Skip to content
Certification Guide

Glossary

Plain-language definitions, each linked to the lesson that teaches the term.

A

Accelerated Site-to-Site VPN
A Site-to-Site VPN option that uses Global Accelerator to carry tunnel traffic over the AWS edge. It is available only on transit gateway attachments and must be chosen when the connection is created.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
ACM public certificate
A certificate from ACM that browsers trust. ACM renews it automatically while its validation record stays in place. Integrated services such as load balancers and CloudFront use it directly, and an exportable public certificate can also run on your own server.
Lesson: Data Security Controls (task 1.3)
Active/passive failover
A failover design in which the recovery site does not serve traffic until a failover event moves it there.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Adaptive capacity
DynamoDB's automatic rebalancing so frequently accessed items don't share a partition. It can't push one key past a partition's limit of 3,000 read units or 1,000 write units per second.
Lesson: High-Performing Databases (task 3.3)
Alias record
A Route 53 record that points at an AWS resource, such as a load balancer or CloudFront distribution. Alias queries to AWS resources have no query charge.
Lesson: Network Cost (task 4.4)
Amazon Cognito identity pool
A service that exchanges a trusted token for temporary, limited-privilege AWS credentials through an IAM role.
Lesson: Secure Workloads and Applications (task 1.2)
Amazon Cognito user pool
A user directory that signs in app users, including through social and enterprise providers, and issues JSON web tokens for your app or API.
Lesson: Secure Workloads and Applications (task 1.2)
Amazon GuardDuty
A threat detection service that analyzes CloudTrail management events, VPC flow logs and DNS logs to report suspicious activity.
Lesson: Secure Workloads and Applications (task 1.2)
Amazon Inspector
A vulnerability management service that scans EC2 instances, ECR container images and Lambda functions for software vulnerabilities and network exposure.
Lesson: Secure Workloads and Applications (task 1.2)
Amazon Macie
A data security service that discovers sensitive data in Amazon S3 and reports S3 buckets that become publicly accessible.
Lesson: Secure Workloads and Applications (task 1.2)
Amazon MSK
Amazon Managed Streaming for Apache Kafka: runs Apache Kafka clusters for you. The exam answer when a team already uses Kafka tools and code.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Archive restore
A request that makes a temporary readable copy of an object stored in S3 Glacier Flexible Retrieval, S3 Glacier Deep Archive or an Intelligent-Tiering archive tier. The object itself stays archived.
Also called: RestoreObjectLesson: Cost-Optimized Storage (task 4.1)
Aurora Capacity Unit (ACU)
The unit Aurora Serverless v2 uses to measure capacity. You set a minimum and a maximum number of ACUs and pay per second for the ACUs used.
Lesson: Cost-Optimized Databases (task 4.3)
Aurora global database
An Aurora deployment with a primary Region and secondary Regions replicated at the storage layer. A secondary cluster can be promoted during a Region-level failover.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Aurora I/O-Optimized
An Aurora cluster storage configuration with no per-request charge for read and write I/O, paid for with higher instance and storage rates.
Lesson: Cost-Optimized Databases (task 4.3)
Aurora Replica
A read-only Aurora instance that reads the same cluster volume as the writer. A cluster supports up to 15 of them, and their lag is typically in the tens of milliseconds within a Region.
Lesson: High-Performing Databases (task 3.3)
Aurora Serverless v2 auto-pause
The Aurora Serverless v2 behavior that pauses a cluster after a period with no connections when its minimum capacity is 0 ACUs, and resumes it on the next connection.
Lesson: Cost-Optimized Databases (task 4.3)
Aurora Standard
The default Aurora cluster storage configuration, with lower instance and storage rates and a charge for each I/O request.
Lesson: Cost-Optimized Databases (task 4.3)
Auto Scaling group
A set of EC2 instances launched from one launch template that Amazon EC2 Auto Scaling keeps between a minimum and a maximum size, replacing unhealthy instances and changing the desired capacity through scaling policies.
Lesson: High-Performing and Elastic Compute (task 3.2)
Automatic key rotation
A KMS setting that creates new key material for a customer managed key on a schedule, by default every year. Old key material stays available to decrypt existing data.
Lesson: Data Security Controls (task 1.3)
AWS Auto Scaling scaling plan
One place to configure scaling across several resource types in an application, such as EC2 Auto Scaling groups, ECS services and DynamoDB tables.
Lesson: High-Performing and Elastic Compute (task 3.2)
AWS Backup vault lock
A protection on an AWS Backup vault that stops deletion of recovery points and shortening of their retention. In compliance mode it becomes immutable after a 72-hour cooling-off period.
Lesson: Data Security Controls (task 1.3)
AWS Batch
A managed service that queues, schedules and runs batch jobs to completion on EC2 or Fargate capacity, without batch software to install or manage.
Lesson: High-Performing and Elastic Compute (task 3.2)
AWS Budgets
A service that tracks cost or usage against an amount you set and sends alerts when actual or forecasted spend crosses a threshold.
Lesson: Cost-Optimized Storage (task 4.1)
AWS Compute Optimizer
An opt-in service that analyzes utilization metrics (14 days by default, 93 days with paid enhanced infrastructure metrics) and recommends sizes for EC2 instances, Auto Scaling groups, EBS volumes, Lambda functions, ECS services on Fargate and more. It sees EC2 memory only through the CloudWatch agent or a supported observability tool.
Also called: Compute OptimizerLesson: Cost-Optimized Compute (task 4.2)
AWS Cost and Usage Report
The most detailed billing data AWS provides: line items for each product, usage type and operation, delivered to an S3 bucket you own and typically analyzed with Amazon Athena.
Also called: CURLesson: Cost-Optimized Storage (task 4.1)
AWS Cost Explorer
A console tool for exploring your AWS costs and usage over time, filtered and grouped by service, account or activated cost allocation tag.
Lesson: Cost-Optimized Storage (task 4.1)
AWS DMS
AWS Database Migration Service: moves data between databases, including ongoing replication so the source can stay live during the move.
Lesson: High-Performing Databases (task 3.3)
AWS Global Accelerator
A global service that gives you two static anycast IP addresses and routes client TCP or UDP traffic from the nearest edge location, over the AWS global network, to load balancer or EC2 endpoints in one or more Regions.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
AWS Graviton
AWS-designed Arm-based (arm64) processors used in EC2 instance types such as m7g and c7g, and available for Lambda and Fargate. Software that runs on Arm usually gets better price performance on them.
Also called: GravitonLesson: Cost-Optimized Compute (task 4.2)
AWS managed key
A KMS key that an AWS service creates and manages in your account, such as aws/s3. You cannot edit its key policy or disable it.
Lesson: Data Security Controls (task 1.3)
AWS PrivateLink
The AWS technology behind interface endpoints, which connects a VPC to AWS services privately, without an internet gateway.
Lesson: Secure Workloads and Applications (task 1.2)
AWS Secrets Manager
A service that stores secrets such as database credentials and API keys, encrypts them with AWS KMS, and can rotate them automatically.
Lesson: Secure Workloads and Applications (task 1.2)
AWS Shield Advanced
A paid subscription that adds enhanced DDoS detection, automatic application layer mitigation, the Shield Response Team and cost protection.
Lesson: Secure Workloads and Applications (task 1.2)
AWS Shield Standard
Automatic DDoS protection at no additional charge for all AWS customers, against common network and transport layer attacks.
Lesson: Secure Workloads and Applications (task 1.2)
AWS Storage Gateway
A service that runs an appliance on premises (or as a VM) to give local applications file, volume or tape access to storage in AWS, with a local cache.
Lesson: High-Performing and Scalable Storage (task 3.1)
AWS WAF web ACL
A set of AWS WAF rules attached to a protected resource, such as a CloudFront distribution, an Application Load Balancer or an API Gateway API, that inspects HTTP requests.
Lesson: Secure Workloads and Applications (task 1.2)

B

Backlog per instance
Queue messages waiting divided by the number of in-service instances. Used as a target tracking metric so queue workers scale in proportion to the work waiting.
Lesson: High-Performing and Elastic Compute (task 3.2)
Backup and restore
The cheapest disaster recovery strategy: nothing runs in the recovery Region, and you rebuild from backups and infrastructure code after a disaster, so RTO is the longest.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Backup copy job
An AWS Backup job that copies a recovery point to a vault in another Region or another account, so that a copy survives the loss of the source.
Lesson: Data Security Controls (task 1.3)
Block storage
Storage presented to one server as a raw disk that its operating system formats and mounts. Amazon EBS and instance store are the AWS examples.
Lesson: High-Performing and Scalable Storage (task 3.1)
Buffer interval
The maximum time Firehose holds data before delivery. For S3 it ranges from 0 to 900 seconds, with a default of 300. The buffer size or interval, whichever is reached first, triggers delivery.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Burstable instance
A T family EC2 instance that runs at a baseline CPU level and spends CPU credits to burst above it. In standard mode it drops to the baseline when credits run out; in unlimited mode it keeps bursting for an extra charge.
Lesson: High-Performing and Elastic Compute (task 3.2)

C

Cache key
The values a cache uses to tell stored objects apart, such as the URL path and selected headers. A new file name gives a new cache key.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Capacity Rebalancing
An Auto Scaling setting that launches a replacement when a Spot Instance gets a rebalance recommendation (elevated interruption risk), before the two-minute interruption notice arrives.
Lesson: Cost-Optimized Compute (task 4.2)
CloudFront invalidation
A request that removes cached copies from CloudFront edge locations before their TTL ends.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Cluster placement group
A placement group that packs instances close together in one Availability Zone for the lowest latency between nodes. A single flow between its instances can reach 10 Gbps instead of 5 Gbps.
Lesson: High-Performing and Elastic Compute (task 3.2)
Cold start
The extra latency of a request that has to wait while Lambda creates and initializes a new execution environment.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Columnar format
A file layout such as Parquet or ORC that stores each column together, so a query reads only the columns it names.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Compliance mode
An S3 Object Lock retention mode that no user, including the root user, can shorten or remove before the retention period ends.
Lesson: Data Security Controls (task 1.3)
Compute Savings Plans
The most flexible Savings Plan, up to 66 percent off On-Demand. It applies to EC2 usage in any instance family, size, Region, operating system or tenancy, and also to Fargate and Lambda.
Lesson: Cost-Optimized Compute (task 4.2)
Consolidated billing
The AWS Organizations feature that pays for all member accounts on one bill. Savings Plans and Reserved Instance discounts are shared across the accounts unless the management account turns sharing off.
Lesson: Cost-Optimized Compute (task 4.2)
Container
An application packaged with everything it needs to run, so it behaves the same on any host. Built from an image described by a Dockerfile.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Container orchestrator
A service that places containers on machines, keeps the requested number running and replaces them when they fail. Amazon ECS and Amazon EKS are the AWS examples.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Control (AWS Control Tower)
A governance rule in AWS Control Tower, also called a guardrail. Preventive controls use SCPs and RCPs, detective controls use AWS Config rules, and proactive controls use CloudFormation hooks.
Also called: GuardrailLesson: Secure Access to AWS Resources (task 1.1)
Control plane
The part of a service used to create or change resources, such as launching instances or editing DNS records. It is less available than the data plane, so keep it off the critical failover path.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Cost allocation tag
A resource tag that you activate in the Billing and Cost Management console so that costs can be grouped and filtered by it in AWS Cost Explorer and cost reports.
Lesson: Cost-Optimized Storage (task 4.1)
CPU credits
What a burstable (T family) instance earns while it runs below its baseline CPU and spends when it bursts above it.
Lesson: High-Performing and Elastic Compute (task 3.2)
Cross-Availability Zone data transfer
Traffic between resources in different Availability Zones of one Region. It is billed for both inbound and outbound data, so placing chatty tiers in one zone can remove the charge.
Lesson: Network Cost (task 4.4)
Cross-zone load balancing
Spreading traffic evenly across registered targets in every enabled Availability Zone. Always on at the load balancer level for an Application Load Balancer (it can be turned off per target group); off by default for a Network Load Balancer.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Custom endpoint
An Aurora endpoint that lists a chosen subset of the cluster's instances, so a workload can use only those instances.
Lesson: High-Performing Databases (task 3.3)
Customer gateway
The AWS resource that represents your on-premises VPN device in a Site-to-Site VPN connection.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Customer managed key
A KMS key that you create and control. You write its key policy, can enable automatic rotation, can disable it, and can schedule its deletion.
Lesson: Data Security Controls (task 1.3)

D

Data key
A key that encrypts the actual data. AWS KMS generates it and returns a plaintext copy for use plus a copy encrypted under the KMS key, which is stored beside the data.
Lesson: Data Security Controls (task 1.3)
Data lake
Raw and processed data kept in Amazon S3 in open file formats so many analytics services can query it. AWS Lake Formation manages access to it.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Data plane
The part of a service that carries the workload's traffic. It has higher availability goals than the control plane, so failover should use data plane operations.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Data Processing Unit (DPU)
The AWS Glue unit of compute: one DPU is 4 vCPUs and 16 GB of memory, and a G.1X worker provides one DPU.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Data repository association
A link between an Amazon FSx for Lustre file system and an S3 bucket. S3 objects appear as files, and changes can be exported back to the bucket.
Lesson: High-Performing and Scalable Storage (task 3.1)
Data transfer out
Data that leaves AWS to the internet or to an on-premises network. Outbound data is billed per GB at a rate set by the path it takes and the Region it leaves from, so compare paths on the pricing pages. Data transfer in over AWS Direct Connect is $0.00 per GB.
Lesson: Network Cost (task 4.4)
Database Savings Plans
A one-year Savings Plan for database spend on Aurora, RDS, DynamoDB, ElastiCache and other database services. You commit to a fixed hourly spend, the discount is up to 35 percent, and it reserves no capacity.
Lesson: Cost-Optimized Databases (task 4.3)
DDoS
Distributed denial of service: a flood of traffic from many sources meant to overwhelm an application. AWS Shield is the AWS answer.
Lesson: Secure Workloads and Applications (task 1.2)
Dead-letter queue
A queue that receives messages that failed processing a set number of times (maxReceiveCount), so they stop blocking the source queue and can be inspected or redriven later.
Also called: DLQLesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Dedicated connection (Direct Connect)
A physical Direct Connect port that you request directly, at 1, 10, 100 or 400 Gbps.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Dedicated Host
A physical EC2 server dedicated to your use. It lets you bring software licenses that are counted per socket, per core or per VM, and helps with some compliance requirements.
Lesson: Cost-Optimized Compute (task 4.2)
Delete marker
The placeholder S3 adds when you delete an object in a versioned bucket. The older versions stay and can be restored.
Lesson: Data Security Controls (task 1.3)
Directory bucket
The S3 bucket type used by S3 Express One Zone. It keeps data in a single Availability Zone for the lowest request latency.
Lesson: High-Performing and Scalable Storage (task 3.1)
Disaster recovery (DR) strategy
A planned approach for restoring a workload after a disaster, chosen by its RPO and RTO: backup and restore, pilot light, warm standby or multi-site active/active.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
DynamoDB Accelerator (DAX)
An in-memory cache for DynamoDB that serves eventually consistent reads in microseconds. It passes strongly consistent and transactional reads through to the table without caching them.
Lesson: High-Performing Databases (task 3.3)
DynamoDB on-demand capacity mode
A DynamoDB billing mode that charges per read and write request and needs no capacity settings.
Lesson: Cost-Optimized Databases (task 4.3)
DynamoDB provisioned capacity mode
A DynamoDB billing mode that charges for the read and write capacity units you set per second, usually with auto scaling.
Lesson: Cost-Optimized Databases (task 4.3)
DynamoDB reserved capacity
A one- or three-year commitment to read and write capacity for provisioned DynamoDB tables, bought in blocks of 100 capacity units. It lowers the rate on the baseline you always use, applies only to provisioned tables on the Standard table class, and is not supported on Standard-IA.
Lesson: Cost-Optimized Databases (task 4.3)
DynamoDB Standard-IA table class
A DynamoDB table class with lower storage pricing and higher read and write pricing, suited to tables where storage is about 50 percent or more of table cost and access is infrequent.
Lesson: Cost-Optimized Databases (task 4.3)
DynamoDB time to live (TTL)
A DynamoDB feature that deletes items after the epoch-seconds time stored in a number attribute. Deletion runs in the background and can lag behind the expiry time.
Lesson: Cost-Optimized Databases (task 4.3)

E

EBS Multi-Attach
An Amazon EBS feature that attaches one io1 or io2 volume to up to 16 Nitro instances in the same Availability Zone. The applications must coordinate writes themselves.
Lesson: High-Performing and Scalable Storage (task 3.1)
EBS Snapshots Archive
A lower-cost tier for EBS snapshots that you keep for months or years. Archived snapshots are stored as full snapshots, bill for at least 90 days and take up to 72 hours to restore before use.
Lesson: Cost-Optimized Storage (task 4.1)
EC2 hibernation
Stopping an instance after saving its RAM to the encrypted EBS root volume, so its processes resume when it starts again. It must be enabled at launch and can't be added to an existing instance.
Also called: HibernationLesson: Cost-Optimized Compute (task 4.2)
EC2 Instance Savings Plans
A Savings Plan tied to one instance family in one Region, up to 72 percent off On-Demand. Size, operating system and tenancy can change within that family.
Lesson: Cost-Optimized Compute (task 4.2)
Edge location
A CloudFront site close to users that serves cached copies of content, so requests don't travel to the origin each time.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
EFS Elastic throughput
The Amazon EFS throughput mode that AWS recommends for spiky or unpredictable workloads. It scales throughput up and down automatically with the workload and bills for the metadata and data read and written.
Lesson: High-Performing and Scalable Storage (task 3.1)
EFS lifecycle management
An Amazon EFS feature that moves files that have not been read for a set time from the Standard storage class to Infrequent Access or Archive, and optionally back on first access.
Lesson: Cost-Optimized Storage (task 4.1)
Elastic Fabric Adapter
A network device for EC2 instances that adds OS-bypass (through Libfabric) to ENA, giving tightly coupled HPC and ML applications lower, more consistent latency and higher throughput for traffic between instances, such as MPI and NCCL.
Also called: EFALesson: High-Performing and Elastic Compute (task 3.2)
Elastic Network Adapter (ENA)
The EC2 network interface that enhanced networking uses on Nitro-based instances, with support for up to 100 Gbps on supported instance types.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Elastic Volumes
An Amazon EBS capability that changes a volume's size, type, IOPS or throughput while it stays attached and in use.
Lesson: High-Performing and Scalable Storage (task 3.1)
ElastiCache reserved node
A pre-purchased ElastiCache node billed at a discounted rate for a one- or three-year term, with No, Partial or All Upfront payment. Its node type and Region are fixed at purchase.
Lesson: Cost-Optimized Databases (task 4.3)
ENA Express
An EC2 network feature that carries TCP traffic on the SRD protocol and raises a single flow to 25 Gbps between supported instances in the same Region, including across Availability Zones.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Endpoint policy
A resource-based policy attached to a VPC endpoint that limits which principals, actions and resources can use the endpoint.
Lesson: Secure Workloads and Applications (task 1.2)
Enhanced fan-out
A Kinesis Data Streams feature that gives each registered consumer its own 2 MB per second per shard, so consumers don't share read throughput.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Envelope encryption
Encrypting data with a data key, then encrypting that data key with a KMS key. Only the small data key goes to KMS, so large data sets can be protected by one key.
Lesson: Data Security Controls (task 1.3)
Ephemeral port
A short-lived, high-numbered port a client picks for the return half of a connection. Stateless network ACLs must allow the range for replies to get back.
Lesson: Secure Workloads and Applications (task 1.2)
Equal-Cost Multi-Path (ECMP)
Routing that spreads traffic across several equal-cost paths, such as several VPN tunnels or connections. It requires dynamic routing.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Event bus
An EventBridge resource that receives events and routes each one to targets whose rules match the event's content.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Eventually consistent read
A read that might not reflect a very recent write but costs half as much. It is the DynamoDB default and the only option on a global secondary index.
Lesson: High-Performing Databases (task 3.3)
Explicit deny
A Deny statement that matches a request. It overrides every Allow in every policy type, which is why adding more allows never fixes it.
Lesson: Secure Access to AWS Resources (task 1.1)
Express workflow
A Step Functions workflow type for high-volume, short work: it runs for up to five minutes. Asynchronous Express runs at-least-once and synchronous Express runs at-most-once. The Standard type runs up to a year with exactly-once execution.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
External ID
A value that a third party must pass when it assumes a role in your account, checked by the role's trust policy. It prevents the confused deputy problem and isn't treated as a secret.
Lesson: Secure Access to AWS Resources (task 1.1)
Extract, transform, load (ETL)
A job that reads data from a source, cleans or reshapes it, and writes it to a target. AWS Glue runs ETL jobs without servers to manage.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)

F

Failover routing policy
A Route 53 routing policy with a primary and a secondary record. Route 53 answers with the secondary record when the primary's health check fails.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Fan-out
Publishing one message to an SNS topic so that every subscriber, usually one SQS queue per consuming system, receives its own copy.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Fargate Spot
A Fargate capacity provider that runs interruption-tolerant Amazon ECS tasks on spare capacity at a discount, with a two-minute warning before a task is stopped.
Lesson: Cost-Optimized Compute (task 4.2)
File storage
Storage shared as a directory tree that many servers mount at once over a network protocol such as NFS or SMB. Amazon EFS and Amazon FSx are the AWS examples.
Lesson: High-Performing and Scalable Storage (task 3.1)
Free backup storage allowance
The backup storage that RDS and Aurora include at no charge: equal to provisioned database storage for RDS, and the latest cluster volume size for Aurora automated backups. Only backup storage above the allowance is billed.
Lesson: Cost-Optimized Databases (task 4.3)
FSx for Lustre
An Amazon FSx file system built on Lustre for HPC, ML and media workloads that need sub-millisecond latency and very high throughput. It can link to an S3 bucket and present its objects as files.
Lesson: High-Performing and Scalable Storage (task 3.1)
FSx for ONTAP capacity pool
The elastic, lower-cost storage tier of Amazon FSx for NetApp ONTAP. Data that is rarely read moves there from the SSD tier automatically.
Lesson: Cost-Optimized Storage (task 4.1)

G

Gateway VPC endpoint
A route-table entry that sends Amazon S3 or DynamoDB traffic from a VPC straight to the service, without an internet gateway or NAT device.
Lesson: Secure Workloads and Applications (task 1.2)
Global secondary index
A DynamoDB index with its own partition and sort keys, used to query by an attribute other than the table's key. It supports only eventually consistent reads.
Lesson: High-Performing Databases (task 3.3)
Golden AMI
A tested machine image with the required software and configuration, copied to the recovery Region so instances can launch quickly during recovery.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Governance mode
An S3 Object Lock retention mode that blocks normal deletes, but lets users with the s3:BypassGovernanceRetention permission remove the lock.
Lesson: Data Security Controls (task 1.3)
gp3
The current General Purpose SSD Amazon EBS volume type. It includes 3,000 IOPS and 125 MiB/s, and you can add IOPS and throughput without enlarging the volume.
Also called: General Purpose SSDLesson: High-Performing and Scalable Storage (task 3.1)

H

Helm chart
A packaged set of Kubernetes configuration files that installs an application in one step.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Heterogeneous migration
Moving a database to a different engine. AWS Schema Conversion Tool (SCT) converts the schema and code, and AWS DMS moves the data.
Lesson: High-Performing Databases (task 3.3)
Horizontal scaling
Adding or removing copies of a component and spreading load across them.
Also called: Scaling outLesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Hosted connection (Direct Connect)
A Direct Connect connection that an AWS Direct Connect Partner provisions and shares with you, at speeds from 50 Mbps to 25 Gbps.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Hot partition
A DynamoDB partition that receives more traffic than its per-partition limit, so requests to its keys are throttled even when the table has spare capacity.
Lesson: High-Performing Databases (task 3.3)

I

IAM role
An IAM identity with permissions but no long-lived password or access keys. A person, application or AWS service assumes it and receives temporary credentials from AWS STS.
Lesson: Secure Access to AWS Resources (task 1.1)
Idempotent consumer
A consumer that produces the same result when it processes the same message more than once, which makes at-least-once delivery safe.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Identity federation
Letting users from an external identity provider, through SAML 2.0 or OIDC, receive temporary AWS credentials for a role instead of having their own IAM users.
Also called: FederationLesson: Secure Access to AWS Resources (task 1.1)
Identity-based policy
A permissions policy attached to an IAM user, group or role that says what that identity may do on which resources.
Lesson: Secure Access to AWS Resources (task 1.1)
Immutable infrastructure
An approach in which servers are replaced from a tested image or template rather than patched or changed in place.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Instance family
The group of EC2 instance types built around one resource, named by the first letter of the type: C for compute, R and X for memory, P, G, Inf and Trn for accelerators, I and D for storage, M for general purpose, T for burstable.
Lesson: High-Performing and Elastic Compute (task 3.2)
Instance profile
The container that attaches an IAM role to an EC2 instance, so applications on the instance receive the role's temporary credentials without stored access keys.
Lesson: Secure Access to AWS Resources (task 1.1)
Instance store
Temporary block storage on disks physically attached to an EC2 instance's host. Very fast, but the data is lost when the instance stops, hibernates or terminates.
Also called: Ephemeral storageLesson: High-Performing and Scalable Storage (task 3.1)
Instance warmup
The time after a new instance enters service during which its metrics are left out of the group's aggregated metrics, so booting instances don't trigger extra scaling.
Lesson: High-Performing and Elastic Compute (task 3.2)
Inter-Region data transfer
Traffic between AWS Regions. It is billed on the outbound side from the source Region at the inter-Region rate; the destination Region's inbound data has no charge.
Lesson: Network Cost (task 4.4)
Interface VPC endpoint
A network interface with a private IP address in your subnet that connects to an AWS service through AWS PrivateLink, controlled by a security group.
Lesson: Secure Workloads and Applications (task 1.2)
Internet gateway
The VPC component that connects a VPC to the internet. A subnet becomes public when its route table points to one.
Lesson: Secure Workloads and Applications (task 1.2)
io2 Block Express
The Provisioned IOPS SSD Amazon EBS volume type for the most demanding workloads, with sub-millisecond latency and the highest IOPS and throughput per volume.
Also called: Provisioned IOPS SSDLesson: High-Performing and Scalable Storage (task 3.1)
IOPS
Input/output operations per second: how many separate reads and writes a storage device completes each second. The key number for small, random I/O such as databases.
Lesson: High-Performing and Scalable Storage (task 3.1)
iSCSI
A protocol that sends disk (block) commands over a network, so a remote volume looks like a local disk. Volume Gateway and Tape Gateway present iSCSI devices.
Lesson: High-Performing and Scalable Storage (task 3.1)

J

Job bookmark
An AWS Glue feature that records which data a job already processed, so a rerun reads only new data.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Jumbo frame
A network packet larger than the standard 1500-byte MTU. Direct Connect private virtual interfaces carry up to 9001 bytes, and Site-to-Site VPN supports no jumbo frames (its tunnel MTU is 1446 bytes).
Lesson: High-Performing and Scalable Network Architectures (task 3.4)

K

Key material
The secret bits inside a key that do the encrypting. Rotation replaces the key material for new data and keeps the old material to decrypt existing data.
Lesson: Data Security Controls (task 1.3)
KMS grant
A delegation that gives a principal a set of KMS key operations without editing the key policy. A grant can be retired when the permission is no longer needed.
Lesson: Data Security Controls (task 1.3)
KMS key policy
The resource policy attached to a KMS key. It must allow access, and IAM policies can grant access only where the key policy allows it.
Lesson: Data Security Controls (task 1.3)

L

Lambda concurrency
The number of requests a Lambda function is handling at the same moment. The default quota is 1,000 concurrent executions per account per Region, shared by all functions.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Lambda execution environment
The isolated sandbox that runs a Lambda function's code. Lambda may reuse it for later requests or discard it, so nothing stored in it is guaranteed to survive.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Landing zone
The multi-account environment that AWS Control Tower sets up: an organization with OUs, shared accounts and controls based on best practices.
Lesson: Secure Access to AWS Resources (task 1.1)
Large Bandwidth Tunnel
A Site-to-Site VPN tunnel option that carries up to 5 Gbps per tunnel, available only on transit gateway or Cloud WAN attachments, and not in five Regions (ap-southeast-4, ca-west-1, eu-central-2, il-central-1, me-central-1).
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Lazy loading
A caching strategy that reads from the cache first and, on a miss, reads the database and writes the result into the cache.
Also called: Cache-asideLesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Link aggregation group (LAG)
A logical bundle of Direct Connect connections of the same speed at one location, which acts as one connection.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Listener rule
An Application Load Balancer instruction that sends requests to a target group based on conditions such as path, host name or header.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Load balancer
A single endpoint that health-checks a group of targets and sends traffic only to the healthy ones. Elastic Load Balancing offers Application, Network and Gateway Load Balancers.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Local secondary index
A DynamoDB index that keeps the table's partition key and uses a different sort key. It must be defined when the table is created and supports strongly consistent reads.
Lesson: High-Performing Databases (task 3.3)
Loose coupling
A design where components talk through a durable intermediary such as a queue, topic or event bus, so each one can fail, scale and change without breaking the others.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)

M

MACsec
An IEEE standard that encrypts traffic on a Layer 2 link. Direct Connect supports it on some dedicated connections.
Lesson: Secure Workloads and Applications (task 1.2)
Managed rule group
A pre-built set of AWS WAF rules maintained by AWS or a vendor, that you add to a web ACL without writing the rules yourself.
Lesson: Secure Workloads and Applications (task 1.2)
Management account
The account that creates and pays for an AWS Organizations organization. SCPs and RCPs don't restrict it, so workloads belong in member accounts.
Lesson: Secure Access to AWS Resources (task 1.1)
Maximum transmission unit (MTU)
The size in bytes of the largest packet a network path can carry. VPC and transit gateway paths support 8500 or 9001 bytes. A transit gateway accepts VPN traffic up to 1500 bytes, but the usable Site-to-Site VPN tunnel MTU is 1446 bytes.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Message broker
A server that relays messages between applications over standard protocols such as AMQP, MQTT or OpenWire. Amazon MQ runs managed ActiveMQ and RabbitMQ brokers.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Message group ID
The SQS FIFO queue attribute that sets the scope of ordering. Messages in one group are processed in order; different groups are processed in parallel.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Minimum billable object size
The smallest size an S3 storage class charges for each object. S3 Standard-IA, S3 One Zone-IA and S3 Glacier Instant Retrieval bill any smaller object as 128 KB.
Lesson: Cost-Optimized Storage (task 4.1)
Minimum storage duration
The shortest period an S3 storage class bills for an object: 30 days for S3 Standard-IA and S3 One Zone-IA, 90 days for S3 Glacier Instant Retrieval and Flexible Retrieval, 180 days for S3 Glacier Deep Archive. Removing or moving the object earlier still bills the remainder.
Lesson: Cost-Optimized Storage (task 4.1)
Mixed instances group
An Auto Scaling group that launches several instance types and can combine On-Demand and Spot Instances, for example an On-Demand base capacity with Spot above it.
Lesson: Cost-Optimized Compute (task 4.2)
Mount target
The NFS endpoint, with its own IP address, that you create for an Amazon EFS file system in an Availability Zone. Clients in that zone mount the file system through it.
Lesson: High-Performing and Scalable Storage (task 3.1)
Multi-AZ DB instance deployment
An RDS deployment with a synchronous standby replica in another Availability Zone that takes over automatically on failure. The standby does not serve reads.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Multi-Region key
A KMS key that has the same key ID and key material in more than one Region, so data encrypted in one Region can be decrypted in another.
Lesson: Data Security Controls (task 1.3)
Multi-site active/active
A strategy that serves production traffic from two or more Regions at the same time, with the lowest RTO and the highest design complexity.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Multipart upload
An Amazon S3 feature that uploads one object as many parts in parallel and retries only the parts that fail. Recommended for objects over 100 MB. Each part must be 5 MiB to 5 GiB, except the last part.
Lesson: High-Performing and Scalable Storage (task 3.1)

N

NAT gateway
A managed network address translation service that lets instances in private subnets start outbound connections and receive replies, without accepting connections from outside.
Lesson: Secure Workloads and Applications (task 1.2)
NAT gateway data processing charge
The per-GB charge a NAT gateway applies to every byte it processes, in either direction, on top of its hourly charge. A gateway endpoint for S3 or DynamoDB avoids it.
Lesson: Network Cost (task 4.4)
NAT instance
An EC2 instance you run yourself as the NAT device. It has no NAT processing fee, but you handle patching, scaling and failover.
Lesson: Network Cost (task 4.4)
Network ACL
A stateless, subnet-level firewall with numbered allow and deny rules. Rules are checked in ascending number order, and the first match wins.
Lesson: Secure Workloads and Applications (task 1.2)
Network File System (NFS)
The file-sharing protocol Linux clients use to mount a shared file system over the network. Amazon EFS and S3 File Gateway speak NFS.
Lesson: High-Performing and Scalable Storage (task 3.1)
Noncurrent version
An older version of an object in a versioned S3 bucket, kept after the object is overwritten or deleted. It is billed until a lifecycle rule or a person removes it.
Lesson: Cost-Optimized Storage (task 4.1)

O

Object Ownership (bucket owner enforced)
An S3 setting that disables object ACLs, so the bucket owner owns every object and access is controlled by policies.
Lesson: Data Security Controls (task 1.3)
Object storage
Storage that keeps each file as an object with metadata in a flat namespace, read and written over an HTTP API rather than mounted as a disk. Amazon S3 is the AWS example.
Lesson: High-Performing and Scalable Storage (task 3.1)
On-demand capacity mode
A Kinesis Data Streams mode with no shard counts to set. The stream scales to up to double the peak write rate of the previous 30 days.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
On-Demand Capacity Reservation
Reserved EC2 capacity for a chosen instance type in one Availability Zone, for any duration. One that starts immediately has no commitment and can be cancelled at any time; a future-dated reservation has a commitment duration and can carry a cancellation charge. It is billed at the On-Demand rate whether or not instances use it, and Savings Plans and regional RIs can discount it.
Also called: Capacity ReservationLesson: Cost-Optimized Compute (task 4.2)
On-Demand Instance
An EC2 instance you pay for by the second (or hour) with no commitment and no discount. It is the default and the fallback for usage that is short, spiky or unpredictable.
Lesson: Cost-Optimized Compute (task 4.2)
Organizational unit (OU)
A container for accounts inside AWS Organizations. Policies attached to an OU apply to every account in it, including accounts added later.
Also called: OULesson: Secure Access to AWS Resources (task 1.1)
Origin (CloudFront)
The source CloudFront fetches content from on a cache miss, such as an S3 bucket, a load balancer or a custom HTTP server.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Origin access control
A CloudFront feature that signs requests to an S3 origin so the bucket can stay private and accept reads only from that distribution.
Also called: OACLesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Origin group
A CloudFront primary and secondary origin. For GET, HEAD and OPTIONS requests, CloudFront retries on the secondary when the primary returns a status code you chose, such as 500, 502, 503 or 504.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)

P

Parameter Store
The Systems Manager capability for configuration data, with standard and advanced tiers and SecureString values. It has no built-in rotation.
Lesson: Secure Workloads and Applications (task 1.2)
Partition key
The DynamoDB attribute whose value decides the partition that stores an item. Many distinct values spread traffic across partitions.
Lesson: High-Performing Databases (task 3.3)
Partition placement group
A placement group that splits instances into partitions on separate racks, up to 7 partitions per Availability Zone. Suits large distributed stores such as HDFS, Cassandra or Kafka.
Lesson: High-Performing and Elastic Compute (task 3.2)
Partition projection
An Athena feature that computes partition values from table properties instead of looking them up in the Glue Data Catalog for each query.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Partition pruning
The engine skipping partitions that a query's filters rule out, so it reads only the matching folders.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Permission set
An IAM Identity Center template of policies. When you assign it to a user or group for an account, IAM Identity Center creates and maintains a matching IAM role in that account.
Lesson: Secure Access to AWS Resources (task 1.1)
Permissions boundary
A managed policy set on an IAM user or role as the maximum its identity-based policies can grant. It grants nothing by itself.
Lesson: Secure Access to AWS Resources (task 1.1)
Pilot light
A DR strategy that keeps data replicated and core infrastructure running in the recovery Region while application servers stay switched off until failover.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Placement group
An EC2 setting that controls how instances are placed on hardware: cluster for low latency in one AZ, partition for rack isolation between groups of instances, spread for distinct hardware per instance.
Lesson: High-Performing and Elastic Compute (task 3.2)
Pod
The smallest unit Kubernetes runs: one or more containers scheduled together. On Amazon EKS, pods run on EC2 nodes or on Fargate.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Poison message
A message that fails every time a consumer processes it. A dead-letter queue moves it aside so it stops blocking or wasting retries.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Pre-launch time
How far ahead of a forecast hour predictive scaling launches instances, so they are ready when the load arrives.
Lesson: High-Performing and Elastic Compute (task 3.2)
Predictive scaling
An Auto Scaling policy that forecasts capacity from up to 14 days of metric history and launches instances ahead of daily or weekly peaks. It scales out only.
Lesson: High-Performing and Elastic Compute (task 3.2)
Prefix list
A named set of CIDR blocks that security group and route table rules can reference instead of listing each address range.
Lesson: Secure Workloads and Applications (task 1.2)
price-capacity-optimized
The recommended Spot allocation strategy for Auto Scaling groups. It picks the Spot pools that are least likely to be interrupted and have the lowest possible price.
Lesson: Cost-Optimized Compute (task 4.2)
Private subnet
A subnet with no route to an internet gateway. Its instances can't be reached from the internet and reach out only through a NAT gateway or VPC endpoints.
Lesson: Secure Workloads and Applications (task 1.2)
Provisioned concurrency
A Lambda setting that keeps a number of execution environments initialized ahead of time so requests don't wait for a cold start.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Public IPv4 address charge
The hourly charge AWS applies to every public IPv4 address, whether it is attached to a running resource or idle, including Elastic IP addresses.
Lesson: Network Cost (task 4.4)
Public subnet
A subnet whose route table sends internet-bound traffic to an internet gateway. Instances in it can be reached from the internet only if they also have a public IP address.
Lesson: Secure Workloads and Applications (task 1.2)

Q

Queue depth
The number of messages waiting in a queue. A common metric for scaling the workers that consume it.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)

R

Rate-based rule
An AWS WAF rule that counts requests per aggregation key, such as IP address, over an evaluation window, and applies its action when a limit is exceeded.
Lesson: Secure Workloads and Applications (task 1.2)
RDS Proxy
A managed connection pool in front of RDS or Aurora that keeps client connections open and routes them to a new primary during failover.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Read capacity unit (RCU)
DynamoDB's unit of read throughput: one strongly consistent read per second, or two eventually consistent reads, of an item up to 4 KB.
Lesson: High-Performing Databases (task 3.3)
Read replica
A read-only copy of a database that receives changes from its source through replication. It takes read traffic; writes still go to the source.
Lesson: High-Performing Databases (task 3.3)
Reader endpoint
An Aurora cluster endpoint that spreads new read connections across the Aurora Replicas. It balances connections, not individual queries.
Lesson: High-Performing Databases (task 3.3)
Record format conversion
A Firehose feature that converts JSON records to Parquet or ORC for S3 delivery, using a schema from a Glue Data Catalog table.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Recovery point objective (RPO)
The most data loss the business accepts, measured as the time between the last recoverable copy and the disaster.
Also called: RPOLesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Recovery time objective (RTO)
The longest a service can stay down after a disaster before it must be working again.
Also called: RTOLesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Redshift Serverless RPU
The unit of compute capacity that Amazon Redshift Serverless bills per RPU-hour while queries run. A usage limit caps how many RPUs can be used, so a burst of queries cannot run up the bill.
Lesson: Cost-Optimized Databases (task 4.3)
Regional NAT gateway
A NAT gateway that is not tied to one Availability Zone. In automatic mode it expands into each AZ that has workloads. It still charges per hour and per GB processed.
Lesson: Network Cost (task 4.4)
Regional Reserved Instance
A Reserved Instance bought for a Region rather than one Availability Zone. Its discount applies in any AZ in that Region and, for Linux/Unix, to any size within the instance family, but it does not reserve capacity.
Lesson: Cost-Optimized Compute (task 4.2)
Registered S3 location
An S3 path registered with Lake Formation so that its data and the folders beneath it are governed by Lake Formation permissions.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Replica lag
The delay between a write on a source database and its appearance on a read replica. For RDS read replicas it is reported by the ReplicaLag CloudWatch metric.
Lesson: High-Performing Databases (task 3.3)
Requester Pays
An S3 bucket setting where the person downloading data pays the request and download charges while the bucket owner pays for storage and for AccessDenied requests made from inside the owner's account or organization.
Lesson: Cost-Optimized Storage (task 4.1)
Reserved concurrency
A Lambda setting that sets aside part of the account's concurrency for one function and also caps that function at the same number.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Reserved DB instance
An RDS or Aurora billing commitment for a one- or three-year term that discounts one instance type, engine and Region. It does not reserve compute capacity.
Lesson: Cost-Optimized Databases (task 4.3)
Reserved Instance (RI)
A 1- or 3-year billing discount for a specific EC2 instance configuration. Standard RIs give the larger discount and can be sold on the Reserved Instance Marketplace; Convertible RIs can be exchanged for other attributes but not sold.
Also called: Reserved Instance, Standard Reserved Instance, Convertible Reserved InstanceLesson: Cost-Optimized Compute (task 4.2)
Resource control policy (RCP)
An AWS Organizations policy that caps what any principal can do to resources of supported services in member accounts. Like an SCP, it never grants permissions.
Also called: RCPLesson: Secure Access to AWS Resources (task 1.1)
Resource-based policy
A policy attached to a resource, such as an S3 bucket policy or an SQS queue policy, that names which principals, including ones in other accounts, may act on it.
Lesson: Secure Access to AWS Resources (task 1.1)
Rightsizing
Matching a resource's type and size to the capacity it actually uses, for example moving an instance that idles at low CPU to a smaller size or a different family. Do it before buying any commitment.
Lesson: Cost-Optimized Compute (task 4.2)
Role chaining
Using the temporary credentials of one role to assume another role. Sessions created this way last at most one hour.
Lesson: Secure Access to AWS Resources (task 1.1)
Root user
The sign-in identity created with an AWS account, using the account email address and password. It has complete access to the account, so you protect it with MFA, create no access keys for it, and use it only for the few tasks that require it.
Also called: AWS account root userLesson: Secure Access to AWS Resources (task 1.1)
Route 53 health check
A check that probes an endpoint and marks it healthy or unhealthy, so DNS routing can skip unhealthy targets.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)

S

S3 Block Public Access
A set of four account-level and bucket-level settings that refuse public ACLs and public bucket policies, whatever the other access settings say.
Lesson: Data Security Controls (task 1.3)
S3 Bucket Key
A bucket-level key that S3 derives from a KMS key, so S3 calls AWS KMS far less often for SSE-KMS objects and KMS request costs drop.
Lesson: Data Security Controls (task 1.3)
S3 Express One Zone
A high-performance Amazon S3 storage class in directory buckets that keeps data in one Availability Zone you choose and gives consistent single-digit millisecond access.
Lesson: High-Performing and Scalable Storage (task 3.1)
S3 File Gateway
A Storage Gateway type that serves NFS or SMB shares on premises and stores each file as an object in Amazon S3, keeping recently used data in a local cache.
Lesson: High-Performing and Scalable Storage (task 3.1)
S3 Glacier Deep Archive
The lowest-cost S3 storage class, for data kept for years and rarely read. Objects must be restored first, which takes up to 12 hours (Standard) or 48 hours (Bulk).
Lesson: Cost-Optimized Storage (task 4.1)
S3 Glacier Flexible Retrieval
An S3 archive storage class whose objects must be restored before they can be read, in minutes (Expedited) to hours (Standard or Bulk). Its API name is GLACIER.
Lesson: Cost-Optimized Storage (task 4.1)
S3 Glacier Instant Retrieval
An S3 archive storage class for data read about once a quarter that must still come back in milliseconds, with a 90-day minimum storage duration.
Lesson: Cost-Optimized Storage (task 4.1)
S3 Intelligent-Tiering
An S3 storage class that moves each object between access tiers based on how recently it was read. It charges a small monitoring fee per object and no retrieval fees, which suits unknown or changing access patterns.
Lesson: Cost-Optimized Storage (task 4.1)
S3 legal hold
An S3 Object Lock setting that protects an object version with no end date. It stays until someone removes it, and it works apart from the retention period.
Lesson: Data Security Controls (task 1.3)
S3 Lifecycle configuration
A set of rules on an S3 bucket that moves objects to cheaper storage classes after a number of days and deletes objects, old versions or incomplete uploads when they are no longer needed.
Also called: Lifecycle ruleLesson: Cost-Optimized Storage (task 4.1)
S3 Object Lock
An S3 setting that stops an object version from being overwritten or deleted for a retention period or while a legal hold is set. It requires versioning.
Lesson: Data Security Controls (task 1.3)
S3 One Zone-IA
An S3 storage class like S3 Standard-IA but stored in a single Availability Zone, so it costs less and is suitable only for data you can re-create.
Also called: S3 One Zone-Infrequent AccessLesson: Cost-Optimized Storage (task 4.1)
S3 prefix
The part of an object key before the object name, for example uploads/2026/ in uploads/2026/photo.jpg. Amazon S3 scales request rate per partitioned prefix.
Lesson: High-Performing and Scalable Storage (task 3.1)
S3 Standard-IA
An S3 storage class for infrequently accessed data that still needs millisecond access. It stores data redundantly across multiple Availability Zones and adds a per-GB retrieval fee.
Also called: S3 Standard-Infrequent AccessLesson: Cost-Optimized Storage (task 4.1)
S3 Storage Class Analysis
An S3 feature that watches how often objects under a prefix or tag are read, and suggests when to move them from S3 Standard to S3 Standard-IA.
Lesson: Cost-Optimized Storage (task 4.1)
S3 Transfer Acceleration
An Amazon S3 bucket feature that routes transfers through the nearest CloudFront edge location and the AWS network to speed up long-distance uploads and downloads.
Lesson: High-Performing and Scalable Storage (task 3.1)
Savings Plans
A billing discount in exchange for committing to a fixed amount of compute spend in USD per hour for 1 or 3 years. A Savings Plan with an hourly commitment of $100 or less can be returned within 7 days of purchase, in the same calendar month. Outside that window the commitment is binding for its full term, and a Savings Plan never reserves capacity.
Lesson: Cost-Optimized Compute (task 4.2)
Savings Plans utilization and coverage
Utilization is how much of your purchased commitment you actually used; coverage is how much of your eligible usage the commitment discounted. Cost Explorer reports both, and an AWS Budgets Savings Plans budget can alert when either drops below a target.
Also called: Savings Plans coverage, Savings Plans utilizationLesson: Cost-Optimized Compute (task 4.2)
Scheduled scaling
Auto Scaling actions that change a group's desired, minimum or maximum capacity at a set date and time or on a recurring schedule.
Lesson: High-Performing and Elastic Compute (task 3.2)
Scratch file system
An Amazon FSx for Lustre deployment type for temporary processing. It costs less because data is not replicated, so it is not kept if a server fails.
Lesson: Cost-Optimized Storage (task 4.1)
Security group
A stateful, allow-only firewall attached to a network interface. Its rules can name another security group as the source, so group membership replaces IP lists.
Lesson: Secure Workloads and Applications (task 1.2)
Server Message Block (SMB)
The file-sharing protocol Windows clients use. Amazon FSx for Windows File Server and S3 File Gateway support SMB.
Lesson: High-Performing and Scalable Storage (task 3.1)
Service control policy (SCP)
An AWS Organizations policy that caps what users and roles in member accounts can do, including the member account root user. It never grants permissions and doesn't apply to the management account or to service-linked roles.
Also called: SCPLesson: Secure Access to AWS Resources (task 1.1)
Session Manager
A Systems Manager capability that opens shell sessions on instances through IAM access control, with no inbound SSH or RDP port.
Lesson: Secure Workloads and Applications (task 1.2)
Shard
The unit of capacity in a provisioned Kinesis data stream. Each shard accepts 1 MB per second or 1,000 records per second of writes and serves 2 MB per second of reads, shared by all standard consumers.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Shared responsibility model
The split of security duties: AWS secures the cloud itself (facilities, hardware, the services' infrastructure), and you secure what you put in it, including identities, permissions and data.
Lesson: Secure Access to AWS Resources (task 1.1)
Single point of failure
A component whose failure stops the whole workload. It is removed with redundancy across zones or Regions, or with a managed service that is redundant by design.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Single-flow bandwidth
The throughput of one 5-tuple TCP or UDP flow. Outside a cluster placement group it is capped at 5 Gbps.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Site-to-Site VPN
An IPsec VPN between your network and a VPC or transit gateway, with two tunnels per connection that run over the internet.
Lesson: Secure Workloads and Applications (task 1.2)
Sort key
The optional second part of a DynamoDB primary key. Items that share a partition key are stored in sort key order, which allows range queries.
Lesson: High-Performing Databases (task 3.3)
Sparse index
A DynamoDB index that contains only the items carrying its key attributes. Items without them are not written to the index.
Lesson: High-Performing Databases (task 3.3)
Spot capacity pool
The set of unused EC2 instances of one instance type in one Availability Zone. Spreading a fleet across many pools makes Spot interruptions less disruptive.
Lesson: Cost-Optimized Compute (task 4.2)
Spot Instance
An EC2 instance that runs on spare capacity for up to 90 percent off On-Demand. EC2 can reclaim it with a two-minute interruption notice, so it suits work that can be interrupted and restarted.
Also called: SpotLesson: Cost-Optimized Compute (task 4.2)
Spread placement group
A placement group that puts each instance on distinct hardware, with at most 7 running instances per Availability Zone. Suits a few critical instances that must not fail together.
Lesson: High-Performing and Elastic Compute (task 3.2)
SQL injection
A web attack that hides database commands inside form fields or URLs. AWS WAF rules can block it; security groups and network ACLs can't see it.
Lesson: Secure Workloads and Applications (task 1.2)
SQS FIFO queue
An SQS queue type that keeps strict order within each message group and removes duplicates sent within a 5-minute interval, at lower throughput than a standard queue.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Standby Region service quota
A per-Region service limit that must be raised in the recovery Region before a standby can scale to production capacity.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
State machine
A defined set of steps and the transitions between them. It is the workflow model AWS Step Functions runs.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Stateful filtering
A firewall that tracks connections, so replies to allowed traffic are allowed automatically. Security groups filter this way.
Lesson: Secure Workloads and Applications (task 1.2)
Stateless filtering
A firewall that checks each packet on its own, so return traffic needs its own rule. Network ACLs filter this way.
Lesson: Secure Workloads and Applications (task 1.2)
Stateless workload
A component that keeps nothing between requests in its own memory or disk, so any copy can serve any request and copies can be added or removed freely.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Static anycast IP address
A fixed IP address that Global Accelerator assigns to an accelerator. Clients and partner firewalls can allowlist it, and it does not change when endpoints change.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Statically stable
A design that keeps serving during a failure without making control plane calls such as scaling, usually by pre-provisioning enough capacity.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Step scaling
An Auto Scaling policy tied to a CloudWatch alarm you create, with step adjustments that add or remove more capacity as the alarm breach grows.
Lesson: High-Performing and Elastic Compute (task 3.2)
Sticky sessions
A load balancer setting that keeps sending a user's requests to the same target. A stopgap for stateful apps: the session is lost if that target fails.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Storage autoscaling
An RDS setting that increases allocated storage automatically up to a maximum storage threshold. It can't reduce storage and can't completely prevent storage-full states during very large data loads.
Lesson: High-Performing Databases (task 3.3)
Storage throughput
How much data a storage device moves per second, usually in MiB/s. The key number for large, sequential reads and writes such as logs and media.
Lesson: High-Performing and Scalable Storage (task 3.1)
Strongly consistent read
A read that returns the most recent data, reflecting every write that succeeded before it. In DynamoDB it costs twice an eventually consistent read and bypasses DAX.
Lesson: High-Performing Databases (task 3.3)

T

Tape Gateway
A Storage Gateway type that presents a virtual tape library over iSCSI so existing backup software can write virtual tapes that are stored in AWS.
Lesson: High-Performing and Scalable Storage (task 3.1)
Target group
The set of targets, such as instances, IP addresses or Lambda functions, that a load balancer routes one kind of request to and health-checks.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Target tracking scaling
An Auto Scaling policy that keeps a metric, such as average CPU or requests per target, near a value you choose. Auto Scaling creates and manages the CloudWatch alarms.
Lesson: High-Performing and Elastic Compute (task 3.2)
Task definition
The Amazon ECS blueprint for running containers: images, CPU and memory, networking mode, roles, volumes and logging.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Tenancy
Whether an EC2 instance runs on shared hardware (the default) or on single-tenant hardware dedicated to one account.
Lesson: Cost-Optimized Compute (task 4.2)
Tight coupling
A design where one component must be up, fast and unchanged for another to work, so a failure or slowdown in one spreads to the other.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Time to live (TTL)
How long a cache keeps a copy before it expires and must be fetched or checked again. CloudFront's default is 24 hours unless the cache policy sets another value.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Traffic policing
Dropping packets that exceed a configured rate. Hosted Direct Connect connections use it, so bursts above the rate lose packets.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Transformation context
A name on a source, transform or sink in an AWS Glue script. Job bookmarks use it to track which data each one has already processed.
Lesson: High-Performing Data Ingestion and Transformation (task 3.5)
Transit gateway
A regional network hub that connects VPCs, VPN connections and Direct Connect gateways. Its route tables decide which attachment can reach which, and it routes transitively.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Transit gateway attachment
A connection from a VPC, VPN, Direct Connect gateway or peer transit gateway to a transit gateway.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Transit gateway route table
A table in a transit gateway that holds routes to its attachments. Association and propagation control which routes each attachment uses and which it advertises.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Transitive routing
Routing that carries traffic through an intermediate network to a third one. Transit gateways support it; VPC peering does not.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)
Trust policy
The resource-based policy attached to an IAM role that says which principals may assume it, optionally with conditions such as MFA or an external ID.
Also called: Role trust policyLesson: Secure Access to AWS Resources (task 1.1)

V

Vertical scaling
Moving a component to a larger or smaller instance size; limited by the largest size and still a single copy.
Also called: Scaling upLesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Virtual private gateway
The VPC's side of a Site-to-Site VPN or Direct Connect connection, attached to one VPC.
Lesson: Secure Workloads and Applications (task 1.2)
Visibility timeout
The period after an SQS consumer receives a message during which other consumers can't see it. If the message isn't deleted before it ends, it becomes visible again and is processed again.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)
Volume Gateway
A Storage Gateway type that presents iSCSI block volumes on premises. Cached volumes keep primary data in AWS; stored volumes keep it locally and back it up to AWS.
Lesson: High-Performing and Scalable Storage (task 3.1)
VPC peering connection
A one-to-one networking connection between two VPCs that uses private IP addresses. It is not transitive, and the two VPCs' CIDR blocks must not overlap.
Lesson: High-Performing and Scalable Network Architectures (task 3.4)

W

Warm pool
Pre-initialized EC2 instances kept beside an Auto Scaling group in a stopped, running or hibernated state, so scale-out can use them instead of booting new instances from scratch.
Lesson: High-Performing and Elastic Compute (task 3.2)
Warm standby
A DR strategy that keeps a scaled-down but fully functional copy of the workload running in the recovery Region, ready to scale out.
Lesson: High Availability and Disaster Recovery: Strategies, Failover and Single Points of Failure (task 2.2)
Warm throughput
A DynamoDB setting that defines the read and write rates a table or index can absorb instantly without throttling. It can be raised before a known traffic event.
Lesson: High-Performing Databases (task 3.3)
Write capacity unit (WCU)
DynamoDB's unit of write throughput: one write per second of an item up to 1 KB.
Lesson: High-Performing Databases (task 3.3)
Write sharding
A DynamoDB pattern that adds a suffix to a hot partition key value so that writes spread across several partition key values. Reads then fan out across the suffixes.
Lesson: High-Performing Databases (task 3.3)
Write-through cache
A caching strategy that updates the cache every time the database is written, so cached data stays fresh at the cost of extra writes.
Lesson: Messaging and Events: Decoupling with Queues, Topics, Buses and Workflows (task 2.1)

Z

Zonal NAT gateway
A NAT gateway in one Availability Zone. Workloads in other zones that use it also pay cross-AZ data transfer, so production designs put one in each AZ.
Lesson: Network Cost (task 4.4)
Zonal Reserved Instance
A Reserved Instance bought for one Availability Zone. Unlike a regional RI, it reserves capacity there, but its discount applies only to that exact instance type in that zone.
Lesson: Cost-Optimized Compute (task 4.2)