Emerging Technologies and Enterprise Applications
Domain 8 asks you to place a technology in the right layer and justify the choice by requirements, not by hype. Questions mix cloud service models, big-data components, NoSQL families and enterprise application integration.
Cloud and virtualization
- NIST service models: IaaS provides virtual machines, storage and networks; PaaS provides a managed runtime and middleware; SaaS provides a finished application. Deployment models are private, public, community and hybrid.
- Essential characteristics: on-demand self-service, broad network access, resource pooling, rapid elasticity and measured service.
- Type 1 hypervisors run on bare metal; type 2 run on a host OS. Containers share the host kernel, start faster and isolate less than virtual machines.
Data platforms and intelligence
- Big data characteristics are volume, velocity, variety, veracity and value. HDFS stores replicated blocks with a NameNode holding metadata; YARN schedules; MapReduce and Spark process. Spark keeps data in memory and plans execution as a DAG, so it outperforms MapReduce on iterative workloads.
- NoSQL families: key-value, document, column-family and graph. CAP forces a choice between consistency and availability during a partition; BASE trades consistency for availability.
- Machine learning pipeline: data preparation, feature engineering, training and validation, evaluation (precision, recall, F1, overfitting checks), deployment and monitoring.
- Blockchain is a replicated ledger with consensus and immutable chaining; it fits multi-party trust without a central authority, not high-throughput private data.
Enterprise applications
- ERP integrates finance, HR, manufacturing and supply chains around a shared database; CRM manages the customer lifecycle; SCM manages supplier networks; BI and data warehouses support decisions.
- SOA exposes coarse-grained, reusable services, often over an ESB; microservices are finer-grained, independently deployable and organised by bounded context, at the cost of distributed-system complexity.
- Middleware families: message-oriented middleware, remote procedure calls, transaction processing monitors, object request brokers. IoT layers: perception, network, platform, application; edge computing moves latency-sensitive processing near devices.
Worked example
A retailer wants a bespoke inventory service with automatic scaling and no operating-system patching; a small firm wants email with no development effort. The first is PaaS, because the provider manages the runtime and OS while the team supplies code; IaaS would leave patching with the team and SaaS cannot host bespoke code. The second is SaaS, since the application is consumed as a service.
Traps
- Describing a virtual machine as a container.
- Claiming blockchain makes data truthful; it only makes recorded data tamper-evident.
- Equating SOA with web services alone.
- Treating partition tolerance as optional in CAP.