Huawei HCIP-Cloud Service Solutions Architect key terms

A geographic area where Huawei Cloud operates data centers, chosen to place workloads near users and to satisfy data-residency rules. Regions are isolated from one another, so cross-region traffic uses the public network or a dedicated connection.

One or more physically separate data centers inside a region, with independent power, cooling and network. AZs in a region are linked by low-latency fiber, enabling cross-AZ redundancy within a single region's management scope.

A virtual machine on Huawei Cloud whose vCPU, memory, image, disks and network are chosen at creation and can be changed later. Instances run inside a VPC, support several billing modes and can be scaled automatically.

A dedicated physical server rented on Huawei Cloud, giving the performance and isolation of a data-center machine without owning hardware. It suits databases, high-performance computing and workloads requiring bare-metal access or specific licensing.

A service that automatically adjusts how many ECS instances a group contains, based on schedules, alarm-triggered policies or a fixed count. It adds capacity at peak load and removes it afterwards, working with ELB to register and deregister instances.

A managed Kubernetes service where Huawei Cloud operates the control plane and you run containerized workloads on worker nodes or serverless containers. It provides auto-scaling, service discovery and integration with VPC, ELB and storage services.

Block storage for ECS and BMS, delivered as network-attached disks that persist independently of the instance lifecycle. Volumes come in performance tiers and can be snapshotted, cloned and, for some types, shared by several servers.

A massively scalable object store reached through HTTP/REST APIs, used for backups, media, static websites, archives and data-lake input. Objects live in buckets and are protected by redundant storage across facilities, with lifecycle rules for aging data.

Tiers that trade retrieval speed and availability against price: Standard for frequent access, Infrequent Access for cooler data, and Archive for long-term retention with hours-long restore. Lifecycle rules can move objects between classes automatically.

A shared file system that many ECS instances or containers can mount simultaneously over NFS or CIFS. Capacity and performance scale elastically, making it suitable for content sharing, web serving and persistent container volumes.

A backup service that copies EVS volumes, BMS disks, SFS file systems and some databases into vaults on a schedule or on demand, enabling point-in-time restore after accidental deletion, corruption or ransomware.

A logically isolated virtual network that you define using IP address ranges, subnets, route tables and gateways. Instances inside it communicate privately, and connectivity to other networks is added explicitly through gateways, peering or tunnels.

A public IP address that can be bound to an ECS, BMS, ELB, NAT gateway or container and later unbound, keeping the same address when reassigned. It provides internet access, with bandwidth billed separately from the resource.

A load balancer that spreads incoming traffic across backend servers using layer 4 (TCP/UDP) or layer 7 (HTTP/HTTPS) listeners. It performs health checks, stops sending traffic to unhealthy backends and can terminate TLS certificates.

A managed service that lets instances in private subnets open outbound internet connections, and optionally allows inbound access through port forwarding, without assigning a public IP to every server.

A link that joins two VPCs so their subnets reach each other privately, even across accounts or regions, without traversing the internet. Route table entries must be added in both VPCs before traffic flows.

A dedicated private physical link between an on-premises data center and Huawei Cloud, avoiding the internet for predictable latency and bandwidth. A virtual gateway and virtual interfaces attach the link to one or more VPCs.

An IPsec-encrypted tunnel over the internet between a customer gateway on premises and a VPN gateway in Huawei Cloud. It is faster and cheaper to deploy than Direct Connect, but quality depends on public internet conditions.

The cloud service that hosts public and private DNS zones and resolves names to IP addresses. Private zones answer queries inside VPCs, letting instances refer to stable names rather than hard-coded addresses.

The service controlling who may use which cloud resources and how. Users, user groups, policies, roles and agencies enforce least privilege, and temporary credentials can be delegated to applications and services.

A stateful virtual firewall attached to instances that filters traffic by protocol, port and source or destination. Only explicitly configured rules are permitted; default rules generally allow outbound traffic and block unsolicited inbound traffic.

A managed service for creating and protecting cryptographic keys. Keys remain inside hardware security modules, and envelope encryption with data keys lets you encrypt OBS objects, EVS volumes and database storage.

A reverse-proxy defense for web applications that inspects HTTP and HTTPS requests and blocks SQL injection, cross-site scripting, malicious crawlers and known exploits. It can also enforce precise access control and rate limits.

An endpoint protection service for servers and containers that detects intrusions, malware, brute-force attacks and risky configurations, and reports file-integrity changes and vulnerabilities to a central console.

A centralized, cloud-native firewall that controls north-south traffic at the VPC border and east-west traffic between VPCs, adding intrusion prevention and unified policy management across multiple accounts.

A managed relational database offering for MySQL, PostgreSQL and SQL Server that automates provisioning, patching, backup and failover. You select instance class and storage, and a standby copy can be kept in another AZ.

A managed in-memory cache compatible with Redis and Memcached, used to offload read pressure from databases and store session state. It supports master/standby and cluster modes with automatic failover.

A managed massively parallel processing warehouse with columnar storage, built for large analytical SQL, reporting and aggregation. Nodes scale horizontally for concurrency, and some editions separate compute from storage.

A middleware layer that shards data across several RDS for MySQL instances and presents them as one logical database. It handles routing, cross-shard queries and scale-out so applications need fewer code changes.

A serverless analytics service that runs SQL and Spark jobs directly against data held in OBS without provisioning clusters. Billing follows compute consumed, which fits ad-hoc exploration and periodic batch processing.

Placing redundant instances or database standbys in separate availability zones of one region so that a single data-center failure does not stop the service. It guards against infrastructure faults but not region-wide disasters.

Recovery point objective is the maximum tolerable data loss expressed in time; recovery time objective is the maximum tolerable downtime. Together they determine which replication method and disaster recovery strategy a design requires.

A service that continuously replicates EVS volumes of protected servers to a DR site and can boot them there on demand. It supports cross-AZ and cross-region protection with orchestrated drills and controlled failover.

A migration tool that installs an agent on a source physical or virtual server, replicates its disks to Huawei Cloud and launches a matching ECS or BMS. Incremental synchronization keeps the copy current before cutover.

A database migration and synchronization service that moves data between on-premises or other-cloud databases and Huawei Cloud RDS with minimal downtime. It performs full plus incremental replication and schema consistency checks.

The monitoring service that collects metrics and events for cloud resources and custom applications and evaluates alarm rules. Alarms can send notifications, trigger auto scaling or launch automated remediation actions.

An audit service that records API operations on your account, showing who did what, when and from where. Traces are retained in a designated bucket for compliance reviews and troubleshooting.

A template-driven service that describes cloud resources and their dependencies so complete environments can be deployed, updated and removed repeatedly. It replaces manual console configuration and keeps environments consistent across regions.

Huawei Cloud charges pay-per-use for flexible short-term workloads and yearly/monthly subscription for steady long-running ones at a discount. Matching each workload to the right mode and right-sizing instances are the largest cost levers.

A key-value label attached to cloud resources that supports grouping, filtering and cost allocation reports. Consistent tagging lets teams attribute spend to projects or departments and drive automated operations and policy.