HCIP-Cloud Computing flashcards
HCIP-Cloud Computing · Build recall with a focused review session.
33 cards. Read the term and the definition here. The deck at the top of this page is optional.
- ECS
- Elastic Cloud Server: a virtual machine on Huawei Cloud with configurable flavor, image, EVS disks, VPC network, and security groups, billed by subscription or pay-per-use.
- Flavor
- The ECS specification defining vCPU count, memory size, and instance family. Choosing a flavor fixes the compute-to-memory ratio and the maximum network and disk performance.
- Image
- A template used to create ECS instances, containing OS, preinstalled software, and disk configuration. Public, private, shared, and marketplace images are supported.
- EVS
- Elastic Volume Service: block storage disks attached to ECS. Types include high I/O, ultra-high I/O, general-purpose SSD, and standard, differing in IOPS and throughput.
- VPC
- Virtual Private Cloud: an isolated logical network with subnets, route tables, and security groups that ECS, BMS, and other resources join at creation time.
- Security group
- A stateful virtual firewall applied at the instance NIC level, controlling inbound and outbound traffic by protocol, port, and source or destination address.
- EIP
- Elastic IP: a public IP address that can be bound to or unbound from an ECS instance or NAT gateway, enabling internet access without rebuilding the instance.
- AS scaling group
- An Auto Scaling group that maintains a minimum, desired, and maximum instance count across subnets, adding or removing ECS instances per attached scaling policies.
- Scaling configuration
- The Auto Scaling template that defines flavor, image, disk, key pair, and security group used when the scaling group launches new ECS instances.
- Alarm-based policy
- An Auto Scaling policy that triggers scale-out or scale-in when a CES metric such as CPU utilization crosses a configured threshold for a defined period.
- BMS
- Bare Metal Server: a physical server offered as a cloud service with no hypervisor layer, used for high-performance databases, HPC, and workloads requiring hardware isolation.
- DeH
- Dedicated Host: a physically isolated host on which ECS instances are placed, supporting bring-your-own-license compliance and host-level resource guarantees.
- Spot instance
- A low-cost ECS billed at a market rate that can be reclaimed when capacity is needed, suited to interruption-tolerant batch and rendering jobs.
- IMS
- Image Management Service: stores and manages public, private, and shared images, supports image creation from ECS or EVS, and enables cross-region replication.
- Instance snapshot
- A point-in-time backup of an EVS disk used to restore data or create new disks and images; snapshots are incremental and stored independently of the instance.
- Availability Zone
- A physically isolated data center region within a cloud region with independent power and network, used to spread ECS instances for high availability.
- ECS lifecycle states
- Instance states include creating, running, stopped, rebooting, and error. A stopped instance still reserves its EVS disks and EIP unless released.
- User data
- A script or cloud-init configuration injected into an ECS at first boot to automate software installation and initial configuration without manual login.
- Cloud-init
- The guest agent that consumes ECS user data and metadata at boot, handling hostname, key injection, network configuration, and custom scripts.
- ECS metadata service
- An internal endpoint reachable from inside the instance that returns instance ID, region, AZ, and user data, used for automation and configuration discovery.
- Key pair
- An SSH public-private key pair used for passwordless Linux ECS login; the public key is stored by the platform and injected into the instance at creation.
- Instance type family
- A grouping of ECS flavors by workload profile, such as general-purpose, memory-optimized, compute-optimized, GPU, and high-performance families.
- GPU-accelerated ECS
- An ECS flavor with attached GPU cards for AI training, inference, rendering, and video transcoding, requiring matching driver and CUDA toolkit installation.
- ECS resizing
- Changing an instance's flavor to add or remove vCPU and memory; it requires stopping the instance and is constrained by the underlying host capacity.
- ELB
- Elastic Load Balance: distributes traffic across backend ECS instances using listeners and health checks, supporting layer 4 TCP/UDP and layer 7 HTTP/HTTPS.
- Health check
- An ELB probe that periodically tests a backend instance and removes it from rotation when it fails, restoring it after consecutive successful probes.
- Anti-affinity group
- An ECS placement rule that spreads instances across different physical hosts to reduce correlated failure, or packs them together for low-latency communication.
- Subscription billing
- A prepaid ECS billing mode purchased for a fixed term, typically monthly or yearly, offering lower cost than pay-per-use for steady workloads.
- Pay-per-use billing
- A postpaid ECS billing mode charged by actual usage time, allowing release at any moment and suited to short-lived, variable, or experimental workloads.
- ECS release
- Deleting an instance. Attached EVS disks and EIPs are released or retained per the chosen option, so data on system disks is lost unless imaged.
- Cloud-init user data limit
- User data is capped in size, commonly around 32 KB after base64 encoding, so large bootstrap payloads should be fetched from OBS instead.
- EVS disk throughput
- The maximum sustained read and write rate of a disk, which scales with disk capacity and type and caps the achievable performance of the attached ECS.
- ECS monitoring
- CES collects ECS metrics such as CPU, memory (with agent), disk, and network, driving alarms and Auto Scaling policies for automated operations.