5.4: Software Architecture and Design Consideration for Cloud

Introduction to Cloud Computing · LibreTexts

0%

Contents

    Text

    Size
    100%
    Theme
    Typeface
    Line length
    Spacing

    Read aloud

    How

    Speed: 1×
    Light up
    Highlight

    Symbols: Mulberry Symbols, © 2018–2026 Steve Lee, CC BY-SA 4.0 (licence, the set), the site's own copies.

    From LibreTexts, under CC BY 4.0, as the source states it; its equations are written in TeX, which the PDF shows typeset. This page at LibreTexts.

    5.4: Software Architecture and Design Consideration for Cloud

    Beyond the high-level principles discussed above, designing an enterprise cloud strategy also requires careful attention to software architecture and engineering considerations. Cloud-based systems have specific design challenges and opportunities. Below, we highlight some key design considerations and how they play into the strategy:

    • Requirements and Geographic Location: When planning a cloud deployment, understanding requirements is paramount – especially those related to data locality and latency. Data sovereignty laws might require certain data to reside in specific regions or countries. Thus, an enterprise might choose cloud provider regions or set up a hybrid cloud to meet geographic requirements. Additionally, user distribution matters: if your users are global, you might design your cloud architecture to use multiple regions or content delivery networks (CDNs) to reduce latency worldwide. In short, align your cloud architecture with both legal requirements and performance needs of your user base.
    • Cost Management: Cloud introduces a utility-based cost model, which is a double-edged sword – you pay only for what you use, but unchecked usage can lead to budget surprises. Architectural decisions greatly affect cost. For instance, choosing a serverless architecture might reduce costs for intermittent workloads (no idle charges), while over-provisioning large VM instances can waste money. A good design leverages auto-scaling to match capacity to demand and uses cost monitoring tools to get visibility. When strategizing, it’s important to plan for cost governance (e.g., setting budgets, using tags to track departmental usage, and optimizing resource sizes).
    • Power and Efficiency (Energy and Redundancy): Although power and cooling are mainly the cloud provider’s concern in their data centers, enterprises focused on sustainability may consider cloud providers that use renewable energy or have more efficient infrastructure. From an architecture perspective, efficiency translates to designing applications that can scale down when not needed (avoiding wasteful usage). Redundancy is a critical design choice for achieving high availability: deploying redundant systems across availability zones or regions so that even if one data center fails, the service remains up. Cloud providers make it easier to build redundancy (with managed multi-AZ databases, load balancers, etc.), but the strategy must plan for it (what tier of redundancy does each application require? e.g., mission-critical apps might be multi-region active-active, whereas a dev/test system might be fine without redundancy).
    • Scalability and Elasticity: Scalability is often a key driver for cloud adoption. Architecturally, this means designing applications to scale horizontally (add more instances under a load balancer) rather than assuming one big server will scale up. Stateless service design, distributed caching, and use of scalable backend services (like cloud-managed databases that scale or queue systems that buffer load) are patterns to embrace. Elasticity – the automated form of scalability – requires integrating with cloud auto-scaling features and designing for quick startup and teardown of components. A well-architected cloud application might use containers or serverless functions to scale out rapidly in response to load. In an enterprise cloud strategy, identify which systems need elasticity (customer-facing ones usually do) and ensure those are built cloud-natively to take advantage of it.
    • Network Topology and Connectivity: Cloud architecture must account for how different components communicate. This involves designing Virtual Private Clouds (VPCs), subnets, and peering or VPN connections for hybrid scenarios. A common strategy is to use a hub-and-spoke network architecture in the cloud: a hub VPC (or virtual network) for shared services like VPN connectivity, and spoke VPCs for different applications or teams, all connected through the hub or a cloud transit gateway. Network design also involves considerations like load balancing (global vs local load balancers), use of CDNs for content, and ensuring low-latency connections where needed (possibly leveraging cloud provider’s backbone between regions). Security groups, firewalls, and segmentation must be architected in from the start to protect the system while allowing necessary communication. Enterprises might also consider multi-cloud connectivity and how to route traffic efficiently between clouds if using more than one provider.
    • Utilization and Resiliency: In traditional data centers, hardware utilization was often low; cloud encourages higher utilization but with resilience. Design for resilience means assuming failures will happen (instance crashes, network glitches, etc.) and handling them gracefully. Patterns like graceful degradation, retry logic with backoff, circuit breakers (in microservices) all contribute to resiliency. At the same time, you want to drive utilization up to get cost efficiency: for example, run more containers per host until the host is reasonably utilized, or use spot instances (excess capacity) for non-critical workloads to reduce costs. A resilient design might deploy workloads across multiple instance types or zones so that even if one type or zone fails, others pick up the slack (improving overall utilization of available resources).
    • Security & Identity Management: Cloud security architecture is a vast topic, but some key elements include:
    • Identity and Access Management (IAM): Design a least-privilege model for both users and services. Leverage the cloud provider’s IAM to control which identities can access which resources. Many enterprises integrate their corporate identity systems (like Active Directory or SSO providers) with cloud IAM for unified authentication.
    • Network security: Use network segmentation (as mentioned), restrict inbound internet access to only what’s necessary, and consider private connectivity (e.g., AWS PrivateLink, Azure Private Endpoints) for connecting to cloud services without exposing traffic to the internet.
    • Data encryption: Enforce encryption at rest and in transit. Most cloud providers have encryption options and key management services (e.g., AWS KMS, Azure Key Vault) to simplify this.
    • Security monitoring: Cloud strategy should include how to monitor and detect security issues – using tools like AWS CloudTrail for auditing API calls, Azure Security Center for threat detection, etc. The architecture might incorporate a SIEM (Security Information and Event Management) system that aggregates logs from cloud resources for analysis.
    • Application security: Continue to follow best practices (e.g., use of HTTPS for all services, regular vulnerability scanning, applying patches to VMs or using managed services that handle patching).
    • Governance (again): Many security aspects tie into the governance principle – such as compliance checks, incident response plans for cloud, etc.

    Designing with security and identity in mind is essential from day one; cloud offers powerful security capabilities but also introduces new considerations (like managing API keys, cloud service roles, etc.). A strong cloud strategy often adopts the concept of a “Zero Trust” network – assuming no part of the network is inherently secure and requiring verification for every transaction – and uses the cloud’s identity features to implement that.

    • Architecture Layers and Cloud Services: It’s useful to recall the layered nature of cloud architecture. We can think of it in terms of layers similar to traditional software architecture [4]:
    • Physical layer: (in cloud, this is abstracted by the provider – but includes data center facilities, hardware, etc.)
    • Virtual/Resource layer: This corresponds to virtualization – the compute instances, virtual networks, storage volumes, etc. that form the basic resources.
    • Control layer: Cloud orchestration services that manage resources (auto-scaling controllers, container orchestrators like Kubernetes, etc.).
    • Service layer: Higher-level services like databases, messaging, AI services, etc., and the actual business services (applications) built on the cloud.
    • Application layer: The end-user applications delivered (which might themselves be composed of multiple cloud services beneath).
    • Presentation layer: The UI or API that the end-user interacts with (could be a web/mobile app consuming cloud APIs).
      Cloud Architecture Layers. Details in caption.
    Figure \(\PageIndex{1}\): Cloud Architecture Layers — From Physical Infrastructure to User Interface. This layered diagram maps the six architectural layers of a cloud system, showing how each builds upon the one below to deliver end-user applications. (CC BY 4.0; Felix Amoruwa) Access to a detailed description of Figure \(\PageIndex{1}\)

    Each layer has design considerations (e.g., at the resource layer, choose right VM sizes; at the service layer, choose between a managed database or self-hosting one on VMs; at the application layer, ensure good practices like caching frequently used data to reduce repeated calls). Earlier in the chapter, we discussed the OSI network model – it’s worth noting that cloud architecture still relies on those network layers (from physical links up to application protocols), but much of it is managed by cloud vendors. As an architect, you should ensure your strategy covers Layer 4-7 concerns (transport security, API design at the application layer, etc.) while the cloud provider handles Layer 1-3 (physical, data link, network routing, etc., although you configure virtual subnets and routing at layer 3/4 sometimes).

    To wrap up, let's consider an example that ties many of these design considerations together:

    Example Scenario: Suppose a retail enterprise is rearchitecting its e-commerce platform in the cloud. The strategy involves a microservices design (service orientation) running on Kubernetes (virtualization of infrastructure via containers). The company sets up in two AWS regions (geographic consideration for global customers, and redundancy for resiliency). They use Infrastructure as Code to provision all cloud resources (extensible provisioning), and each environment is configured via config files so it can be tweaked for development, testing, production (configurable enablement). They define standard data formats (JSON APIs for all services, unified info exchange) and implement a central event bus for communication between services. They also enforce quality and governance by using AWS Config rules to ensure, for example, no S3 buckets are open to the public and all data is encrypted (security governance), CloudWatch alarms to monitor performance (QoS governance), and AWS Budgets for cost control. The architecture includes an auto-scaling group for web servers (scalability), multiple availability zones usage for database and caching layers (redundancy and resiliency), and a continuous integration pipeline that automatically deploys tested code (to maintain quality in deployments).

    This scenario shows how principles and design considerations come together: the enterprise’s cloud strategy document would explicitly cover these aspects, often mapping each to specific implementations or policies. The result is a cloud environment that is robust, flexible, and aligned to business needs.


    This page titled 5.4: Software Architecture and Design Consideration for Cloud was last modified on Tue, 22 Sep 2026 21:13:25 GMT and is shared under a CC BY 4.0 license and was authored, remixed, and/or curated by Felix Amoruwa.