Alibaba ACA Cloud Business flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

ECS

Elastic Compute Service: Alibaba Cloud's IaaS virtual machine service offering scalable CPU, memory, and GPU instances across multiple instance families and regions.

OSS storage classes

Standard, Infrequent Access, Archive, Cold Archive, and Deep Cold Archive. Lower-cost classes have longer minimum storage durations and slower retrieval times.

SLB

Server Load Balancer distributes incoming traffic across multiple ECS instances. Types include Classic (L4), Application (L7), and Network (L4 high performance).

VPC

Virtual Private Cloud is an isolated network with a CIDR block, subnets (vSwitches) bound to one AZ, route tables, and security groups.

RDS

Relational Database Service is a managed database supporting MySQL, SQL Server, PostgreSQL, MariaDB, and PPAS, with backups, read replicas, and multi-AZ high availability.

ApsaraDB for Redis

Managed in-memory key-value store used for caching, session management, leaderboards, and pub/sub. Supports cluster mode and read/write splitting.

MaxCompute

A big-data warehouse service for batch SQL processing and analytics over petabytes of data, integrated with DataWorks for scheduling and governance.

CDN

Content Delivery Network caches static and dynamic content at edge nodes close to users, reducing latency and origin load. Origin can be OSS or ECS.

ESS

Elastic Auto Scaling automatically adjusts ECS capacity based on scaling rules such as target tracking, scheduled actions, or custom metrics.

Multi-AZ high availability

Deploying instances across multiple Availability Zones within a region so that a single zone failure does not take down the application.

Decoupling with message queue

Using Message Queue or Kafka to buffer and asynchronously process traffic between tiers, preventing cascading failures and smoothing load spikes.

Security groups vs ACLs

Security groups are stateful instance-level firewalls; network ACLs are stateless subnet-level filters. Both are used together for layered network security.

RTO and RPO

Recovery Time Objective is the maximum acceptable downtime; Recovery Point Objective is the maximum acceptable data loss. They drive DR strategy selection.

Reserved instances

A pricing model where committing to a term (e.g., one or three years) yields significant discounts versus pay-as-you-go, suited to steady workloads.

NAT Gateway

A managed service that lets ECS instances in private subnets access the internet for outbound traffic without exposing them to inbound connections.

Horizontal vs vertical scaling

Horizontal scaling adds more instances behind a load balancer; vertical scaling resizes a single instance. Horizontal is preferred for elasticity and fault tolerance.

CloudMonitor

Alibaba Cloud's monitoring service that collects metrics for cloud resources, sets alarms, and triggers notifications or automated actions.

ActionTrail

An audit service that records API calls and console actions in your account for security analysis and compliance reporting.

Server Migration Center

SMC migrates physical, virtual, or cloud servers to Alibaba Cloud ECS using a client agent, supporting incremental synchronization and controlled cutover.

DTS

Data Transmission Service migrates and synchronizes data between databases, supporting schema migration, full plus incremental sync, and continuous replication.

Migration Hub

A centralized console to plan, track, and manage migration projects, including server, database, and storage migrations to Alibaba Cloud.

Express Connect

A dedicated private network connection between an on-premises data center and Alibaba Cloud VPC, offering stable bandwidth and low latency for hybrid cloud.

E-commerce reference architecture

SLB plus auto-scaled ECS, Redis for caching and sessions, RDS for transactions, MQ for order decoupling, and CDN for static assets to absorb traffic spikes.

Pilot migration

Migrating a small representative subset of workloads first to validate design, performance, and rollback procedures before bulk migration.

Data Online Migration

A service that migrates large datasets from local storage or other cloud object stores into OSS using parallel transfer and checksum validation.

PAI

Platform for AI provides model development, distributed training (DLC), and inference deployment (EAS) with notebook environments and GPU compute.

Qwen

Alibaba Cloud's family of large language models available in multiple sizes (e.g., Turbo, Plus, Max) trading off cost, latency, and capability.

Model Studio

A platform (Bailian) offering API access to Qwen models plus tools for RAG, agent building, and model fine-tuning.

RAG

Retrieval-Augmented Generation retrieves relevant documents from a knowledge base, adds them to the prompt, and generates a grounded answer, reducing hallucination.

Vector search

Converting text to embeddings and storing them in a vector database to find semantically similar items for RAG, recommendation, or search.

EAS

Elastic Algorithm Service is PAI's inference deployment component that hosts models behind endpoints with auto scaling and monitoring.

GPU ECS instances

ECS instance families with NVIDIA GPUs for deep learning training and inference, often used with PAI for AI workloads.