Cloud Computing

Cloud Computing Case Study

 At one extreme, sharing basic IT infrastructure
 E.g Amazon’s EC2: an EC2 instance appears physical HW, users
can control nearly the entire sw stack, from the kernel upwards
 At the other extreme, sharing application domain-specific
platforms
 E.g. Google AppEngine, a platform for building and hosting web
applications on Google Web servers. Limited range of applications
are supported,
 In the middle, sharing programming environments
 Microsoft’s Azure written using .NET libraries, and compiled to
the Common Langurage Runtime (CLR).

Cloud Computing 2
Amazon Web Services
 AWS is a collection of remote computing services that together make up a
cloud computing platform, offered over the Internet by Amazon.com.
 List of AWS products
 Amazon Elastic Compute Cloud (EC2) provides scalable virtual private servers
using Xen.
 Amazon Simple Storage Service (S3) provides Web Service based storage.
 Amazon SimpleDB allows developers to run queries on structured data. It provide
"the core functionality of a database."
 Amazon Elastic MapReduce allows developers to easily and cheaply process vast
amounts of data. It uses a hosted Hadoop framework running on the web-scale
infrastructure of EC2 and Amazon S3.
 Amazon Virtual Private Cloud (VPC) creates a logically isolated set of Amazon EC2
instances which can be connected to an existing network using a VPN connection.
 And More…

Cloud Computing 3
Amazon’s EC2
 Amazon Elastic Compute Cloud (EC2)
 Web service that provides resizable compute capacity in the cloud

 An EC2 instance appears physical HW, provides users

complete control over nearly entire sw stack, from the
kernel upwards
 Load Variety of operating system
 Install Custom applications
 Manage network access permission
 Run image using as many/few systems as you desire

C. Xu @ Wayne State Cloud Computing 4

Amazon’s EC2 features
 Elastic capacity
 Elastic resource config/reconfig; Elastic num of instances
 Completely Control
 Root access/access to console output/data store/ reboot
 Reliable
 Multiple locations
 Elastic IP addresses
 Secure
 Firewall config
 Virtual Private Cloud
 Performance
 Auto Scaling
 Auto local balancing

Cloud Computing 5
Amazon’s EC2 Instances
 On-Demand Instances
 Pay for capacity without long-term commitment

 Reserved Instances
 Standard Instances
 Micro Instances
 High-Memory Instances
 High-CPU Instances
 High-I/O instance
 High Storage Instances

 Spot Instances
 Bit on unused Amazon EC2 capacity, run those instances for as long as
their bid exceeds the current Spot Prices

C. Xu @ Wayne State Cloud Computing 6

Amazon’s EC2 Usages
 Create an Amazon Machine Image (AMI) containing your
applications, libraries, data and associated configuration settings. Or
use pre-configured, templated images to get up and running
immediately.
 Choose the types of instances and OS, then start, terminate, and
monitor as many instances of your AMI as needed, using the web
service APIs or the variety of management tools provided.
 Determine whether you want to run in multiple locations, utilize static
IP endpoints, or attach persistent block storage to your instances.
 Pay only for the resources that you actually consume, like instance-
hours or data transfer.
 http://www.youtube.com/watch?v=JPFoDnjR8e8

Cloud Computing 7
Amazon’s S3
 Amazon Simple Storage Service (S3)
 Storage for the Internet.

 Features
 Unlimited Storage
 Highly scalable
• in terms of storage, request rate and concurrent users
 Reliable
• Store redundant data in multiple facilities and on multiple devices
 Secure
• Flexibility to control who/how/when/where to access the data
 Performance
• Choose region to optimize for latency/minimize costs

 Work with other AWS products

 EC2/Elastic MR/Amazon Import/Export…

C. Xu @ Wayne State Cloud Computing 8

Example : online photo processing service
 Photo operation
 red eye reduction/cropping/customization/re-coloring/teeth whitening, etc

 Procedure
 Web server receive request
 Put request message in the queue
 Pictures stored in S3
 Multiple EC2 instances run photo processing
 Put back in the queue
 Return

Cloud Computing 9
Google’s AppEngine
 A platform for building and hosting web applications on
Google data servers.
 support Python/Java
 Upload and ready to run
 No need to maintain your own servers, easy to scale as your traffic
and storage grow
 Kevin Gibbs, App Engine Tech Lead
“AppEngine is a system that exposes various pieces of
Google’s scalable infrastructure so that you can write
server-side applications on top”

Cloud Computing 10
AppEngine Does One Thing Well
 AppEngine handles HTTP(S) requests, nothing else
 Think RPC: request in, processing, response out
 Dynamic web serving, with full support for common web technologies
 App configuration is dead simple
 No performance tuning needed
 Everything is built to scale
 “infinite” number of apps, requests/sec, storage capacity
 Automatic scaling
 Automatic management
 Load balancing/monitoring/scaling

 Local development Environment: simulation

 Google AppEngine Software Development Kit

Cloud Computing 11
AppEngine Service
 URLFetch
fetch web resources/services
 Images
 manipulate images: resize, rotate, flip, crop
 Google Accounts
 Allow users to sign in their product by using G account
 Mail
 Send message using Google infrastructure.
 XMPP
 send instant messages
 Datastore
 managing data objects

 Memcache
 Distributed in-memory data cache in front of constant storage

Cloud Computing 12
App Enginer Sandbox
 Appl run in a secure environment that provides
limited access to the underlying OS.
 Allow AppEngine to distribute web requests for the
application across multiple servers; Start and stop
servers to meet traffic demands
 Isolates your appl in its own secure, reliable
environment that is independent of the hw, OS, and
physical location of the web server

Cloud Computing 13
App Enginer Sandbox
 Limited Access:
 Appl can only access other computers on the Internet
through provided URL fetch and email API
 Cannot write to file system, can read file uploaded by
the appl; The app must use the App Engine datastore
for all data that persists between requests
 Run in response to a web request, and must return
response data within 30 seconds. A request handler
cannot spawn a sub-process or execute code afte the
response has been sent.

Cloud Computing 14
App Engine Architecture (python)
req/resp
stateless APIs

urlfech Python stdlib

mail VM
process app
images

stateful datastore
APIs memcache

Cloud Computing 15
App Engine Architecture

Cloud Computing 16
Why Not LAMP?
 Linux, Apache, MySQL/PostgreSQL,
Python/Perl/PHP/Ruby
 LAMP is the industry standard
 But management is a hassle:
 Configuration, tuning
 Backup and recovery, disk space management
 Hardware failures, system crashes
 Software updates, security patches
 Log rotation, cron jobs, and much more

More Important…

Cloud Computing 17
Automatic Scaling
 Low-usage apps: many apps per physical host
 High-usage apps: multiple physical hosts per app

 Stateless APIs are trivial to replicate

 Datastore built on top of Bigtable; designed to scale well

 Abstraction on top of Bigtable

Cloud Computing 18
Automatic Scaling
 You don’t need to configure your resource needs
 One CPU can handle many requests per second
 Apps are hashed onto CPUs:
One process per app, many apps per CPU
 Creating a new process is a matter of cloning a generic “model” process
and then loading the application code (in fact the clones are pre-created
and sit in a queue)
 The process hangs around to handle more requests (reuse)
 Eventually old processes are killed (recycle)
 Busy apps (many QPS query per sec) get assigned to multiple CPUs
 This automatically adapts to the need
• as long as CPUs are available

Cloud Computing 19
 AppEngine’s Resource Limit
Quota Limit
Apps per developer 10
Time per request 30 sec
Quota Free Quota
Files per app 1,000 Emails per day 2,000
HTTP response size 10 MB BW in per day 10,000 MB

Datastore item size 1 MB BW out per day 10,000 MB

Secure BW in per day 2,000 MB
Appl code size 150 MB
Secure BW out per day 2,000 MB
CPU megacycles per day 200,000,000
HTTP Requests per Day 1,333,328*
Datastore API calls per day 10,368,000*
Data stored 1 GB

URLFetch API calls per day 657,084*

Cloud Computing 20
AppEngine Development Demo
 http://www.youtube.com/watch?v=tcbpTQXNwac

Cloud Computing 21
Microsoft’s Azure
 Cloud service platform hosted in MS data centers,
including an OS, Windows Azure, and a set of developer
services like .NET services, Live services, and SQL data
services
 Window Azure: serving as the development, service hosting, data
storage and service management environment

Cloud Computing 22
Components of Azure
Compute，Storage ，Fabric Controller， Content Delivery
Network (CDN)， Connect

Cloud Computing 23
Compute
 Web role: web-based application
 Worker role: various windows-based code
 VM role: Windows Server image, customized software installations on
OS will be kept.

Cloud Computing 24
Compute
 Submit an application with configuration of type and
number of instances
 Fabric controller creates VM and runs the code
 Requests through HTTP/HTTPS/TCP are load balanced
across all instances without affinity.
 Exposed API that allows
 automatically scale up and down.
 application logging
 Monitoring resource usage: CPU/storage

Cloud Computing 25
Storage: Blob, Table and Queue
 Blob
 Simple unstructured hierarchy: each container holds one or more blobs,
which is up to terabyte
 Access from URI:
http://<StorageAccount>.queue.core.windows.net/<QueueName>
 Basic storage

Cloud Computing 26
Storage: Blob, Table and Queue
 Tables
 Structured storage.
 Stored in a group of entities that contain properties.
 Entity with different num of properties in various types.
 Not relational
tables, support
massively scalable
applications.

Cloud Computing 27
Storage: Blob, Table and Queue
 Queues
 Provide a way for communication between instances

Three replications for all kinds

Cloud Computing 28
Fabric Controller
 All Azure application and data reside in Microsoft data
center
 All Azure components are managed by the fabric
controller.

Cloud Computing 29
Fabric Controller
 Distributed application that’s replicated across
machines.
 It owns all of the resources: computer, switches,
load balancers and is aware of each azure
application
 Autonomic management
 Monitoring the system status
 Optimizing hardware utilization
 Create VMs and intelligent deployment
 Update instances without shutdown
 Managing the OS for Web and Work roles

Cloud Computing 30
Content Delivery Network
 The CDN stores copies of a blob at sites closer to the
clients that use it
 Servers store videos, Flash, HTML5…

Cloud Computing 31
Connect
 Effectively connecting on-premises environments with Windows
Azure
 Installing a Endpoint Agent on each on-premises machines and
communication through IP-level connection
 Azure applications appears to be on the same IP network as the on-
premises machines.

Cloud Computing 32
Case Study 1
 Large scale web applications with occasional huge spikes
and background processing.
 Video sharing site

 Deployment
 A number of web instances
based on demand
 Table storage
for information
 Many works for
processing
 Blobs storage for
large data set

Cloud Computing 33
Case Study 2
 Parallel processing applications
 Financial modeling at a bank
 New drug testing simulations in a pharmaceutical company

 Deployment
 Web role for access interface
 Many workers
for processing
 Large data set
stored in blobs

Cloud Computing 34
Case Study 3
 Using storage from an on-premises or hosted
application
 Archive old email
 User log file

 Deployment
 Connect on-premises
application with Azure

Cloud Computing 35
Case Study

Cloud Computing 36