Cloud Computing Patterns
[Link]
Cloud Computing Patterns
Patterns are a widely used concept in computer
science to describe good solutions to reoccurring
problems in an abstract form. Such conceptual
solutions can then be applied in concrete use cases
regardless of used technologies, such as software,
middleware, or programming languages. We employ
patterns to describe cloud service models and cloud
deployment types in an abstract form to categorize
the offerings of cloud providers. Furthermore, we
give reoccurring cloud application architectural
patterns on how to design, build, and manage
applications that use these cloud offerings. The
abstraction of these patterns make them applicable
to challenges faced by developers regardless of the
actual technologies and cloud services that they are
using.
Cloud Computing Patterns – Overview
[Link] 2
Cloud Computing Fundamentals
Application Workloads
[Link]
Static Workload
IT resources with an equal utilization over time experience static workload.
How can an equal utilization be characterized and how can
applications experiencing this workload benefit from cloud
computing?
Fundamental Patterns – Application Workloads
[Link] 4
Periodic Workload
IT resources with a peaking utilization at reoccurring time intervals experience
periodic workload.
How can an equal utilization be characterized and how can
applications experiencing this workload benefit from cloud
computing?
Fundamental Patterns – Application Workloads
[Link] 5
Once-in-a-lifetime Workload
IT resources with an equal utilization over time disturbed by a strong peak occurring
only once experience once-in-a-lifetime workload.
How can equal utilization with a one-time peak be characterized and
how can applications experiencing this workload benefit from cloud
computing?
Fundamental Patterns – Application Workloads
[Link] 6
Unpredictable Workload
IT resources with a random and unforeseeable utilization over time experience
unpredictable workload.
How can random and unforeseeable utilization be characterized and
how can applications experiencing this workload benefit from cloud
computing?
Fundamental Patterns – Application Workloads
[Link] 7
Continuously Changing Workload
IT resources with a utilization that grows or shrinks constantly over time experience
continuously changing workload.
How can a continuous growth or decline in utilization be characterized
and how can applications experiencing this workload benefit from
cloud computing?
Fundamental Patterns – Application Workloads
[Link] 8
Cloud Computing Fundamentals
Cloud Service Models
[Link]
Infrastructure as a Service (IaaS)
Providers share physical and virtual hardware IT resources between customers to
enable self-service, rapid elasticity, and pay-per-use pricing.
How can different customers share a physical hosting environment so
that it can be used on-demand with a pay-per-use pricing model?
Fundamental Patterns – Cloud Service Models
[Link] 10
Platform as a Service (PaaS)
Providers share IT resources providing an application hosting environment between
customers to enable self-service, rapid elasticity, and pay-per-use pricing.
How can custom applications of the same customer or different
customers share an execution environment so that it can be used on-
demand with a pay-per-use pricing model?
Fundamental Patterns – Cloud Service Models
[Link] 11
Software as a Service (SaaS)
Providers share IT resources providing human-usable application software between
customers to enable self-service, rapid elasticity, and pay-per-use pricing.
How can customers share a provider-supplied software application so
that it can be used on-demand with a pay-per-use pricing model?
Fundamental Patterns – Cloud Service Models
[Link] 12
Cloud Computing Fundamentals
Cloud Deployment Models
[Link]
Public Cloud
IT resources are provided as a service to a very large customer group in order to
enable elastic use of a static resource pool.
How can the cloud properties be provided to a large customer group?
Fundamental Patterns – Cloud Deployment Models
[Link] 14
Private Cloud
IT resources are provided as a service exclusively to one customer to meet require-
ments on privacy, security, and trust while enabling elastic use of a static resources.
How can the cloud properties be provided in environments with high
privacy, security and trust requirements?
Fundamental Patterns – Cloud Deployment Models
[Link] 15
Community Cloud
IT resources are provided as a service to a group of customers trusting each other in
order to enable collaborative elastic use of a static resource pool.
How can the cloud properties be provided exclusively to a group of
customers forming a community of trust?
Fundamental Patterns – Cloud Deployment Models
[Link] 16
Hybrid Cloud
Different clouds and static data centers are integrated to form a homogeneous
hosting environment.
How can the cloud properties be provided across clouds and other
environments?
Fundamental Patterns – Cloud Deployment Models
[Link] 17
Cloud Offering Patterns
Cloud Environments
[Link]
Elastic Infrastructure
Hosting of virtual servers, disk storage, and configuration of network connectivity is
offered via a self-service interface over a network.
How do cloud offerings providing infrastructure resources behave and
how should they be used in applications?
Cloud Offering Patterns – Cloud Environments
[Link] 19
Elastic Platform
Middleware for the execution of custom applications, their communication, and
data storage is offered via a self-service interface over a network.
How do cloud offerings providing execution environments behave and
how should they be used in applications?
Cloud Offering Patterns – Cloud Environments
[Link] 20
Node-based Availability
A cloud provider guarantees the availability of individual nodes, such as individual
virtual servers, middleware components or hosted application components.
How can providers express availability in a node-centric fashion, so
that customers may estimate the availability of hosted applications?
Cloud Offering Patterns – Cloud Environments
[Link] 21
Environment-based Availability
A cloud provider guarantees the availability of the environment hosting individual
nodes, such as virtual servers or hosted application components.
How can providers express availability in an environmental-centric
fashion, so that customers may estimate the availability of hosted
applications?
Cloud Offering Patterns – Cloud Environments
[Link] 22
Cloud Offering Patterns
Processing Offerings
[Link]
Hypervisor
To enable the elasticity of clouds, the time required to provision and decommission
servers is reduced through hardware virtualization.
How can virtual hardware that has been abstracted from physical
hardware be used in applications?
Cloud Offering Patterns – Processing Offerings
[Link] 24
Execution Environment
To avoid duplicate implementation of functionality, applications are deployed to a
hosting environment providing middleware services and common functionality.
How can multiple application components share a hosting
environment efficiently?
Cloud Offering Patterns – Processing Offerings
[Link] 25
Map Reduce
Large data sets to be processed are divided into smaller data chunks and distributed
among processing application components. Individual results are later consolidated.
How can the performance of complex processing of large data sets be
increased through scaling out?
Cloud Offering Patterns – Processing Offerings
[Link] 26
Cloud Offering Patterns
Storage Offerings
[Link]
Block Storage
Centralized storage is integrated into servers as a local hard drive to enable access to
this storage via the local file system.
How can central storage be accessed as a local drive by servers and
hosted applications?
Cloud Offering Patterns – Storage Offerings
[Link] 28
Blob Storage
Data is provided in form of large files that are made available in a file system-like
fashion.
How can large files be stored, organized, and made available over a
network?
Cloud Offering Patterns – Storage Offerings
[Link] 29
Relational Database
Data is structured according to a schema that is enforced during data manipulation
and enables expressive queries of handled data.
How can data elements be stored so that relations between them can
be expressed and expressive queries are enabled to retrieve required
information effectively?
Cloud Offering Patterns – Storage Offerings
[Link] 30
Key-value Storage
Semi-structured or unstructured data is stored with limited querying support but
high-performance, availability, and flexibility.
How can key-value elements be stored to support scale out and an
adjustable data structure?
Cloud Offering Patterns – Storage Offerings
[Link] 31
Strict Consistency
Data is stored at different locations to improve response time and to avoid data loss
in case of failures while consistency of replicas is ensured at all times.
How can data be distributed among replicas to increase availability,
while ensuring data consistency at all times?
n<w+r
Cloud Offering Patterns – Storage Offerings
[Link] 32
Eventual Consistency
Performance and availability of data in case of network partitioning are enabled by
ensuring data consistency eventually and not at all times.
How can data be distributed among replicas with focus on increased
availability and performance, while being resilient towards
connectivity problems?
Cloud Offering Patterns – Storage Offerings
[Link] 33
Cloud Offering Patterns
Communication Offerings
[Link]
Virtual Networking
Networking resources are virtualized to empower customers to configure networks,
firewalls, and remote access using a self-service interface.
How can network connectivity between IT resources hosted in a cloud
be configured dynamically and on-demand?
Cloud Offering Patterns – Communication Offerings
[Link] 35
Message-oriented Middleware
Asynchronous communication is provided while hiding complexity of addressing,
routing, or data formats to make interaction robust and flexible.
How can communication partners exchange information
asynchronously with a communication partner?
Cloud Offering Patterns – Communication Offerings
[Link] 36
Exactly-once Delivery
The messaging system ensures that each message is delivered exactly once by
filtering possible message duplicates automatically.
How can it be assured that a message is delivered only exactly-once to
a receiver?
Cloud Offering Patterns – Communication Offerings
[Link] 37
At-least-once Delivery
In case of failures that lead to message loss, messages are retransmitted to assure
they are delivered at least once.
How can a message-oriented middleware ensure that messages are
received successfully at least once?
Cloud Offering Patterns – Communication Offerings
[Link] 38
Transaction-based Delivery
Clients retrieve messages under a transactional context to ensure that messages are
received by a handling component.
How can it be ensured that messages are only deleted from a message
queue if they have been received successfully?
Cloud Offering Patterns – Communication Offerings
[Link] 39
Timeout-based Delivery
Clients acknowledge message receptions to ensure that messages are received
properly.
How can it be ensured that messages are only deleted from a message
queue if they have been received successfully at least once?
Cloud Offering Patterns – Communication Offerings
[Link] 40
Cloud Application Architecture Patterns
Fundamental Cloud Architectures
[Link]
Loose Coupling
A broker encapsulates concerns of communication partner location, implementation
platform, time of communication, and data format.
How can dependencies between distributed applications and between
individual components of these applications be reduced?
Cloud Application Architecture Patterns – Fundamental Cloud Architectures
[Link] 42
Distributed Application
A cloud application divides provided functionality among multiple application
components that can be scaled out independently.
How can application functionality be decomposed to be handled by
separate application components?
Cloud Application Architecture Patterns – Fundamental Cloud Architectures
[Link] 43
Cloud Application Architecture Patterns
Cloud Application Components
[Link]
Stateful Component
Multiple instances of a scaled-out application component synchronize their internal
state to provide a unified behavior.
How can applications components that are scaled-out maintain a
synchronized internal state?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 45
Stateless Component
State is handled external of application components to ease their scaling-out and to
make the application more tolerant to component failures.
How can elasticity and robustness of an application component be
increased?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 46
User Interface Component
Customizable synchronous user interfaces are accessed by humans, while
application-internal interaction is realized asynchronously to ensure loose coupling.
How can user interface components be accessed interactively by
humans while being configurable and decoupled from the remaining
application?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 47
Processing Component
Processing functionality is handled by elastically scaled components. Functionality
is made configurable to support different customer requirements.
How can processing be scaled out elastically while being configurable
regarding the supported functions?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 48
Batch Processing Component
Requests are delayed until environmental conditions make their processing feasible.
How can asynchronous processing requests be delayed to be handled
when conditions for their processing are optimal?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 49
Data Access Component
Access to data is provided by special components that isolate complexity, enable
additional data consistency, and ensure adjustability of data elements.
How can the complexity of data storage be hidden while ensuring data
structure configurability?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 50
Data Abstractor
Data is abstracted to inherently support eventually consistent data storage through
the use of abstractions and approximations.
How can eventually consistent data be presented, so that possible
inconsistencies are hidden?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 51
Idempotent Processor
Application functions detect duplicate messages and inconsistent data or are
designed to be immune to these conditions.
How can an application component cope with message duplicates or
data inconsistencies that could lead to duplicate function execution?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 52
Transaction-based Processor
Components receive messages or read data and process the obtained information
under a transactional context to ensure successful processing.
How can a processing component ensure that messages are processed
and data is persisted successfully?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 53
Timeout-based Message Processor
Clients acknowledge message processing to ensure that messages are handled. If a
message is not acknowledged after a certain timeout, it is processed again.
How can an application guarantee that all handled messages are
processed at-least-once?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 54
Multi-component image
Virtual servers host multiple application components that may not be active at all
times to reduce provisioning and decommissioning operations.
How can a virtual server provide the functionality of multiple
application components to be used flexibly in applications?
Cloud Application Architecture Patterns – Cloud Application Components
[Link] 55
Cloud Application Architecture Patterns
Multi-Tenancy
[Link]
Shared Component
A component is accessed by multiple tenants to leverage economies of scale.
How can an application component be shared between multiple
tenants enabling some individual configuration?
Cloud Application Architecture Patterns – Multi-Tenancy
[Link] 57
Tenant-isolated Component
A shared component avoids influences between tenants regarding assured
performance, available storage capacity, and accessibility of functionality and data.
How can an application component be shared while enabling
individual configuration and tenant-isolation regarding performance,
data volume, and access privileges?
Cloud Application Architecture Patterns – Multi-Tenancy
[Link] 58
Dedicated Component
Components providing critical functionality are provided exclusively to tenants while
still allowing other components to be shared between tenants.
How can application components that cannot be shared be integrated
into a multi-tenant application?
Cloud Application Architecture Patterns – Multi-Tenancy
[Link] 59
Cloud Application Architecture Patterns
Cloud Integration
[Link]
Restricted Data Access Component
Data provided to clients from different environments is adjusted based on access
restrictions.
How can an application component alter provided data based on
access restrictions imposed on different environments?
Cloud Application Architecture Patterns – Cloud Integration
[Link] 61
Message Mover
Messages are moved automatically between different cloud providers to provide
unified access to application components using messaging.
How can message queues of different providers be integrated without
an impact on the application components using them?
Cloud Application Architecture Patterns – Cloud Integration
[Link] 62
Application Component Proxy
An application component is made available in an environment from where it
cannot be accessed directly by deploying an application component proxy.
How can an application component be accessed if direct access to its
hosting environment is restricted?
Cloud Application Architecture Patterns – Cloud Integration
[Link] 63
Compliant Data Replication
Data is replicated among multiple environments. Data is automatically obfuscated
and deleted to meet laws and security regulations.
How can data be replicated between environments if some
environments may only handle subsets of the data due to laws and
corporate regulations?
Cloud Application Architecture Patterns – Cloud Integration
[Link] 64
Integration Provider
Integration functionality such as messaging and shared data is hosted by a separate
provider to enable integrate of otherwise separated hosting environments.
How can application components that reside in different
environments, possibly belonging to different companies, be
integrated through a third-party provider?
Cloud Application Architecture Patterns – Cloud Integration
[Link] 65
Cloud Application Management Patterns
Management Components
[Link]
Provider Adapter
Provider interfaces are encapsulated and mapped to unified interfaces to separate
concerns of interactions with the provider from application functionality.
How can the dependencies of an application component on a provider-
specific interface be managed?
Synchronous to Asynchronous Adapter
Cloud Application Management Patterns – Management Components
[Link] 67
Provider Adapter
Provider interfaces are encapsulated and mapped to unified interfaces to separate
concerns of interactions with the provider from application functionality.
How can the dependencies of an application component on a provider-
specific interface be managed?
Asynchronous to Synchronous Adapter
Cloud Application Management Patterns – Management Components
[Link] 68
Managed Configuration
Scaled-out application components use a centrally stored configuration to provide a
unified behavior that can be adjusted simultaneously.
How can the configuration of scaled out application component
instances be controlled in a coordinated fashion?
PULL: PUSH:
Cloud Application Management Patterns – Management Components
[Link] 69
Elasticity Manager
The utilization of IT resources on which an application is hosted is used to determine
the number of required application component instances.
How can the number of required application component instances be
determined based on the utilization of hosting IT resources?
Cloud Application Management Patterns – Management Components
[Link] 70
Elastic Load Balancer
The number of synchronous accesses to an elastically scaled-out application is used
to determine the number of required application component instances.
How can the number of required application component instances be
determined based on monitored synchronous accesses?
Cloud Application Management Patterns – Management Components
[Link] 71
Elastic Queue
The number of accesses via messaging is used to adjust the number of required
application component instances.
How can the number of required application component instances be
adjusted based on monitored asynchronous accesses?
Cloud Application Management Patterns – Management Components
[Link] 72
Watchdog
Applications cope with failures automatically by monitoring and replacing
application component instances if the provider-assured availability is insufficient.
How can applications automatically detect failing application
components and handle their replacement?
Cloud Application Management Patterns – Management Components
[Link] 73
Cloud Application Management Patterns
Management Processes
[Link]
Elasticity Management Process
Application component instances are added and removed automatically to cope
with increasing or decreasing workload, respectively.
How can the number of resources to which application components
are scaled-out be adjusted efficiently to the experienced workload and
anticipated future workload?
Cloud Application Management Patterns – Management Processes
[Link] 75
Elasticity Management Process
Application component instances are added and removed automatically to cope
with increasing or decreasing workload, respectively.
How can the number of resources to which application components
are scaled-out be adjusted efficiently to the experienced workload and
anticipated future workload?
Cloud Application Management Patterns – Management Processes
[Link] 76
Elasticity Management Process
Application component instances are added and removed automatically to cope
with increasing or decreasing workload, respectively.
How can the number of resources to which application components
are scaled-out be adjusted efficiently to the experienced workload and
anticipated future workload?
Cloud Application Management Patterns – Management Processes
[Link] 77
Resiliency Management Process
Application components are checked for failures and replaced automatically without
human intervention.
How can the overall availability of an application be ensured
automatically even if individual application component instances fail?
Cloud Application Management Patterns – Management Processes
[Link] 78
Resiliency Management Process
Application components are checked for failures and replaced automatically without
human intervention.
How can the overall availability of an application be ensured
automatically even if individual application component instances fail?
Cloud Application Management Patterns – Management Processes
[Link] 79
Resiliency Management Process
Application components are checked for failures and replaced automatically without
human intervention.
How can the overall availability of an application be ensured
automatically even if individual application component instances fail?
Cloud Application Management Patterns – Management Processes
[Link] 80
Feature Flag Management Process
If the cloud cannot provide required resources in time, application features by are
degraded gracefully in order to keep vital features operational.
How can the performance of an application degrade gracefully, if the
experienced workload increases but additional cloud resources are
unavailable or take too long to provision?
Cloud Application Management Patterns – Management Processes
[Link] 81
Feature Flag Management Process
If the cloud cannot provide required resources in time, application features by are
degraded gracefully in order to keep vital features operational.
How can the performance of an application degrade gracefully, if the
experienced workload increases but additional cloud resources are
unavailable or take too long to provision?
Cloud Application Management Patterns – Management Processes
[Link] 82
Feature Flag Management Process
If the cloud cannot provide required resources in time, application features by are
degraded gracefully in order to keep vital features operational.
How can the performance of an application degrade gracefully, if the
experienced workload increases but additional cloud resources are
unavailable or take too long to provision?
Cloud Application Management Patterns – Management Processes
[Link] 83
Update Transition Process
When a new application component version, middleware versions etc. become
available, running application components are updated seamlessly.
How can application components of a distributed application be
updated seamlessly?
Cloud Application Management Patterns – Management Processes
[Link] 84
Update Transition Process
When a new application component version, middleware versions etc. become
available, running application components are updated seamlessly.
How can application components of a distributed application be
updated seamlessly?
Cloud Application Management Patterns – Management Processes
[Link] 85
Update Transition Process
When a new application component version, middleware versions etc. become
available, running application components are updated seamlessly.
How can application components of a distributed application be
updated seamlessly?
Cloud Application Management Patterns – Management Processes
[Link] 86
Standby Pooling Process
Application component instances are kept on standby to increase provisioning
speed and utilize billing time-slots efficiently.
How can required provisioning times for application component
instances be ensured while utilizing pay-per-use resources in an
optimal fashion?
Cloud Application Management Patterns – Management Processes
[Link] 87
Standby Pooling Process
Application component instances are kept on standby to increase provisioning
speed and utilize billing time-slots efficiently.
How can required provisioning times for application component
instances be ensured while utilizing pay-per-use resources in an
optimal fashion?
Cloud Application Management Patterns – Management Processes
[Link] 88
Standby Pooling Process
Application component instances are kept on standby to increase provisioning
speed and utilize billing time-slots efficiently.
How can required provisioning times for application component
instances be ensured while utilizing pay-per-use resources in an
optimal fashion?
Cloud Application Management Patterns – Management Processes
[Link] 89
Standby Pooling Process
Application component instances are kept on standby to increase provisioning
speed and utilize billing time-slots efficiently.
How can required provisioning times for application component
instances be ensured while utilizing pay-per-use resources in an
optimal fashion?
Cloud Application Management Patterns – Management Processes
[Link] 90
Composite Cloud Application Patterns
Native Cloud Applications
[Link]
Two-Tier Cloud Application
Presentation and business logic is bundled to one stateless tier that is easy to scale.
This tier is separated from the data tier that is harder to scale.
How can application functionality be separated from data handling to
scale them independently?
Composite[Link]
Cloud Application Patterns – Native Cloud Applications 92
Three-Tier Cloud Application
Presentation, business logic, and data handling are separated to scale stateless
presentation and compute-intensive processing independent of the data tier.
How can presentation logic, business logic, and data handling be
decomposed into separate tiers that are scaled independently?
Composite[Link]
Cloud Application Patterns – Native Cloud Applications 93
Content Distribution Network
Applications component instances and data handled by them are globally
distributed to meet the access performance required by a global user group.
How can timely access to an application be ensured for a globally
distributed user group?
Composite[Link]
Cloud Application Patterns – Native Cloud Applications 94
Composite Cloud Application Patterns
Native Cloud Applications
[Link]
Hybrid User Interface
Varying workload from a user group interacting asynchronously with an application
is handled in an elastic environment.
How can a user interface for asynchronous interaction be hosted in a
cloud while being integrated with an application otherwise hosted in a
static data center?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 96
Hybrid Processing
Processing functionality that experiences varying workload is hosted in an elastic
cloud while the remainder of an application resides in a static environment.
How can processing components that experiences varying workload be
hosted in an elastic cloud while the remainder of an application is
hosted in a static data center?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 97
Hybrid Backend
Backend functionality (data-intensive processing and storage) is experiencing
varying workloads and is hosted in an elastic cloud.
How can processing components that experience varying workload
and access large amounts of data be hosted in an elastic environment
while the remainder of the application is hosted in a static
environment?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 98
Hybrid Data
Data of varying size is hosted in an elastic cloud while the remainder of an
application resides in a static environment.
How can a data handling functionality that experiences varying
workload be hosted in an elastic cloud while the rest of the application
is located in a static data center?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 99
Hybrid Backup
Data is periodically extracted from an application to be archived in an elastic cloud
for disaster recovery purposes.
How can data be archived in a remote environment while the
remainder of the application is hosted in a static environment?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 100
Hybrid Application Functions
Some application functionality provided by user interfaces, processing, and data
handling is experiencing varying workload and is hosted in an elastic cloud.
How can arbitrary functionality of an application be distributed among
static data centers and elastic clouds best matching its requirements?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 101
Hybrid Development Environment
A production runtime environment is replicated and mocked in an elastic
environment where new applications can be developed and tested.
How can an application use different computing environments during
its development, test, and production stages?
Composite[Link]
Cloud Application Patterns – Hybrid Cloud Applications 102
The End.
[Link]
