What is the Amazon MQ migration guide?
Amazon MQ is a managed message broker service that makes it easy to migrate to a message broker in the cloud.
Amazon MQ currently supports Apache ActiveMQ
Concepts for migrating to Amazon MQ
Before migrating, review the following key concepts to consider when migrating a commercial message broker to Amazon MQ.
Messaging protocols
You can connect your broker to Amazon MQ without any code changes if you currently use one of the following industry-standard protocols:
For more information about connecting to an Amazon MQ managed broker, see Working Examples of Using Java Message Service (JMS) with ActiveMQ in the Amazon MQ Developer Guide.
Message persistence
To replicate persistence mode or sync point control options with Amazon MQ, you can deploy your brokers as active/standby brokers. In the active/standby deployment, brokers use shared storage across multiple Availability Zones, with an optional time to live (TTL).
For more information about how Amazon MQ ensures message durability, see Availability options.
Network options
Depending on the interlopability of your applications and the type of access that they need, you can permit public access, VPN access, or VPC access using Amazon Virtual Private Cloud (Amazon VPC).
In a hybrid architecture where on-premises systems
need access to resources in the cloud, we recommend setting up your Amazon MQ managed
brokers with
public
network access. You can also achieve a hybrid solution
by using
Amazon VPN
Tip
If your resources are primarily deployed within the Amazon Cloud, we recommend configuring your Amazon MQ brokers with Amazon VPC. For network access across multiple VPCs, you can use VPC peering.
Availability options
Amazon MQ supports durability-optimized brokers backed by Amazon Elastic File System (Amazon EFS)
Note
In the event of a broker or Availability Zone failure, active/standby brokers automatically fail over to a standby instance in another Availability Zone.
To achieve high availability and message durability, you can use a network of brokers. A network of brokers is a series of simultaneously active single-instance or active/standby brokers that allows you to rapidly scale your throughput and connection count. You can configure a network of brokers in a variety of topologies depending on your application's needs.
Messaging patterns
Amazon MQ offers the following topology options to support a variety of messaging patterns:
-
Point-to-point
-
Request-response
-
Hub and spoke
-
Mesh
-
Enterprise service bus
For more information about using Amazon MQ to set up the right broker topology for your cloud architecture, see Amazon MQ Broker Architecture in the Amazon MQ Developer Guide.
Important
Revising a broker configuration or an ActiveMQ user does not immediately apply those changes. For your changes to take effect, you must wait for the next maintenance window or reboot the broker. For more information, see Amazon MQ Broker Configuration Lifecycle in the Amazon MQ Developer Guide.
Performance and scalability
With Amazon MQ you can scale your messaging middleware horizontally, vertically, or in a hybrid model.
Horizontal scaling enables you to increase your throughput and connection count without interruptions, because your resources remain active and online. To scale horizontally, you can deploy a network of brokers in an active/standby configuration across multiple Availability Zones.
To
scale your resources vertically, you can
increase the compute capacity of your broker instances from
mq.t2.micro
(1 vCPU and 1 GiB) up to mq.m5.4xlarge
(16 vCPU and 64
GiB). For more information about Amazon MQ instance types, see Instance
Types in the Amazon MQ Developer Guide.
Note
Choosing larger broker instance types might not improve overall system throughput. Overall latency is due to many factors, such as message size, the type of protocol, the number of active producers and consumers, consumption speed, and message persistence.
Amazon MQ also supports creating throughput-optimized message brokers backed by Amazon Elastic Block Store (Amazon EBS)
To instruct Amazon MQ to optimize for queues with slow consumers, set the
concurrentStorageAndDispatchQueues
attribute to
false
.
The following table shows the throughput of an mq.m5.2xlarge
broker
configured with these options:
-
Broker instance –
mq.m5.2xlarge
-
Persistent –
TRUE
-
Client –
m5.xlarge
-
CSAD –
TRUE
-
Protocol – OpenWire
Producers/Consumers | |||||
---|---|---|---|---|---|
Message size | Metrics | 25 | 50 | 100 | 200 |
1 KB | TPS | 2,250 | 4,300 | 8,467 | 16,334 |
CPU% | 8% | 15% | 27% | 58% | |
5 KB | TPS | 2,067 | 3,834 | 7,150 | 14,516 |
CPU% | 10% | 17% | 30% | 63% | |
10 KB | TPS | 1,900 | 3,467 | 7,083 | 11,334 |
CPU% | 15% | 24% | 48% | 80% | |
50 KB | TPS | 1,592 | 2,917 | 4,500 | 4,917 |
CPU% | 30% | 52% | 83% | 92% | |
100 KB | TPS | 1,250 | 2,184 | 2,513 | 2,770 |
CPU% | 42% | 72% | 85% | 92% |
Note
Performance numbers can vary depending on multiple configuration parameters. For more information on Amazon MQ throughput measurements, see Throughput Benchmarks.
You can measure the throughput of your Amazon MQ brokers using JMS
Benchmark
Latency
You can set up your Amazon MQ brokers for low-latency messaging, with latency often as low as single-digit milliseconds. Use an always-on connection to help reduce the amount of time that it takes to deliver messages to a consumer.
Using in-memory storage can further reduce overall latency across your messaging architecture. For more information on how different storage types can affect latency, see Differences Between Storage Types in the Amazon MQ Developer Guide.
Destination options
You can set up Amazon MQ managed brokers as queues or
topics. Amazon MQ queues are, by default, first in, first out (FIFO)
queues,
also known as ordered queues. You can scale FIFO queues using Message Groups
Topics in Amazon MQ use the publisher/subscriber pattern and can be
durable or non-durable. Amazon MQ topics also support Virtual
Destinations
Tip
You can optimize and fine-tune the performance of your topics. For more
information, see Performance Tuning
Security and authentication
With Amazon MQ you control who is
allowed
to create or modify brokers, and which applications are allowed to send and receive
messages. For more information about authentication options, and how to integrate the
Lightweight Directory Access Protocol (LDAP)
Connections to Amazon MQ brokers use Transport Layer Security (TLS). To isolate your brokers in a private virtual network, you can restrict access to a private endpoint within a VPC. To control network access to your brokers, you can configure security groups in the VPC. For more information, see Security Best Practices for Amazon MQ in the Amazon MQ Developer Guide.
Amazon MQ encrypts messages at rest and in transit using encryption keys that it manages
and stores securely in Amazon Key Management Service (Amazon KMS)
Tip
For additional security, we highly recommend designing your application to use client-side encryption.
For more information about Amazon MQ security and how messaging data is encrypted, see Data Protection in Amazon MQ in the Amazon MQ Developer Guide.
Broker quotas
By default, each Amazon MQ broker can support 1,000 connections (or 100 connections for
mq.t2.micro
brokers). To allow multiple consumers to share connections
to your Amazon MQ brokers and to improve overall performance, we recommend using
pooled connections. For more information, see Always Use Connection Pooling in the
Amazon MQ Developer Guide.
You can request an increase for many broker usage quotas for your Amazon account. For more information, see Amazon service quotas in the Amazon Web Services General Reference.
Configuration options
Amazon MQ supports standard JMS features including point-to-point (message queues), publish-subscribe (topics), request/reply, persistent and non-persistent modes, JMS transactions, and distributed (XA) transactions.
Amazon MQ brokers can also support more complex messaging patterns such as composite destinations, which enable producers to send the same message to multiple destinations, and virtual destinations, which enable publishers to broadcast messages via a topic to a pool of receivers subscribing through queues.
For more information, see Amazon MQ Broker Configuration Parameters in the Amazon MQ Developer Guide.
Cost estimation
With Amazon MQ, you pay only for the provisioned capacity that you use. Factors such as
broker instance type, the amount of data stored on each broker instance,
and the Amazon Region in which you deploy your brokers can affect your total cost of ownership.
To estimate your broker costs for Amazon MQ, you can us the Amazon Pricing Calculator
Note
For data transferred in and out of Amazon MQ, you pay standard Amazon data transfer charges.
To get started, Amazon MQ offers a Free Tier, which includes up to 750 hours of a
single-instance mq.t2.micro
or
mq.t3.micro
broker per month, and up to 5 GB of
durability-optimized storage per month for one year. For more
information on the Free Tier, pricing, and associated costs, see Amazon MQ Pricing