Popular
Cloud
Deployment
Models

Part 2 of Principia Cloud Series

The flexibility of services and applications determines the biggest advantages of cloud platforms. Quite in resonance, different enterprises from variant sectors have diverse requirements paired with complex environments and stringent rules and regulations that need to be adhered to. Hence, while a traditional digital product company won’t face any heat with a normal lift and shift to the cloud, organizations like Banking and Finance need to embrace stringent security, data residency, and information locked-in policies that mandate a different type of cloud environment.

Cloud Services Providers hence extend multiple cloud deployment models with a mix of private and shared computing, networking, datacenter, storage, software, application resources basis business needs and policies. More often than not, owing to its large-scale global operations, a client company might choose to adopt more than one cloud platform and administer it all within a single intuitive dashboard or platform. Service providers need to carefully assess a client’s present and future business needs, craft a tailored cloud delivery strategy and blueprint, implement the environment with zero disruption and data loss, and ensure there are nil workload failovers. Also, service providers need to map the cloud-native tools, solutions, and service models needed (IaaS, PaaS, SaaS) services to address the client’s business, operational, and client delivery goals.

Following are the most common cloud deployment models:

Public Cloud

The most popular cloud deployment model and the centerpiece offering of Global Cloud Service Giants such as Amazon Web Services (AWS), Microsoft Azure, Oracle Cloud, Google Cloud Platform, IBM Cloud, and more. Public cloud defines a distributed IT architecture hosted on the provider’s securely located mammoth computing infrastructure (servers, storage, datacenters, compute, networks, and more) that can be accessed by a shared pool of users/tenants (consumers, enterprises alike) on an as-needed, pay-as-you-go basis. This model is ideal for enterprises that face extensive losses on maintaining IT platforms, have medium-sized IT environments that could be easily modernized or virtualized, need extra scalability to meet consumer demands, and require an IT transition to the cloud at speed.

The major difference between rented cloud infrastructure companies and dedicated cloud organizations is the latter’s extra IT service offerings on top of the backend centralized computing systems. Cloud providers hence, in addition to IT infra or hardware renting (typical IaaS), adopts cutting-edge solutions (in-house or third-party systems) to virtualize infra and software resources for clients, create hypervisors or containers to safely package-categorize-run-maintain virtualized assets, add an administrative layer of control with backend integrations and architectural features to create an intuitive IT ecosystem, and an array of APIs, protocols, interfaces, automation solutions to allow users leverage the cloud architecture and the underlying infra brute with unprecedented agility, scalability, and security. Some key characteristics of public clouds include on-demand computing and self-service provisioning, resource sharing or pooling, hyper scalability and extra flexibility-elasticity, billed basis usage, high uptime and availability, centralized protection and security, and exhaustive network access.

How to Deploy Public Cloud?

Businesses looking to migrate on the public cloud need to first lock-in on the service model: rent computing, server, networks, datacenter infrastructure only (IaaS), require additional software development platforms and backend architectures to fast-track product development (PaaS) or embrace the end-to-end SaaS approach to modernize organizational workflows and achieve customer servicing excellence. The enterprise can opt for a live, on-network transfer of data and assets if the dataflow size is comparatively lower or adopt a complete offline transition. The hosting location (availability zone) is chosen with respect to client location proximity and compliance, regulatory requirements.

Post this exercise, the vendor maps the requirements and virtualizes the required assets, abstracts infra, pools information into data lakes, and orchestrates operational software and applications basis the model chosen. TA lift and shift model directly virtualizes applications on the cloud, re-factor requires certain models to be revamped to leverage the cloud backend, re-architect requires significant backend chunks to be re-hauled to synchronize on the cloud basis microservices architectures or replace the entire existing application backend with a new re-written version on the cloud. The transition is often in synonymy with the client’s vision with cloud computing and results in either partial transformation of IT on the cloud or complete, end-to-end modernization.

Benefits of Public Cloud:

• Truncated Expenses from a CapEx to OpEx Model: pay only for what you utilize and rent any IT infra, hardware, software basis needs with unprecedented scalability
• Cost-efficient multi-tenant architecture that reduces overall infrastructure costs with zero IT idle expenses. Avoid all IT maintenance, security, storage, database management, and related hassles with ease.
• Highly reliable and available architecture with near-zero downtime, zero disruptions, and DR-ready infrastructure. Scale your operations up or down as and when needed, based on operational needs.
• Contrary to usual notions, leverage cutting-edge security with advanced intelligence on the adopted public cloud platform. Most cloud providers are now spending billions on data security and threat prevention, management that dispels fears of enterprises regarding workload and information protection on the cloud.
• Embrace advanced analytical and intelligence capabilities with cloud-native solutions, leveraging the cloud’s colossal computing power and data resources. Avail cutting-edge business performance insights and make informed strategic decisions with ease.

Private Cloud

A private cloud defines a cloud environment wherein IT resources such as compute, networking, storage, databases, applications, software, architectures, and platforms are delivered on-demand to an enterprise in a specialized environment that’s isolated from the contemporary shared pool/multi-tenant environments. In other words, when an enterprise or organization opts for private cloud infrastructure, the provider crafts an exclusive IT environment, albeit at higher prices, and dedicates required virtualization infra, assets, applications that won’t be accessed to anyone but the client organization.

While such an arrangement constraints companies from leveraging the low price-points of public cloud platforms and their seemingly unending benefits, such IT environments become a dire necessity for core industrial organizations. Firms delivering healthcare, banking, manufacturing, utilities, and related offerings often harbor gargantuan databases filled with sensitive information that can’t be risked on a shared IT pool. Besides, most nations make such organizations liable to stringent regulations and compliance practices that would be difficult to adhere to in a full-fledged public cloud environment.

How to Deploy Private Cloud?

Instead of knocking on the doors of a cloud service provider, organizations can also build their private cloud networks. Secure, preserved datacenters and centralized processing systems would address storage, computing, networking needs whereas the developed cloud platform would virtualize combined resources from all physical infrastructure using automated virtualization technologies at one go. An additional administrative and control layer safeguards the data, workload perimeters enabling admins to track, monitor, and usage data-workflows with ease. On top, this private cloud infrastructure, clients could also opt to centralize backend software architectures using Linux kernels, Kubernetes or Docker containers, Microservices to abstract, allocate and scale application distribution across all end users of the private cloud network. In such a set-up, once the cloud infrastructure is installed, the client is in full control and responsible for running costs, resources, on-prem risk management, etc.

Public cloud giants such as Amazon, Microsoft, Google, IBM, Oracle have also come up with virtual private cloud environments designed on their public cloud architectures. Herein, dedicated infra and IT resources are allocated for the clients that are protected from the shared pool using Virtual Private Networks, VLAN, encrypted communications, subnets, and more. While some cloud providers don’t charge separately for a private cloud set-up and simply bill basis usage, many cloud providers instead charge a separate fee owing to the exclusivity of the environment. User organizations can leverage on-cloud security policies and compliant-ready architectures such as Governance, policy checkers, Identity and Access Management, Security and Monitoring, etc.

Client companies can also leverage private cloud-like advantages at less orchestration, abstraction, and management hassles with Community Cloud platforms. Community Cloud is a specialized case of a private cloud environment wherein user companies belonging to a similar vertical are grouped under an isolated, heavily protected cloud environment. For example, banking organizations needing private cloud architectures could share a single community cloud platform dedicated to the BFSI industry.

Benefits of Private Cloud:

• Avail highly personalized architecture with customized resources to meet exclusive business, operational demands. All infra services and solutions on the cloud network are dedicated to one organization only.  
• Exert unparalleled security controls across all levels, workflows, databases. Resource privacy and cloud security are maintained at the highest levels in compliant-ready architectures.
• With all compute resources dedicated to one organization’s requirements, gain ultra scalability, agility, and flexibility of workloads. Scale, evolve, upgrade, and transform architectures as and when needed without a single glitch
•  Enhanced, universal visibility across all resources, instances, workflows with ease

Hybrid Cloud

As the name suggests, Hybrid cloud platforms deliver the best of both worlds to companies: tight-knot security and extra agility of private clouds combined with cost-effectiveness, hyper scalability, and the colossal cloud-native solutions pool of public clouds. Hybrid clouds, by definition, details a scenario wherein there’s at least one public or private cloud environment. In large IT scenarios, there could be multiple private and public cloud architectures seamlessly connected within a hybrid umbrella.

As digital transformation needs and IT complexities go off the roof, the boundaries between private and public cloud are hard to realize. Earlier public cloud systems could be distinguished as off-premise, distantly located computing infra accessed by numerous enterprises together. However, the aggressive growth and demand of cloud have brought public cloud giants to the doorstep, delivering public cloud infrastructure on-premises. On the other hand, enterprises in a bid to shove off the datacenter, on-prem IT hassles, are now opting for exclusive private cloud infrastructure managed by private cloud providers on distant locations. To clear the air hence, the hybrid cloud must have the following characteristics: connecting multiple computing infra through networks, consolidate IT resources from variant environments under a single umbrella architecture, scale up and down as needed and provide extra resources on-demand, seamlessly transfer workloads-data between variant cloud environments, integrate unified administration tools, and automate key functionalities for seamless operations.

How to Deploy Hybrid Cloud?

Just like the previous two cloud models, hybrid clouds to virtualize infra including compute, networking, storage, datacenter assets, software platforms, backends, applications basis the adopted SaaS, IaaS, PaaS models. Abstraction of hardware and connecting on to the cloud, perimeter protection through special firewalls, data and applications packaging via containers and microservices, assets management, and all are delivered via the cloud architecture. However, there are primarily two types of hybrid cloud deployment that uphold the promise of this model:

The traditional hybrid cloud architecture essentially follows the combination of one or more public cloud environments with one or more private cloud environments through bulky systems, middleware, OS, containers, architectures, interfaces, APIs, and more. The platform facilitates easy transfer of data and workloads between the hybrid components with ease and agility. Most hybrid cloud solutions including AWS Outposts, Azure Stack, Google Cloud Hybrid are built and delivered using this model. To facilitate seamless connectivity, AWS offers Direct Connect, Microsoft Azure offers ExpressRoute, OpenStack provides OpenStack Public Cloud Passport, and Google delivers Dedicated InterConnect.

The second innovative hybrid cloud model instead stems from a ground-breaking realization. Instead of patching two distinct cloud environments together and allows seamless integration-connectivity-transfer of assets, why not build applications that universally run on multiple private or public cloud environments? As highlighted on Red Hat, this is akin to building a car that commutes on the road, air, and water instead of building multiple lane roads, bridges to connect different cities! Under this model, IT teams deploy the same operating systems, backend architectures across all cloud environments to abstract all hardware requirements, orchestrate application environments with unified platforms leveraging orchestration engines like Kubernetes or Red Hat OpenShift. This delivers a universal, consistent, and seamlessly interconnected computing environment across all cloud architectures without the need for special middleware, APIs, interfaces, and application customizations to enable disruption-free workload processing in the hybrid hub. Akin to a DevOps model, resources can work collaboratively without any hitch across multiple systems and architectures with uninterrupted scalability, agility, and uncompromised security.

Owing to spectacular methodologies and features, hybrid cloud environments have gained immense popularity especially for organizations with distinct IoT, Edge environments to leverage the best of digital transformation benefits. The hybrid cloud market is expected to explode to 128 billion US dollars by 2025!

Benefits of Hybrid Cloud:

• Optimize IT investments and leverage the best of both worlds: private cloud’s uninterrupted security and public cloud’s economical resources, unparalleled scalability and flexibility
• Stringent policy and compliance adherence: Privacy of important workloads and applications can now be transferred to or distributed across compliant-ready, secure hybrid cloud components with complete ease
• Easy connectivity between hybrid cloud components: Seamlessly transition workloads, data, applications, systems between adopted private and public cloud environments without disruption, based on security and regulatory requirements. Leverage a universal, consistent computing infra across all hubs to gain immense delivery and operational flexibility.
•  Top-notch security: With hybrid cloud environments, gain an extra layer of security to your IT infra and applications, workloads, data. In addition to the private cloud environment’s exclusive protection, embrace advanced cloud-native security tools of public cloud environments for enhanced risk administration over the entire IT landscape.
• Embrace advanced analytical and intelligence capabilities with public cloud-native solutions, leveraging the cloud’s colossal computing power and data resources. Avail cutting-edge business performance insights and make informed strategic decisions with ease.

Multicloud

Consider a scenario. Suppose your legacy IT systems and infrastructure are considerably failing to meet your growth expectations. After multiple deliberations, the board decisions to move the entire IT landscape to a dedicated cloud platform. Upon communication with a cloud service provider, the entire assessment, objectification, blueprinting, and implementation is completed with all infra assets, networks, servers, applications, platforms being safely migrated to the cloud under a SaaS framework. Your firm witnesses extra agility on your workloads, ultra-low latencies, no operational delays, and streamlined workflows. However, within a few months, though your enterprise is successfully operating on the cloud, you require some special HRM, ERP, or CRM software to fine-tune some key relations offerings. You shortlist another vendor who delivers this application on-net via their infrastructure hosting. You avail the packaged solutions suite and choose to administer the entire workload for the same on the latter’s cloud environment. Kudos, now you have a multicloud architecture.

How to Deploy Multicloud?

As the term highlights, multicloud refers to the usage of two or more cloud environments, either public, private, or mixed by an enterprise. Unlike a hybrid cloud environment, there is no architectural collaboration, unification, or integration-orchestration between the adopted cloud environments. Each platform operates independently and the client utilizes both cloud systems to achieve variant daily operations. Often, this is the ideal case scenario with medium-sized or large companies looking for IT flexibility as in reality no single cloud platform can address all requirements perfectly: embed best functionalities from multiple, exciting cloud providers. Apart from this situation, organizations might also face an additional burden if the adopted cloud platform is located thousands of miles away beyond borders and some key operations are mandated to be hosted in-country, adhered to key regulations, or need extra scalability and availability. A third popular instance to adopt multicloud environments stems from the jitters of modern, fast-growing companies to reduce reliance on single cloud providers to avoid vendor lock-in. A single disaster could jeopardize entire operations and the presence of multiple cloud environments could keep sustaining operations during accidental failovers of one or two cloud platforms.

Dedicated automation solutions, microservices architectures, and containers using Linux Kernels, Kubernetes, or Docker could facilitate the design of multiple cloud IT frontends. With containers, key libraries, applications, platforms, the software could be packaged into different cloud environments based on architectural needs, synchronicity, and business requirements. Microservices, the ‘art’ of slashing down large software codes into smaller components for portability and transition ease, could catalyze a seamless, distributed IT structure among the adopted cloud platforms. For instance, a single application could utilize the scalability and compute of one cloud system with AWS EC2 Instances, for example, state-of-the-art AI protection solutions like that of Azure Sentinel, and dynamic functionality libraries from the third vendor such as S/4 HANA ERP for seamless customer servicing. A step further, more often than not, multicloud managed service providers arm client organizations with a single, multi-purpose, and intelligent dashboard to administer all adopted cloud platforms and their workloads with ease. Admins can monitor performances, address risks, and gain unique performance insights for better decision-making.

Benefits of Multicloud:

• Avoid vendor lock-in: Reduce reliance on single cloud vendors and mitigate issues such as obsolescence, vendor insolvency, failovers, low supports, non-alignment with future business needs and products
• Reduce Disaster Impact on Operations: With IT infra and applications distributed across multiple cloud platforms, reduce the possibilities of total operation disruption and business impact due to failovers of one or two cloud platforms. Get necessary backup to lower risks with ease.
• Cost-effective approach: Gain the most benefits at the least investments across multiple cloud vendors. Leverage the solutions, services that each deliver best and administer everything with ease.
• Data analytics and AI capabilities: Embrace advanced analytical and intelligence capabilities with public cloud-native solutions, leveraging the cloud’s colossal computing power and data resources. Avail cutting-edge business performance insights and make informed strategic decisions with ease.
• Enhanced, universal visibility across all resources, instances, workflows across all cloud environments with ease through intuitive, intelligent, and multi-purpose dashboards

A Peek into Trends: Function-as-a-Service (FaaS) or Serverless Computing

A relatively new as-a-service model on the block, adding to the fundamental trident of SaaS, PaaS, and IaaS, Function-as-a-Service (FaaS) was first made popular by hook.io and is now widely integrated into the services of the largest cloud platforms: AWS Lambda, Google Cloud Functions, IBM OpenWhisk, Microsoft Azure Functions to name a few. Before we dive into the meaning and immense promise, utilities of FaaS, one has to re-vitalize the microservices architecture of software development.

Highlight into Microservices Architecture

Without precise realization, modular architectures are what make life possible. Modular utilities are like independent chunks organized into a powerful whole. If there’s an issue, instead of breaking down the entire system, one can pinpoint the risk to a particular module within the system, modify and re-orient the same, and place it back neatly within the system. Talk about our biological architectures with organ and tissue modules for instance. When we have a heart problem, all we need to do is operate on that particular organ and everything falls back to place. How would it have been like if we had to repair every organ accordingly after a heart operation is made so that all the former systems work in sync with the revitalized heart? This seems like a no-brainer but this was precisely the challenge with age-old monolithic software architectures. A bug in a storage request module, for instance, would require the entire architecture to be torn down (assuming that the issue is recognized to a particular segment in the first place) and adjustments made accordingly to make it work again. With a microservices architecture, the software or application is developed on tens of shorter, independent modules that could be updated, repaired, or revitalized upon wish without hampering the overall platform.

The Function-as-a-Service Architecture: Introduction to Serverless

Function-as-a-Service moves a step further to this already useful architecture. Breaking down the microservice modules into further function segments, developers and coders could now organize, revitalize applications at extremely minute, sub-module functional levels without worrying about the changes to be made on the overall modules and server architectures executing the modules or applications. Hence, a firm’s IT team is abstracted from the underlying infrastructure, server complications down at the functionality levels while the same is managed, maintained, scaled, and updated by the cloud provider.

Simply put, developers now can upload their codes right on the cloud platform as a function and the cloud provider would automatically extend a server/infra to run the code whenever the same is called to use. (A crude example: Instead of renting a server to run/manage appointment handling in an app, the server is directly assigned by a cloud provider whenever the appointment form is active due to submissions). Therefore, clients don’t need to rent servers on the cloud for a full-time basis and billing occurs only when the servers/infra are called to action due to the deployed functional codes running live. In contrast, IaaS, PaaS, or SaaS requires full-time renting of servers that harbor at least some ground activity during idle conditions (When a specific software module is not in action). Common examples include data processing (e.g., batch processing, stream processing, extract-transform-load (ETL)), and Internet of things (IoT) services for Internet-connected devices, mobile applications, and web applications.

As the phenomena dictate above, this is why FaaS is also known as serverless computing. In reality, serverless is a subset of FaaS and the latter extends to all kinds of computing infra. This cloud service mode transforms IT into a utility and grants real freedom to developers, coders, software engineers to focus on application and software development without spending a second on the infra requirements end-to-end.

Benefits of FaaS:

• Complete developer freedom — all underlying server and infrastructure is managed by the cloud provider and assigned to the client automatically whenever a deployed application/software function is called to action
• Higher Developer Bandwidth: Exceedingly more time to focus on developing the code architecture and the frontend end-user experiences rather than worrying about tweaking, modifying backend server and infra architectures.
• Freedom to Hyperscale: Since servers and underlying infra are automatically assigned to the client’s application functions whenever the latter is called to action, there’s no bar for scalability. The specific function can run on higher resources without the entire application being called, thereby minimizing delays and lags.  
• Extremely cost-effective: Compared to all other service models, this generally is the most cost-effective approach as clients don’t need to incur any idle expenses for the infra. The same is being paid only when specific functions in an application are active and resources allocated hence (No need to pay for the extra servers or infra running that would have occurred if the entire application/module had to run for that one functional segment)
• Business/Delivery Advantage: By drilling down the microservices architecture into further functions, it becomes easier to manage, maintain, deploy, and run code chunks independently whenever and wherever needed. This accelerates the delivery pipeline and makes it easier to respond to risks and mitigate them at functional levels without jeopardizing the entire module or application.

Next in this series is the Common concerns and deadlocks, myths & misconceptions of cloud computing. Keep watching this space.