How can you evaluate cloud architectures for specific use cases?

Some key topics to explore at this level - 1. Do you need orchestration? 2. What is cloud security architecture ? 3. Data strategy and associated data patterns ? 4. Cloud monitoring and cost governance

Indeed, the starting point is discerning Who is the user? and What are their use cases? A common oversight by many (often me too in the beginning), even by seasoned technologists, is neglecting non-functional requirements from the outset, which can significantly impact the software product’s adoption down the line. Been there did it and learnt from it ??

In My experience, we always needs to work backwards from what customer wants. This in combination to well architected frameworks can help one build better cloud architectures

One angle to consider is also presence or not of existing on-premises which will interact with cloud components. This interaction, captured as a hybrid architecture, is very common in enterprise environments and leads to different set of integration or shared services patterns.

First I will see is the business benefit. Second is the cost-benefit analysis. Third will be identifying component / system by risk. Once I have these in place I will look at the skill / experience in the company who can manage the systems. Other thing to consider is also govt regulation and security.

Absolutely one thing I liked most is that the architecture style selection is the most important and critical aspect. It is a multi dimensional matrix decision and is not only dependent on technology selection but also capabilities, resource availability to execute, timelines, budget and other priorities.

In my experience, I've rarely found monolithic architectures as the way to go unless it's a throw-away project. Serverless architectures are usually lower-cost and faster-to-market. They work really well for event-driven workloads and real-time applications. However, different cloud vendors offer different capabilities and performance characteristics, so you have to be more careful about your vendor choice and you can end up being tied in to one vendor. Serverless components also can be harder to monitor and debug and need some careful thought while being designed. Most complex systems these days will likely use a hybrid architecture with a combination of "serverless microservices" and some independently deployed microservices.

Craft KPI/KPA metrics with the business, based on their business outcomes and tie those metrics to the relevent cloud solutions that was implemented.

Break the architecture components into Infrastructure, Data Management, AI and ML, Business Applications, Security layers. Given the business problem, you would need to make building block decisions in each layer. e.g. some considerations for infrastructure layer can be, Redundancy, Availability, Latency, RPO, RTO, Multiparty . Likewise, for data management layer, based on business usecase an appropriate type of database, data processing building block should be selected.

System architecture and design articulation: Vendor identification/management, site/location, compute, storage, database/application, compatibility matrix, networking, interfaces, security, monitoring, load balancing, load sharing, caching, high availability, BAR, Identity management, auto-scalability, KPI, KRA, SLA and frameworks. Consider service and maintenance: (optimisation, standardisation, adaptability, reliability, continuity): SDLC management, architecture tweaking, asset/contract management, reliability, availability, extensibility, maintainability, manageability, security, collaboration, testing, performance, business communications, vendor management, deployment, validation and versioning.

Along with cost, its also availability. Depending on what type of infra and what location it is needed in makes a big impact. Certain cloud vendors do not offer certain services in some regions or they are more expensive in some regions. Sometime they do not have HA/DR in some regions.

Cost is not only the cost of physical components of architecture, but also cost of deploying the architecture, skilling people up, if you are using a new technology. Cost of adoption of new technology by the organization should also be considered.

Cost Planning:operational expenditure (OPEX) and capital expenditure (CAPEX) costs. Infrastructure costs: pre and post-production infrastructure setup, maintenance, and operational expenses. Software costs: Licensing Security costs: Security measures Services costs: Cloud based, third-party and APIs Data management costs: Access, usage and location Scalability and demand costs: As per application or infra based demands Governance costs: Monitoring tools and cost utilization Total cost of ownership(TCO): Calculate overall cost by aggregation above all factors ROI: Ensure the benefits, justify the costs incurred.

In my experience, utilizing a hybrid approach using multiple cloud providers is the best way to take advantage of the best tool and not put in a box. This strategy may sometimes cost more initially, however the return on investment will make it worth it when done by an expert.

Choosing Cloud Services: - Consider control, ease of development, maintenance, scalability for Cloude models(IaaS, PaaS, SaaS). - Each fits specific use cases with trade-offs in control and ease of use. Evaluation Tools: - Scoring Models: Quantitative evaluation based on criteria, aiding decision-making. - Decision Matrices: Structured tool, weighted criteria for best option visualization. - SWOT Analysis: Evaluates strengths, weaknesses, opportunities, threats, guiding planning. - Best Practices: Involve stakeholders, define criteria, update models, consider expert opinions. EA Frameworks: - TOGAF:Comprehensive, structured approach, promotes reusability. - SAFe:Framework scales Agile/DevOps/Lean, enabling coordinated planning across teams.

I think based on the internal skillset of the organization's Cloud administration, devops, software teams plays a major role along with the use cases that are needed to be done.

An organisation's ESG strategy should always be taken into consideration as well. Architectures and deployments in the could should help to reduce a company's environmental footprint, improve the social responsibility, and enhance the governance practices of cloud-based solutions.