Standardisation in the Cloud

Overview

As we know, cloud computing enable to share & manage pool of computing resources, configurations and application services with more convenient, ubiquitous way by minimizing interactions with its provider(s). Cloud model meant for mass consumption by simplifying its access points for end usage, but it also abstract and hide its big backend and middleware engines and its complexities from end users\consumers. I always think, What enterprise architects would be expecting out of cloud platform while choosing right cloud provider and to serve that what type of architecture and design principals and standards a cloud providers' engineering teams need to factors into their services' implementation. However, the abundance of easy and on-demand access of cloud comes with its own challenges, like in scenario of combining & integrating multiple application hosting platforms like on-premises to cloud & across multiple cloud platforms. How easily I can move my applications and data to the cloud? What if I move from one cloud provider to another, what are challenges and would my existing interfaces continue to work as is? How easy to consume and publish data across multiple enterprise cloud platforms? Will my organization get into provider lock-in trap? , these are few questions enterprise architects would analyse before choosing the right cloud provider. The applications & data interoperability and portability plays a key role when moving from on-premises to the cloud as well as one cloud to another cloud platform and this will be the focus area of this post.

Importance of Standardisation

 Apps & data interoperability and portability fully depend on the type of architecture patterns and standards , a particular cloud provider follows with its service fabric.  Almost all public cloud providers' service fabric follow Service Oriented Architecture (SOA), with standard service orchestration and choreography based architecture pattern. How many standards, cloud providers are following in implementation of their services design & architecture is the most important factor to preserve maximum level of interoperability and portability. The cloud interoperability and portability not only allow best use & integration of multiple heterogeneous cloud services by avoiding vendor-lock in but also play a vital role in cloud adoption by realizing its value as computing utility (Cloud Standards Customer Council,2014). The portability helps in onboarding / migration of enterprise components to the cloud while interoperability makes consumptions, integration, deployment of such component to heterogeneous interfaces and services. Getting stuck into specific cloud providers proprietary architecture and pattern not only trap enterprises in vendor lock-in but also reduces possibilities of collaboration and integration between services across diverse platforms by increasing cost and security risks.

Interoperability

The degree to which components coming from diverse platforms (like multiple public cloud, private cloud, on-premises) work together successfully is measured as interoperability. Achieving the higher ratio of interoperability will not only smoothen applications' migration but also enable integration between multiple cloud providers. The semantic interoperability plays a crucial role in cloud computing by enabling it as promising  IT paradigm and allowing it as efficient global service collaboration platform.

Portability

Portability plays a critical role in moving data to the cloud platform in onboarding process as well as providing full control over personal data.  It can be categorized in below areas

• Data portability: standards in this area provides reusability of data across application platform.
• Application Portability: standards in this area provides reusability of application components across cloud PaaS services and other platforms.
• Platform portability: standards in this area provides reusability of platform components across on-premises and IaaS cloud platforms (also known as platform source portability). As well as support reusability of application and data packages to the respective supported platforms ( also known as machine image portability).

Key Standardisations

Below are key areas of Interoperability & Portability standardisation to get maximum benefits while onboarding, hosting, moving and managing workloads into and across cloud platforms.

Security Standards

User identity plays a key role to authenticate users /consumers to the cloud platform to provide the right level of access to requested services.  End users are more registered with the application then infrastructure/services and are not suitable candidates from the interoperability perspective, but cloud resource consumers are admin users, are the right candidate for interoperability context. By using right user identity interoperability standard, a scenario like federation between identity providers and cross-platform single sign-on capabilities are possible. Below are few examples.       

• OAuth: IETF(the Internet Engineering Task Force) defined the open protocol allows the client to access resources on behalf of the resource owner. Also, allow authorized their party server resources access without sharing credentials.
• OpenID Connect, LDAP & SAML: Standards that enable their party access and management capabilities.
• WS-Security:  OASIS standard specification allows the method to secure SOAP message using XML encryption and security tokens and being used in MS Azure and Amazon EC2.        

Workload & Platform Migration and Management

Application interoperability & portability helps to migrate enterprise workload (mostly deployed as virtual machine image), from one cloud provider to another by making easy extraction, uploading and management of workload across cloud platforms. Below are key workload migration interoperability & platform portability standard followed by big cloud providers.       

• Open Virtualization Framework (OVF): A DMTF supported packaging standard for the virtual machine supported by AWS, OpenStack, Eucalyptus.
• Virtual Machine Hard Disk (VHD): Supported by Microsoft, a file format for the virtual machine. This is also supported by AWS cloud platform.
• Amazon Machine Image (AMI): AWS proprietary virtual machine format, now becoming de facto standard and adopted by OpenStack and Eucalyptus as well.
• OpenStack: Open source cloud IaaS provider provides great example allowing portability of platform source code across cloud platforms.
• Common Information Model (CIM): Defined by DMTF (The Distributed Management Task Force ) provides information management standard definition across cloud platform systems.
• The Virtualization Management (VMAN): provide standards for virtual resource management.

Data Migration and Management

Data interoperability & portability standards help in migrating data from one cloud provider to another by providing standard data extraction and upload capability. Also, provide standards to perform CRUD (create, retrieve, update, delete) operations on data stored in the cloud, so switching to another provider, application CRUD interfaces remain same. Also, availability of standard data access interfaces (like REST, SOAP, etc.) not only provide rapid development advantages but also grow wider cross-platform tool sets ecosystem of cloud providers.       

• Standard Data Models: Provide standard data storage structure with standard data querying interfaces to query and process data across cloud platform by keeping critical query & processing interfaces unchanged. SQL for relational databases, NoSQL database, Resource Description Framework (RDF) and the Web Ontology Language (OWL), etc.
• Cloud Data Management Interface (CDMI): Standards defined by SNIA (Storage Networking Industry Association) provide unified standard API to perform CRUD operations on cloud storage. As well as the standard for data discovery and management.
• Representational State Transfer (REST): Not data specific standards but widely adopted as standard data access interface by most of the cloud providers.
• SOAP: Again not data specific standard but widely adopted protocol used by most of cloud providers' storage management interfaces.