ACA CCE key terms

A geographic area where Alibaba Cloud deploys its data centers, chosen by customers to place resources close to users and to satisfy data residency requirements.

An isolated physical location with independent power and network within a region, used to spread resources so a single failure does not take down an entire application.

The ability to automatically add or remove computing resources as demand rises or falls, so capacity tracks workload instead of being fixed in advance.

An architecture pattern that places redundant instances in two or more zones of one region so the service survives the loss of a single zone.

A logically isolated private network in the cloud where you define IP ranges, subnets, route tables, and gateways for your resources.

A subnet inside a VPC that holds a range of private IP addresses and determines which zone the attached cloud resources are created in.

The division of security duties between the cloud provider and the customer, where the provider secures the platform and the customer secures what they put on it.

Alibaba Cloud's core virtual machine service, providing resizable compute capacity with a choice of instance types, images, disks, and networking.

A predefined hardware profile for an ECS instance that bundles CPU, memory, and network capability, grouped into families such as general purpose or compute optimized.

A template containing an operating system and optional preinstalled software used to create ECS instances quickly and consistently.

A service that automatically adjusts the number of ECS instances based on rules or schedules, matching capacity to load and reducing cost during quiet periods.

A logical collection of ECS instances managed together by Auto Scaling, sharing configuration, minimum and maximum size, and scaling rules.

A managed service that distributes incoming traffic across multiple backend servers to improve availability and handle higher request volumes.

A highly durable, scalable storage service for unstructured data such as images, videos, and backups, organized into buckets and accessed over HTTP APIs.

The top-level container in OSS that holds objects, has a globally unique name, and is associated with a region and access policy.

A durability and cost tier for stored objects, such as Standard, Infrequent Access, Archive, and Cold Archive, chosen to match access frequency.

A managed relational database service that handles provisioning, backups, patching, and failover for engines such as MySQL, SQL Server, and PostgreSQL.

A replica of an RDS primary database that serves read queries, offloading traffic from the primary and improving read scalability.

A content delivery network that caches static and dynamic content at edge nodes near users, reducing latency and origin bandwidth consumption.

A managed domain name resolution service that maps names to IP addresses and supports routing policies such as latency-based and geo-based resolution.

A stateful virtual firewall attached to ECS instances that allows or denies inbound and outbound traffic by protocol, port, and source or destination.

An identity and permission service for controlling which users, groups, and roles can perform which actions on which Alibaba Cloud resources.

An identity assumed temporarily by trusted entities such as ECS instances or applications, granting permissions without storing long-term access keys.

A JSON document that specifies allowed or denied actions, resources, and conditions, attached to RAM users, groups, or roles to define permissions.

A credential pair of an identifier and secret used by programs to call Alibaba Cloud APIs, which must be protected and rotated like a password.

A free baseline protection automatically provided to Alibaba Cloud resources that detects and scrubs common volumetric denial-of-service attacks.

A service that inspects HTTP and HTTPS traffic to block attacks such as SQL injection and cross-site scripting before they reach the application.

A managed platform that gives developers API access to foundation models for text, image, and other modalities, with tools for tuning and deployment.

A large pretrained model, such as the Qwen family, that can be adapted to many downstream tasks through prompting or fine-tuning rather than training from scratch.

Further training of a pretrained model on domain-specific examples so its outputs better match a particular task, style, or body of knowledge.

A unit of text, roughly a word fragment, that a language model processes; usage and billing are usually measured in input and output tokens.

A billing model where charges accrue based on actual resource usage over time, with no upfront commitment and no long-term contract.

A billing model where a resource is paid for upfront for a fixed term, typically at a lower unit price than pay-as-you-go.

The top-level identity that owns all resources and billing, signs contracts, and can create RAM users under it for day-to-day access.

A tiered support offering that determines response times, channels, and technical account services available to a customer, from basic to enterprise levels.