidly deploy new servers
 Easy to deploy
 Reconfigurable while services are running
 Optimizes physical resources by doing more with less
 Cons
 Slightly harder to conceptualize
 Slightly more costly (must buy hardware, OS, Apps, and now the abstraction layer)
Virtualization Status
 Offerings from many companies
 e.g. VMware, Microsoft, Sun, ...
 Hardware support
 Fits well with the move to 64 bit (very large memories) multi-core (concurrency) processors.
 Intel VT (Virtualization Technology) provides hardware to support the Virtual Machine Monitor layer
 Virtualization is now a well-established technology
So what about Cloud Computing?
 Suppose you are Forbes.com
 You offer on-line real time stock market data
 Why pay for capacity weekends, overnight?
 Forbes' Solution
 Host the web site in Amazon's EC2 Elastic Compute Cloud
 Provision new servers every day, and deprovision them every night
 Pay just $0.10* per server per hour
 * more for higher capacity servers
 Let Amazon worry about the hardware!
 Cloud computing takes virtualization to the next step
 You don’t have to own the hardware
 You “rent” it as needed from a cloud
 There are public clouds
 e.g. Amazon EC2, and now many others (Microsoft, IBM, Sun, and others ...)
 A company can create a private one
 With more control over security, etc.
 Goal 1 – Cost Control
 Cost
 Many systems have variable demands
 Batch processing (e.g. New York Times)
 Web sites with peaks (e.g. Forbes)
 Startups with unknown demand (e.g. the Cash for Clunkers program)
 Reduce risk
 Don't need to buy hardware until you need it
 Goal 2 - Business Agility
 More than scalability - elasticity!
 Ely Lilly in rapidly changing health care business
 Used to take 3 - 4 months to give a department a server cluster, then they would hoard it!
 Using EC2, about 5 minutes!
 And they give it back when they are done!
 Scaling back is as important as scaling up
 Goal 3 - Stick to Our Business
 Most companies don't WANT to do system administration
 Forbes says:
 We are is a publishing company, not a software company
 But beware:
 Do you really save much on sys admin?
 You don't have the hardware, but you still need to manage the OS!
 How Cloud Computing Works
 Various providers let you create virtual servers
 Set up an account, perhaps just with a credit card
 You create virtual servers ("virtualization")
 Choose the OS and software each "instance" will have
 It will run on a large server farm located somewhere
 You can instantiate more on a few minutes' notice
 You can shut down instances in a minute or so
 They send you a bill for what you use
 Any Nasty Details?
(loads of them!)
 How do I pick a provider?
 Am I locked in to a provider?
 Where do I put my data?
 What happens to my data when I shut down?
 How do I log in to my server?
 How do I keep others from logging in (security)?
 How do I get an IP address?
 Etc.
 And One Really Important Caveat*
 (footnote)
How come Amazon?
 Grew out of efforts to manage Amazon’s own services
 (Each time you get a page from Amazon, over a hundred servers are involved)
 See reference Amazon Architecture on their service design concepts
 They got so good at it that they launched Amazon Web Services (AWS) as a product
 Cloud Computing Status
 Seems to be rapidly becoming a mainstream practice
 Numerous providers
 Amazon EC2 imitators ...
 Just about every major industry name
 IBM, Sun, Microsoft, ...
 Major buzz at industry meetings
 So What’s the Take-Away?
 There seems to be a major revolution underway in how we manage hardware
 Specify (machine per service or one big machine with many virtual servers
 Purchase (own it yourself or rent from a public cloud)
 Use (always-on, or flexible provisioning as needed ...)
 We may need to rethink both our research and teaching
For UWF:
What About Research?
 The Eucalyptus Project
 From University of California Santa Barbara
 An open source collection of tools to build your own cloud
 Linux using Xen for virtualization
 An apparently open research area: handling data
 Regular databases apparently don't scale well
 Especially hard to make elastic (scale up / scale down)
For UWF:
What About Teaching?
 Our graduates should know about cloud computing / virtualization
 It will be useful for some applications, though not for all
 But what are the right learning objectives?
 Awareness (its there ...)
 Mechanics (here’s how to instantiate a server ...)
 Design (how to make a scalable service ...)
 ???
For Fall 2009 ...
 Currently developing a Virtualization / Cloud Computing “module”
 1 – 2 class sessions plus an exercise
 Target courses (November):
 COP 6990 – Multi-Process Computing (Simmons)
 CTS 4817 – Web Server Administration (Owsnicki-Klewe)
 Objectives
 Awareness and mechanics of EC2