ELL 887 - CLOUD COMPUTING

Cloud Enabling Technologies

 2

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology
 3

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology
 4

Internet
 Cloud consumers and cloud providers
typically use the Internet to
communicate
 Internet is based on a decentralized
provisioning and management model
• Not controlled by any centralized
entities
 Established and deployed by Internet
Service Providers (ISPs), the
Internet’s largest backbone networks
are interconnected by core routers
that connect the world’s multinational
networks.
 Government and regulatory laws
dictate the service provisioning
conditions of organizations and ISPs

Internetworking structure of the internet

 5

Internet Service Providers

 The Internet’s topology has
become a dynamic and complex
aggregate of ISPs that are highly
interconnected
• Tier 1: Large scale international cloud
providers
• Tier 2: Large regional providers
• Tier 3: Local ISPs

 Cloud providers and consumers can

connect directly using a Tier 1
provider
Worldwide connectivity is  Communication path that connects
enabled thru a hierarchical a cloud provider and consumer may
topology involve multiple ISP networks.
 6

Components of Internetworking Architecture

 Connectionless Packet Switching
(Datagram Networks)
• End to end data flows are divided into
packets carrying the necessary location
information
• Received and processed thru network
switches and routers
• Queued and forwarded from one
intermediary node to the next
 Routers
• Device that is connected to multiple
networks thru which it forwards packets Packets travelling thru the internet are
directed by a router that arranges them
• Routers process and forward each packet
into a message
individually
• It determines the most efficient hop for
packet delivery
 7

Internet Reference Model and Protocol Stack

• IP packets are transferred thru the underlying physical networks

• Transport layer protocols provides standardized end-to-end communication
support that facilitates the navigation of data packets
• Application layer protocols use transport layer protocols to standardize and
enable specific data packet transferring methods
 8

Internetworking Architecture of a Private Cloud

• The physical IT resources that constitute the Cloud are located and managed
within the organization.
• Enterprise applications and various IT solutions are hosted on centralized
servers and storage devices in the organization’s data center
• End user devices access the data center thru the corporate network
• TCP/IP facilitates both internet access and on-premise data exchange over
LANs
• Organizations have complete control over and can safeguard their corporate
networks.
 9

Internetworking Architecture of an Internet based Cloud Deployment model

The internet is the connecting agent between the cloud consumers,

roaming end users and the cloud providers own network
 10

Internetworking Comparison
On Premise IT Resources Cloud based IT Resources

• Internal end user devices • Internal end user devices

access corporate IT services access corporate IT services
thru the corporate network thru an internet connection
• Internal users access corporate • Internal users access corporate
IT services thru the corporate IT services thru the Cloud
internet connection while provider’s internet connection
roaming in external network while roaming in external
network
• External users access
corporate IT services thru the • External users access
corporate internet connection corporate IT services thru the
Cloud provider’s internet
connection
 11

Network Bandwidth
• End-to-end bandwidth is determined by:
• Bandwidth of the data link that connects network to ISPs
• Transmission capacity of the shared data links that connect intermediary
nodes
 ISPs need to use broadband technology to implement the core
network required to guarantee end-to-end connectivity
 Bandwidth is critical for applications that require large amounts of
data to be transferred to and from the cloud
 12

Latency
• Latency is the time required by a packet to travel from one data node
to another
• Increases with every intermediary node
• Transmission queues in the network infrastructure can result in heavy
load conditions that also increase network latency
• Internet latency is highly variable since networks are dependent on
traffic conditions in shared nodes
• Packet networks with best effort QoS typically transmit traffic on a
first come first serve basis
• Using congested network paths results in service level degradation.
• Latency is critical for applications that require swift response time.
 13

Cloud Carrier
 The service levels of Internet connections between cloud consumers
and providers and determined by their ISPs which are usually
different
• Include multiple ISP networks in their paths
 QoS management across multiple ISPs is difficult to achieve
• Requires collaboration of the cloud carriers on both sides
 Cloud consumers and providers may need to use multiple cloud
carriers to achieve the necessary level of reliability
 14

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology
 15

Data Centers
 Advantages of grouping IT resources in a data center:
• Power sharing
• Higher efficiency in shared IT resource usage
• Improved accessibility for IT personnel
 16

Data Center features

 Virtualization
• Physical and virtualized IT resources
 Standardization and Modularity
• Built upon standardized commodity hardware
• Designed with modular architectures
• Aggregating multiple identical building blocks of facility infrastructure and equipment
• Key requirements for reducing investment and operational costs enabling economy of scale
 Automation
• Automate tasks like provisioning, configuration, patching and monitoring without supervision
 Remote Operation and Management
• Most of the operation and administrative tasks are commanded thru the network’s remote control and
management systems
 High Availability
• Designed to operate with redundancy to sustain availability
• Outages impact business continuity
 Security
• Data centers store and process business data
• Housing data exclusively at one data center facility means much more can be compromised by a
successful security incursion
• Requirements for security such as physical and logical access controls and data recovery strategies
needed
 17

Data Center components

 18

Facilities – Computing Hardware

• Heavy processing in data centers is executed by standardized
commodity servers that have substantial computing power and
storage capacity.
• Computing hardware technologies are integrated into module servers
such as:
• Rackmount form factor server design composed of standardized racks with
interconnects for power, network and internal cooling
• Support for different hardware processing architecture
• Power efficient multi-core CPU architecture that houses hundreds of
processing cores in a space for a single unit of standardized rack
• Redundant and hot-swappable components
• Support industry-standard operational and management software
systems that configure, monitor and control hardware IT resources
from remote management consoles
 19

Server Network Connection in a Data center (Example)

 20

Data Center H/W - Storage

 Data centers have specialized storage systems that store enormous
amount of information
 Involve the following technologies:
• Hard disk arrays:
− Storage systems are containers housing numerous hard disks that are organized
into arrays
− These arrays divide and replicate data among multiple physical devices
− Increase performance and redundancy by including spare disks
− This technology implemented using Redundant Arrays of Independent Disks
(RAID) schemes implemented through hardware disk controllers
• I/O caching: Hard disk array controllers enhance disk access times and
performance by data caching
• Hot-swappable hard disks: Can be safely removed without powering down
• Storage virtualization
• Fast data replication mechanisms include:
− Snapshotting: Saving a VM’s memory into hypervisor readable file for future
reloading
− Volume Cloning: Copying virtual or physical hard disk volumes
 21

Types of Storage
 Direct-attached Storage (DAS): Storage system directly connected to
the computing IT resource using a Host Bus Adapter (HBA).
 Networked Storage Devices: Storage system connected to one or
more IT resources through a network
• Storage Area Network (SAN):
− Physical data storage media are connected thru a dedicated network (fiber
channel)
− Provide block-level data storage
− Appears to OS as an attached storage
− Expensive
• Network Attached Storage (NAS):
− Network-attached storage is a file-level computer data storage server
− Connected to a computer network using ethernet
− Appears to OS as a shared folder
− Inexpensive
 Storage systems provides tertiary redundancies using removable
media such as robotized tape libraries
• Used as backup and recovery systems.
 22

Storage System Network Connection in a Data center

(Example)
 23

Data Center H/W - Network

 Data centers require extensive network hardware to enable multiple levels of connectivity
 Requirement for scalability and high availability fulfilled by redundant, fault-tolerant
configurations
 Data Center broken down into 5 network subsystems:
I. Carrier and External Network Interconnection:
• Comprised of backbone routers that provide routing between external WAN connections and the data center’s
LAN as well as perimeter network security devices
II. Web-tier Load Balancing and Acceleration:
• Consists of Web acceleration devices that perform content aware routing
III. LAN Fabric:
• Constitutes the internal LAN
• Provides high-performance and redundant connectivity for all of the data center’s network enabled IT resources.
• Implemented with multiple network switches that operate at speeds upto 10GB/sec
• Can also perform several virtualization functions
IV. SAN Fabric:
• Related to the implementation of SAN that provide connectivity between servers and storage systems
• Implemented with Fibre Channel (FC), Fiber Channel over Ethernet (FCoE) and Infiband Network switches
V. NAS Gateways:
• Supplies attachment points for NAS based storage devices
• Implements protocol conversion h/w that facilitates data transmission between SAN and NAS devices.

 Ultra high-speed network optical links can be used to aggregate individual gigabit-per
second channels into single optical fibers using multiplexing technologies.
 24

Internetworking between 2 data centers

(Example)
 25

Technological Obsolescence
• IT hardware lifecycles typically lasts between 5-7 years.
• The on-going need to replace equipment frequently results in a mix of
hardware whose heterogeneity can complicate data center’s
operations.
 26

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology
 27

Web Technology
 Web technology is generally used both as the implementation
medium and management interface for cloud services because:
• Cloud computing relies on internetworking
• Web browser universality
• Ease of web-based service development
 WWW is a connection of interlinked IT resources that are accessed
thru the internet
 Uniform Resource Locator (URL): Standard syntax of identifiers for
Web-based resources
 Hypertext Transfer Protocol (HTTP): Primary communication protocol
for transmitting URL in the WWW.
 Markup Languages (HTML, XML): A lightweight means of specifying
Web-based data and metadata
 28

Web Applications
 A distributed application that
uses Web-based technologies
• Presentation Layer: User
interface
• Web servers interact with
application servers to execute
the requested application logic
• Typically involves interaction with
the data layer that is comprised
of data stores

•
3 tier Web application
 29

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Virtualization Technology
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology
 30

Multitenant Application
 Enables multiple users (tenant) to access the same application logic
simultaneously.
 Each tenant has a dedicated instance of the software.
 Tenants can individually customize features of the application such
as:
• User interface
• Business process
• Data model
• Access control
 Multi-tenant application architecture is more complex than a single-
tenant application
• Need to support sharing of various artifacts among multiple users
 31

Multitenant Application Characteristics

• Usage isolation: One user usage does not affect availability and
performance of other users
• Data security: Tenants cannot access data of another tenant
• Recovery: Backup and recovery procedures are separately executed
• Application upgrades: Tenants not affected by the synchronous
upgrade of shared software artifacts
• Scalability: Application can scale by increase in usage by existing
tenants or increase in number of tenants
• Metered usage: Tenants charged only for their own consumption
• Data tier isolation: Databases isolated among tenants. In some cases
can be designed to be shared among tenants
 32

Multitenant Application

Multitenant application serving multiple cloud service consumers

 33

Multitenancy vs Virtualization
Virtualization Multitenancy

• Multiple virtual copies of the server • A physical or virtual server hosting

environment can be hosted by a an application that can be used by
single physical server different users
• Each copy can be provided to • Each user feels as though they
different users, configured have exclusive usage of the
independently and can contain application
their own OS and applications
 34

Overview
• Virtualization
• Containerization
• Broadband Network and Internet Architecture
• Virtualization Technology
• Data Center Technology
• Web Technology
• Multitenant Technology
• Service Technology (Basis of as-a-service Cloud delivery
models)
 35

Web Services
 Represent an established and common
medium for Web-based service logic.
 Core technologies for Web services are
represented by the following industry
standards:
• Web Service Description Language
(WSDL): Defines the API of a Web
service
• XML Schema Definition Language:
XML schemas are created to define
the data structure of XML based
messages exchanged by Web
services.
• SOAP: Defines a common
messaging format used for request
and response messages exchanged
by Web services.
• Universal Description, Discovery and
Integration (UDDI): Regulates service
registries in which WSDL definitions
can be published as part of service
catalog for discovery purposes.
 36

REST Services
 Service implementations that rely on the use of core Web technologies
 REST services do not have individual technical interfaces
 Share a common technical interface that is known as the uniform contract
which is typically established by the use of HTTP methods
 Design constraints:
• Client server: Client and server applications must be able to evolve separately
without any dependency on each other.
• Stateless: The server will not store anything about the latest HTTP request the client
made. It will treat every request as new. No session, no history.
• Cache: Caching shall be applied to resources when applicable, and then these
resources must declare themselves cacheable. Caching can be implemented on the
server or client side.
• Uniform Interface: All resources should be accessible through a common approach
such as HTTP GET and similarly modified using a consistent approach.
• Layered System: REST allows use of a layered system architecture where you
deploy the APIs, store data and authenticate requests in separate systems.
• Code-on-Demand (Optional): Can return executable code to support a part of
the application
 37

Service Agents
 Event driven programs designed to intercept messages at runtime.
 Two types are common in Cloud environments
• Active:
− Perform an action upon intercepting and reading the contents of a message
− Action typically requires making changes to the message content or changes to
the message path
• Passive:
− Do not change message content
− Read the message and capture certain parts of the message for monitoring,
logging or reporting
 Cloud-based environments rely heavily on the use of system-level or
custom service agents to perform runtime monitoring required to
ensure features like elastic scaling and pay for use billing can be
carried out instantaneously.
 38

Service Middleware
 Two most common types of middleware platforms relevant for
services computing are:
• Enterprise Service Bus (ESB): Encompasses a range of intermediary
processing features including service brokerage, routing and message
queuing.
• Orchestration Platform: Designed to host and execute workflow logic that
drives service composition.
 39

Readings
• Cloud Computing: Concepts, Technology & Architecture. Chapter 5