CAP-C01 Alibaba Cloud Certified Professional: Cloud Architect Questions and Answers

Tair (Redis OSS-Compatible) is the appropriate centralized session-state layer for an elastic, distributed application. When ECS instances frequently scale in and out, application sessions must not depend on the lifecycle of any individual compute node. By storing session tokens and session state in Tair, every ECS instance can retrieve the same session information, making the web/application tier effectively stateless and compatible with Auto Scaling.

Alibaba Cloud Tair is a fully managed Redis-compatible database designed for high throughput, low latency, elastic scalability, and high availability. Alibaba Cloud also documents session-oriented Tair use cases in which session tokens are stored as Redis/Tair data structures with TTL values, demonstrating precisely why the service fits distributed session management.

Session stickiness is weaker because it attempts to keep a user on a specific backend and therefore creates affinity to an instance that might be removed during scaling or failover. CloudMonitor is an observability service, not a session store. STS issues temporary Alibaba Cloud credentials and does not maintain application user sessions.

Consequently, separating session state from ECS compute by storing it in Tair gives the architecture the elasticity and fault tolerance required.

Study Guide reference: Distributed Architecture on Alibaba Cloud; Performant Architecture; Core Infrastructure Deep Dive.

==================================================

The application needs asynchronous decoupling between high-volume ingestion and database writes. Simple Message Queue provides exactly this load-leveling pattern. The first Function Compute function can acknowledge incoming requests quickly and place each item into SMQ. A second function consumes the queued messages and writes them to RDS at a rate the database can sustain.

SMQ is a distributed managed messaging service specifically designed to decouple producers and consumers. Alibaba Cloud identifies traffic-spike buffering, asynchronous processing, and independent component scaling as core SMQ use cases. Queue messages are delivered at least once during their validity period, significantly improving durability compared with directly coupling every incoming API request to a database write.

Function Compute also integrates directly with SMQ queues and topics as event sources, reducing custom integration and infrastructure management.

Kafka could also provide durable asynchronous messaging, but it is a more elaborate streaming platform than necessary for this workload and therefore carries additional configuration complexity. Moving to ECS reintroduces server administration, while replacing RDS does not solve the ingestion decoupling problem.

Study Guide reference: Core Infrastructure Deep Dive — Function Compute, messaging architectures, asynchronous processing, and scalable application integration.

==================================================

Option B reduces interference between transaction processing and read-heavy workloads inside PolarDB. PolarDB ' s read/write splitting automatically routes writes to the primary node while distributing eligible reads across read-only nodes. This allows reporting and ordinary SELECT traffic to consume separate compute capacity instead of competing with critical transactional writes on the primary node. Alibaba Cloud specifically identifies reporting, analytics, and read-heavy applications as appropriate read/write-splitting scenarios.

Option D provides stronger isolation for computationally expensive analytics. DTS can continuously synchronize PolarDB for MySQL data into MaxCompute using initial full synchronization followed by ongoing incremental change capture. Analysts can then execute large-scale analytical workloads against MaxCompute without placing those queries on the production transactional database.

Running queries on ECS does not inherently remove load from PolarDB because the SQL operations would still execute against the database. Introducing MongoDB merely to execute analytical queries creates a second database technology without an appropriate replication or analytical architecture.

Together, PolarDB read/write splitting optimizes operational read traffic while DTS-to-MaxCompute isolates heavyweight analytical processing, producing an effective HTAP-style separation of concerns.

Study Guide reference: Core Infrastructure Deep Dive — PolarDB, read/write splitting, DTS, MaxCompute, and analytical workload offloading.

==================================================

CDN should form the content-delivery layer for cacheable video assets. Alibaba Cloud CDN distributes content to geographically dispersed edge nodes so viewers retrieve cached objects from nearby POPs rather than repeatedly connecting to a distant origin. Alibaba Cloud explicitly supports video-oriented CDN acceleration and edge delivery, reducing both end-user latency and origin bandwidth demand.

Auto Scaling addresses peak-time compute demand. It monitors workload conditions and automatically increases ECS capacity during traffic spikes, then scales resources back when traffic decreases. Alibaba Cloud specifically identifies streaming platforms with unpredictable demand as a typical Auto Scaling scenario.

Global Accelerator complements CDN and Auto Scaling where viewers or application endpoints are separated by long-distance or cross-border networks. GA routes traffic through Alibaba Cloud ' s optimized global BGP network, reducing latency, jitter, and packet loss. Alibaba documentation explicitly identifies video streaming as a latency-sensitive scenario that benefits from this architecture.

Tair can improve authentication or session performance, but it does not directly address video delivery quality or regional network performance. Therefore, CDN + Auto Scaling + Global Accelerator most directly satisfies the streaming requirement.

Study Guide reference: Delivering Services and Content on Alibaba Cloud — CDN, Global Accelerator, Auto Scaling, video delivery, and global traffic optimization.

==================================================

The core problem is durability during transient processing or network failures. A topic provides publish/subscribe notification delivery, but placing an SMQ queue behind the topic creates a persistent buffer between the notification producer and the Function Compute consumer. Alibaba Cloud supports subscribing a queue to an SMQ topic so that every message published to the topic is copied into the queue and retained for consumption.

Function Compute should then process records from the queue. Queue semantics provide the required failure recovery: after a consumer receives a message, it becomes temporarily invisible. The consumer deletes the message only after successful processing. If processing fails—for example because of a temporary connectivity problem—the message is not deleted and becomes visible again after the visibility timeout, allowing another processing attempt.

Function Compute also supports SMQ queue triggers with configurable retries, fault-tolerance behavior, and dead-letter handling, which can further strengthen this architecture.

Increasing memory does not fix network reliability, and multi-region deployment alone does not guarantee message persistence. A durable queue therefore prevents transient failures from silently losing ingestion work.

Study Guide reference: Core Infrastructure Deep Dive — Simple Message Queue, Function Compute, durable asynchronous processing, retries, and decoupled event-driven architecture.

Global Accelerator is designed for precisely this class of global interactive application. Clients connect through Alibaba Cloud ' s nearby acceleration access points, after which traffic traverses Alibaba Cloud ' s optimized global network toward the most appropriate backend region. This reduces Internet routing variability and improves end-to-end latency for latency-sensitive workloads such as gaming.

GA listeners define the application protocol and ports, while endpoint groups represent backend deployments in different regions. With intelligent routing, GA can select a nearby healthy endpoint group and forward requests accordingly. Alibaba Cloud also provides endpoint health checks; when an endpoint becomes unhealthy, new requests can automatically be directed to healthy endpoints.

CDN is highly effective for static cacheable content but is not the correct primary mechanism for arbitrary real-time game traffic and regional backend health-based routing. RDS and Tair improve application data performance but do not solve global network acceleration.

Therefore, GA in front of regional ALBs provides the correct architecture: nearest-entry acceleration, Alibaba Cloud backbone transport, regional endpoint groups, health monitoring, and failover.

Study Guide reference: Delivering Services and Content on Alibaba Cloud — Global Accelerator, ALB, multi-region traffic management, and high availability.

The ECS security group must allow inbound HTTPS traffic. For an inbound security-group rule, the Internet clients are the source, so the correct rule permits TCP destination port 443 from source 0.0.0.0/0. This makes Option C correct. Alibaba Cloud security groups are stateful, which means response traffic belonging to an allowed connection is automatically permitted and does not require a matching outbound security-group rule.

Network ACLs behave differently. Alibaba Cloud VPC network ACLs are stateless. If the ACL permits an inbound HTTPS request but does not permit the corresponding return packet, the TCP session cannot operate correctly. The inbound ACL therefore needs TCP 443 from Internet clients, while the outbound ACL must permit return traffic to client ephemeral destination ports. Alibaba Cloud explicitly recommends 1024–65535 when all client ephemeral-port implementations must be supported.

Option D permits only the incoming side of the ACL and therefore fails because response packets remain blocked. Option E incorrectly assumes that outbound server responses use destination port 443; the server ' s source port is 443, but the destination is the client ' s ephemeral port.

Study Guide reference: Securing Workloads on Alibaba Cloud — stateful Security Groups, stateless network ACLs, HTTPS exposure, and ephemeral ports.

==================================================

VPC peering provides direct private communication between two VPCs without exposing either application or database traffic to the public Internet. Alibaba Cloud VPCs are isolated by default. After creating a peering connection and configuring the required bidirectional routes, resources in VPC A and VPC B can communicate using private IP addresses. VPC peering supports same-account and cross-account connections as well as same-region and cross-region configurations.

This is materially more secure than exposing the database publicly. The application ' s existing EIP does not need to participate in database connectivity; the application can use its ECS private address when communicating across the peering link. Security-group and database whitelist rules should then restrict access to only the required source CIDRs and ports.

Option A intentionally routes database access through public addressing. Option B introduces an unnecessary proxy server, additional management overhead, and another failure point. Option D creates the largest attack surface by making the database Internet-accessible.

The design principle is straightforward: when workloads in two Alibaba Cloud VPCs require direct communication, establish private network connectivity and keep sensitive database traffic off the Internet.

Study Guide reference: Building Enterprise-grade Networks on Alibaba Cloud — VPC isolation, VPC peering, routing, and private database access.

Time Series Database is specifically designed for timestamped telemetry generated by IoT devices, monitoring systems, and industrial equipment. Alibaba Cloud identifies IoT device monitoring as a core TSDB scenario, including continuously collecting, storing, analyzing, and querying device-status and business-event data in real time.

ApsaraMQ for Kafka provides the high-throughput ingestion and buffering layer. It is a managed distributed Kafka platform intended for real-time data pipelines, monitoring-data aggregation, log collection, streaming processing, and large-scale analytics. Its partitioned distributed architecture lets producers ingest very large event volumes while downstream consumers process the streams independently.

Function Compute supplies elastic event processing without persistent compute servers. Alibaba Cloud supports ApsaraMQ for Kafka triggers that invoke Function Compute automatically whenever new records arrive in Kafka topics. This enables functions to perform transformations, filtering, enrichment, alert generation, or routing in near real time.

EMR is valuable for Hadoop/Spark-style batch analytics, but the question specifically emphasizes low-latency real-time ingestion and analysis. It therefore does not replace the streaming components selected above.

Study Guide reference: Core Infrastructure Deep Dive — IoT telemetry, TSDB, ApsaraMQ for Kafka, Function Compute, and real-time streaming architectures.

==================================================

Option B provides the simplest consolidated content-delivery architecture. Alibaba Cloud CDN can use multiple origins, and current CDN functionality supports origin-selection rules based on parameters such as request path, headers, query strings, and cookies. This allows static requests to be associated with an OSS origin while application requests are forwarded toward the ALB-backed application tier.

Static objects stored in OSS benefit directly from CDN edge caching. Requests are routed to nearby CDN points of presence, and cached resources can be returned without repeatedly accessing the OSS origin, reducing both latency and origin load.

Dynamic responses are generally not cached because they are generated uniquely for requests. They can still be routed to the application origin according to the configured origin rules.

The Global Accelerator alternatives create an unnecessarily fragmented architecture and attempt to use separate acceleration paths where one CDN-facing domain and origin-routing configuration can satisfy the question with fewer components. They also introduce additional DNS and acceleration configuration.

Therefore, a single CDN delivery layer with OSS and ALB-backed origins, combined with Alibaba Cloud DNS directing the application domain to CDN, produces the lowest configuration overhead.

Study Guide reference: Delivering Services and Content on Alibaba Cloud — CDN, OSS acceleration, ALB origins, and intelligent content delivery.

==================================================

Express Connect with redundant connections is the strongest match for a mission-critical hybrid network requiring private connectivity, high availability, and enterprise-scale bandwidth. Express Connect establishes dedicated, secure, high-speed private connectivity between an on-premises data center and Alibaba Cloud VPC resources rather than transporting application traffic over the public Internet.

Alibaba Cloud ' s current Express Connect guidance identifies disaster recovery and high availability as primary enterprise scenarios. Redundant circuits and connection paths can protect connectivity against failures involving leased lines, devices, fiber cuts, and access points. Alibaba also identifies highly elastic, high-availability architectures based on dedicated connectivity and ECMP for environments with significant traffic requirements.

VPC peering connects VPC networks and does not independently provide an on-premises dedicated access circuit. VPN Gateway can connect an on-premises network to a VPC securely, but it uses Internet-based VPN transport and is generally not the strongest option where dedicated private connectivity, large scalable bandwidth, and robust path redundancy are explicit requirements. NAT Gateway provides address translation and Internet access functions rather than private hybrid connectivity.

Therefore, redundant Express Connect paths provide the enterprise-grade hybrid networking characteristics described.

Study Guide reference: Building Enterprise-grade Networks on Alibaba Cloud; hybrid connectivity, resilient networking, and Express Connect design.

==================================================

Option C provides the lowest-change migration path for the application tier. Rehosting the existing Windows workload on Windows Server ECS instances preserves its operating-system environment while distributing instances across availability zones improves resilience against zone-level infrastructure failures.

For the database tier, Option A provides the stronger managed architecture. Alibaba Cloud explicitly supports using DTS to migrate self-managed Oracle databases to PolarDB for PostgreSQL (Compatible with Oracle), including schema, full-data, and incremental-data migration for minimal migration downtime.

PolarDB ' s Oracle-compatible edition is specifically engineered to run Oracle workloads with minimal application modification. It supports Oracle-compatible data types, SQL constructs, PL/SQL features, OCI connectivity, packages, sequences, synonyms, and related functionality.

It also provides built-in high availability: failures are detected automatically and workloads can fail over to healthy nodes. This provides a more complete managed availability architecture than simply deploying independent self-managed Oracle servers across zones, which would additionally require Oracle-level replication/failover configuration.

Options B and D require substantial application refactoring and therefore conflict with the explicit requirement to minimize development changes.

Study Guide reference: Building Highly Available, Performant Cloud Architecture — workload migration, ECS multi-zone design, DTS, and PolarDB Oracle compatibility.

==================================================