Skip to content

Route53.md

Latest commit

 

History

History
113 lines (74 loc) · 5.23 KB

Route53.md

File metadata and controls

113 lines (74 loc) · 5.23 KB

What is Amazon Route 53?

  • Managed DNS & Domain Registry: AWS’s highly available and scalable Domain Name System (DNS) service, launched in December 2010. The name refers to DNS’s port (TCP/UDP 53).
  • DNS resolution for AWS & external endpoints: Routes users to AWS services (EC2, ELB, S3, CloudFront, API Gateway, WorkMail, Lightsail, VPC endpoints) and non-AWS resources.
  • Alias record type: Proprietary server-side equivalent of CNAMEs—resolves at apex domains and supports AWS targets without extra cost.

Integration with Other AWS Services

  • CloudTrail: Logs all Route 53 API calls for auditing.
  • CloudWatch: Monitors health-check status and metrics like DNS query volume, latency.
  • Well-Architected integrations: Use Alias records to point to ELB, CloudFront, S3 static sites, API Gateway, WorkMail, VPC endpoints.
  • Route 53 Resolver: Provides recursive DNS within VPCs, plus inbound/outbound endpoints for hybrid (on-prem ⇄ AWS) DNS, forwarding rules, and firewall features.
  • Traffic Flow / Application Recovery Controller: Advanced routing (latency-based, geo, geoproximity, weighted, failover) and DNS-level failover and recovery controls.

Routing Policies & Use Cases

  • Simple: One-to-one mapping—good for straightforward setups.
  • Weighted: Distribute traffic (e.g., blue-green, A/B testing).
  • Latency-based: Route to lowest-latency region for global apps.
  • Geolocation / Geoproximity: Route based on user location or proximity; “bias” allows traffic steering.
  • Multi-value Answer: Returns multiple healthy IPs; simple load distribution.
  • Failover: Active-passive high availability with health checks.
  • IP-based: Route based on client IP ranges—useful for internal vs external traffic logic.

Use Cases:

  • Global applications: Improve performance and availability using latency or geo routing.
  • Disaster recovery: Failover routing with health checks.
  • Testing/deployment: Weighted routing for traffic shifting.
  • Hybrid DNS: Use Resolver endpoints for on-premises integration.
  • Apex domain support: Alias records for root domain to AWS targets.

Similar Services (AWS & External)

  • Within AWS:

    • Route 53 Resolver (intra-VPC DNS and hybrid scenarios).

  • External DNS providers:

    • Google Cloud DNS, Azure DNS, Cloudflare DNS, Dyn, NS1.

Limits & Quotas

Route 53 has service quotas you can view or request increases via AWS Service Quotas :

Hosted zones:

  • Default: 500 per account (increased on request).
  • Delegation sets: 100 per reusable set.

Records:

  • Max 10,000 records per hosted zone; >10,000 incurs extra charges.
  • 400 records per record set; 100 geolocation/weighted/etc. of same name+type.
  • Geoproximity: 30 same name/type; CIDR collections limited.

Resolver (Hybrid):

  • Endpoints: 4 per region; 6 IPs per endpoint; 1000 rules; 2000 associations.
  • UDP QPS: ~10,000 per IP; use multiple ENIs for higher loads.

Costs:

  • Resolver endpoints: ~$0.125 per ENI per hour + query charges ($0.40/million up to 1B; $0.20 thereafter).

Exceptions:

  • LimitExceededException when quotas are exceeded—handle via optimization or Service Quotas request.

Best Practices

  • Use Alias records to AWS targets for cost and flexibility.
  • Balance TTL: low TTL = fast changes, high traffic; high TTL = cache efficiency but slower updates.
  • Monitor DNS health and metrics via CloudWatch and set alerts.
  • Use multi-AZ subnets for Resolver endpoints; avoid routing loops; specify static IPs for resiliency.
  • For Resolver, enable capacity utilization metrics and integrate alarms for scaling.

Wrong Use Cases (When not to use Route 53)

  • Not a substitute for ELB: Use multi-value answer only for simple distribution; not load balancing.
  • Low QPS at Resolver endpoints: Cheap approach but may cost more; better to use private hosted zone associations if within AWS.
  • Complex routing internal only: Avoid if overcomplicating; internal tools or simpler DNS may suffice.
  • Using Route 53 where regional DNS or internal corporate DNS is preferred: e.g., for internal-only resolution without VPC/VPC endpoints.

Quick-Reference Summary

Topic Key Points
Definition Managed DNS + Domain registrar; supports AWS & non-AWS
Routing Simple, Weighted, Latency, Geo, Geoproximity, Multi-value, Failover, IP-based
Integration Works with ELB, CloudFront, S3, API GW, VPC, CloudWatch, CloudTrail
Limits Hosted zones, records, Resolver rules/endpoints, QPS, cost quotas
Best Practices Use Alias, monitor metrics, choose TTL wisely, use multiple AZs, capacity metrics
Wrong Uses Not for load balancing, internal DNS only, cost-heavy hybrid usage, over-engineering