Routing policies

Routing policies can be configured in both public and private hosted

zones on Amazon Route 53. Some key points:
• Routing policies allow you to control how Amazon Route 53
responds to DNS queries and routes traffic to resources. Supported
policies include simple, failover, geolocation, latency-based, and weighted
routing.
• Private hosted zones are useful when you want to route traffic using
Route 53 but only within a VPC, not to resources accessible on the public
internet. This provides a private DNS system for your VPC.

• All the standard routing policies (simple, failover, geolocation etc.)

can be used within private hosted zones to control routing between
resources in the same VPC or across multiple VPCs.
• You would typically use a public hosted zone when you want to
route traffic to resources that have public IP addresses or public domain
names accessible from the internet. The same routing policies apply in
public hosted zones
Routing policies in Route 53 allow you to control how traffic is routed to
your resources. The main routing policies are:
1. Simple routing - Used for a single resource like a web server, serving
content for a domain.

2. Failover routing - Used to configure active-passive failover between

two resources like server instances.

3. Geolocation routing - Routes traffic based on the location of users,

allowing you to localize content delivery.

4. Geoproximity routing - Similar to geolocation but also allows

shifting traffic between resources based on their locations.
5. Latency routing - Routes traffic to resources in different AWS
regions based on lowest latency.
6. Weighted routing - Distributes traffic to multiple resources in
proportions that you specify.

7. Multivalue answer routing - Returns up to 8 healthy records

randomly for a query.

8. IP address-based routing - Routes based on IP addresses of users in

addition to their location.
You can use routing policies to control traffic routing in both public and
private hosted zones on Route 53. Private hosted zones are useful when
you want routing only within a VPC. The appropriate routing policy
depends on your specific requirements and architecture.

Here are the explanations of the different routing policies supported by

Amazon Route 53:

1. Simple Routing Policy:

o Used for a single resource like a web server, serving content for a
domain. Route 53 responds with the resource record set specified.

2. Failover Routing Policy:

o Used to configure active-passive failover between two resources like

EC2 instances. One resource receives all traffic when available, the other
receives traffic when the first is unavailable.
o You can specify whether to fail over all traffic or a percentage of
traffic upon health check failure.

o Allows for more control over failover process than simple failover
policy.

3. Geolocation Routing Policy:

o Routes traffic based on the location of users, allowing localization of

content delivery. Provides continent, country and state-level granularity.

o You specify regions/locations of resources and Route 53 routes

users to the closest resource based on their location.
4. Geoproximity Routing Policy:

o Similar to geolocation but also allows shifting traffic between

resources based on their locations. Specifying a bias can send more or less
traffic to a resource.

5. Latency Routing Policy:

o Routes traffic to resources in different AWS regions based on lowest

latency measurement to improve end-user experience.

6. Weighted Routing Policy:

o Distributes traffic to multiple resources in proportions that you

specify using weights.

o Resources with higher weights receive more traffic than those with
lower weights.
7. Multivalue answer routing policy:
o Allows associating multiple values (e.g. IP addresses) with a single
DNS name (domain), and Route 53 will return all values to the caller in a
random order.
o Useful for load balancing, A/B testing, rolling deployments etc. by
exposing multiple variants of resources behind a single name.
IP address-based routing

IP address-based routing allows you to customize routing based on the IP

addresses of clients sending DNS queries. This gives you more control
over routing decisions compared to routing policies that rely on data
collected by Route 53.
Some common use cases for IP-based routing include:
1. Routing end users from certain internet service providers (ISPs) or
networks to preferred endpoints. This can help optimize network transit
costs and performance.
2. Overriding default routing like geolocation-based routing by adding
custom rules based on your knowledge of client IP addresses and physical
locations.

To configure IP-based routing, you need to create a traffic policy in Route

53 that maps IP addresses or ranges to resource record sets. You then
associate this traffic policy with a hosted zone.
When a query comes from a matching IP/range, it will be routed to the
specified endpoint. This allows you to customize routing at a more
granular level compared to policies like geolocation routing.
Multivalue answer routing

Multivalue answer routing policy allows you to associate multiple resource

record sets with a single DNS name (domain). Route 53 will then return all
the values to the DNS query originator in a random order.
Some key differences compared to a simple routing policy:

1. Each resource is represented as an individual record set with

optional health checks.
2. Route 53 will only include healthy records in the response based on
the health check status.

3. This ensures availability as if one resource fails, clients can try the
other healthy ones from the response. With simple routing, all values are
always returned regardless of health.
4. Latency is reduced as clients don't have to retry an unhealthy record.

Weighted routing

Weighted routing policy in Route 53 allows you to associate multiple

resource record sets with a single domain or subdomain and specify how
much traffic is routed to each resource by assigning weights.

The key aspects of weighted routing are:

1. You create multiple records with the same name and type for each
resource you want to route traffic to.

2. Each record is assigned a relative weight that corresponds to the

percentage of traffic to route to that resource.

3. Route 53 calculates the proportion of traffic to send to each resource

based on its weight as a fraction of the total weights. Higher weight means
more traffic.

4. This allows flexible traffic distribution that can be gradually adjusted

over time by changing weights. Resources can also be disabled by
assigning zero weight.

5. Weighted routing does not consider the client IP address, so the

same client may not always be routed to the same endpoint on
subsequent queries.
To achieve session stickiness, one approach is to start by routing a small
percentage of traffic using geolocation or latency-based routing policies
instead of weights. These policies will route the same client back to the
same region/endpoint based on source IP. The percentage can then be
gradually increased over time.
Latency routing

Latency routing policy in Route 53 allows you to route traffic to the

optimal endpoint based on network latency. Here is how it works:

1. You create latency resource record sets for your resources such as
EC2 instances, ELBs etc. hosted in different AWS regions.

2. Route 53 measures latency from various locations worldwide to

these resources continuously.

3. When a DNS query comes from a client, Route 53 determines the

client's location and refers to its latency data to select the lowest latency
endpoint.
4. The IP address or alias of the selected endpoint is returned in the
DNS response.

5. Route 53 continues monitoring latency and may route the same

client to a different endpoint over time if network conditions change.

This ensures your users are always routed to the fastest regional endpoint
for optimal performance. You don't need to manage complex
configurations - Route 53 handles it automatically based on real-time
latency data.

Geoproximity routing
Geoproximity routing policy in Route 53 allows you to route traffic based
on the geographic location of your users and your resources.

With geoproximity routing, you specify the location of each resource using
AWS regions/local zones or latitude-longitude coordinates.

Route 53 continuously monitors the network latency from various

locations to these resources. When a DNS query comes in, it determines
the location of the user and selects the lowest latency resource location
to route the traffic to.
You can also optionally apply a bias value (+/- integer from 1-99) to each
resource. A positive bias expands the size of the geographic region from
where traffic is routed to that resource, while a negative bias shrinks the
region. This allows shifting some percentage of traffic between resources.

The main benefit is that users are always routed to the fastest endpoint
based on their location, ensuring high performance. Route 53 handles the
complex configurations automatically based on real-time latency data.

Some key points about geoproximity routing:

• Requires Route 53 Traffic Flow to be enabled

• Locations can be AWS regions/zones or lat-long
• Bias expands or shrinks routed geographic region size

Geolocation routing
Geolocation routing allows you to route traffic based on the geographic
location of your users. With geolocation routing you can create multiple
records for the same domain or subdomain, with each record associated
with a different geographic location like continent, country or state.

When a DNS query is received, Route 53 determines the location of the

user and returns the record associated with the smallest matching
geographic location. This allows you to localize content and restrict
distribution based on location.

Some key benefits of geolocation routing include:

1. Presenting content in the language of the user's location.

2. Restricting distribution of content only to certain locations.

3. Load balancing across endpoints in a predictable way so the same

user location is consistently routed to the same endpoint.

4. Combining with other routing types like latency-based routing for

low latency architectures.
To get started with geolocation routing, you can create geolocation
resource record sets using the AWS Management Console or Route 53
API. The record set is marked as geolocation and associated with the
applicable geographic location.

Refer to the Amazon Route 53 documentation for more details on

configuring geolocation routing and geographic locations supported.

Failover routing

Failover routing policy in Route 53 allows you to configure active-passive

failover between multiple resources like EC2 instances or ELB load
balancers. Here's how it works:
1. You create primary and secondary resource record sets for your
active-passive resources. For example, a primary record for your web
application on EC2 and a secondary record for a backup application.

2. You associate health checks with the records to monitor the health
and performance of the resources.

3. When the primary resource is healthy, Route 53 returns the IP

address of the primary record in responses to DNS queries.

4. If the health check for the primary resource fails, Route 53 will
automatically start returning the IP address of the secondary record
instead.

5. This failover happens within minutes and ensures your users are not
impacted by any outages and are seamlessly redirected to the backup
resource.

6. You can optionally associate a health check with the secondary

record as well. If both primary and secondary fail health checks, Route 53
enters a "failover emergency" mode and may return records for unhealthy
resources.
Simple routing

The simple routing policy is used for a single resource that performs a
given function for your domain, such as a web server that serves content
for a website.

With simple routing, Route 53 responds to DNS queries with the value
(e.g. IP address) for the record set. It is useful when you have a single
resource like a web server and want all traffic to be directed to that
resource.

Some key points about simple routing:

• Used for a single resource like a web server
• Route 53 returns the value of the record set in response to DNS
queries

• Useful for directing all traffic to a single resource

• Can be used to create records in private hosted zones

For more details on choosing the appropriate routing policy, please refer
to the Route 53 documentation. It provides details on all routing policies
and how to configure them based on your specific use case and
architecture.
***