NAVER Cloud Platform Certified Associate flashcards

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

Region and Zone

A Region is a geographic location hosting NCP services; a Zone is an isolated datacenter inside a Region. Deploying across Zones protects against single-datacenter failure.

VPC

Virtual Private Cloud is the isolated network boundary on NCP. Servers, load balancers, and databases are placed in VPC subnets, replacing the older Classic network model.

Public versus private subnet

A public subnet has a route to the Internet Gateway and can host internet-facing resources. A private subnet reaches the internet only through a NAT Gateway and is used for backends and databases.

Console, API, and CLI

Three access surfaces for NCP that share one identity and permission model. The console is the web UI, the API is REST-based, and the CLI wraps the API for scripting.

Server spec families

NCP server specs are grouped by vCPU-to-memory ratio, such as Standard, High Memory, High CPU, and GPU types. Choose based on workload profile, then resize by stopping or redeploying.

Auto Scaling components

A Launch Configuration defines the instance template, a Scaling Group sets min, max, and desired capacity plus subnets, and Scaling Policies define metric, schedule, or manual triggers.

Scaling cooldown

A cooldown period after a scaling action during which further scaling is suppressed. It prevents rapid oscillation when a metric hovers around the threshold.

Init script (user data)

A script executed at first boot of a server. It is the standard mechanism to install agents, mount storage, and register instances with monitoring in an Auto Scaling group.

NCP Kubernetes Service

A managed Kubernetes offering where NCP operates the control plane and customers manage node pools built from server specs. Worker nodes are billed as servers.

Object Storage

Bucket-based, S3-compatible object storage for files and blobs. It is Region-scoped, supports storage classes and lifecycle rules, and is accessed through an endpoint.

Block versus File Storage

Block Storage attaches to one server in the same Zone, like a disk. File Storage is shared over NFS by multiple servers and suits common filesystem needs.

Security Group versus NACL

Security Groups are stateful and attached to resources with allow-only rules. NACLs are stateless and applied at subnet level, requiring explicit inbound and outbound rules.

Load Balancer types

Application LB handles HTTP/HTTPS with L7 routing; Network LB handles TCP/UDP at L4 with high throughput. Health checks remove failing targets from rotation.

VPC Peering limits

Peering connects two VPCs privately but is not transitive. A third VPC cannot reach the peered pair through an existing peering without its own peering connection.

NAT Gateway

A managed gateway that lets resources in private subnets initiate outbound internet traffic without inbound exposure. Data processing through it is a billing factor.

Cloud DB for MySQL

A managed MySQL service provisioned as a separate DB server inside the VPC. It supports HA standby, automated backups, and access control through ACG rules.

DB HA standby

A standby database instance in a different Zone from the primary, with automatic failover. Applications connect through an endpoint so failover does not require code changes.

DB restore behavior

Restoring a managed DB backup creates a new instance rather than overwriting the existing one. Cutover to the restored instance is an application or DNS change.

Read replica tradeoff

A read replica offloads read queries from the primary but lags behind it. Only route queries that tolerate stale data to replicas, never critical read-after-write checks.

Cloud DB for Redis

A managed in-memory data store used for caching, session storage, and hot data. It complements, rather than replaces, a durable relational database.

Sub Account

An additional identity created under the main account with its own login and assigned policies. It avoids sharing root credentials and enables least-privilege access.

Policy-based authorization

Permissions are expressed as policies that allow specific actions on specific resources, with default deny. Custom policies grant narrower rights than administrator policies.

Access key rotation

API access uses an access key ID and secret key. Keys must be rotated periodically, scoped per application, and never committed to source code or embedded in clients.

ACG rules

Access Control Groups act as the firewall for VPC resources. Rules are allow-based, evaluated separately for inbound and outbound, and attached to resources.

MFA for Sub Accounts

Multi-factor authentication adds a second verification factor at login. Enforcing it for Sub Accounts reduces risk from stolen passwords, especially for administrator-level identities.

Monitoring agent requirement

Memory and disk metrics on a server require the monitoring agent installed on the OS. Without it, only hypervisor-level metrics such as CPU and network are visible.

Alarm notification channel

An alarm rule triggers on a metric threshold and must be bound to a channel such as email, SMS, or webhook. Without a channel the alarm fires silently.

Common cost drivers

Frequent billing surprises come from always-on servers, unattached block storage, retained snapshots, public IPs, NAT Gateway processing, and outbound data transfer.

Tagging for cost allocation

Consistent tags such as environment, owner, and cost center let billing and inventory reports be filtered. Untagged resources cannot be attributed to a team.

CLOVA AI APIs

Managed AI services such as Speech, OCR, Face Recognition, and Translation are called as APIs with credentials. Keys belong server-side, never in client code.

API Gateway role

A managed front door for backend APIs that terminates authentication, applies rate limits, and supports versioning. It centralizes cross-cutting concerns outside application code.

Hybrid connectivity choice

IPsec VPN Gateway suits quick, lower-cost links over the internet. A dedicated connection provides predictable bandwidth for steady, high-volume hybrid traffic.

Cutover with rollback

Run source and target in parallel, replicate data, validate, then switch DNS with a low TTL. Keep the old environment until validation passes so rollback stays possible.