Magento 2.

0
Site Performance and
Scalability Optimizations
Table of Contents

1 Introduction
2 Recommended Magento Configurations
4 Layered Architecture
5 Performance and Scalability Improvements
5 Client-side Improvements
6 Page Caching
7 Application Enhancements
7 Asynchronous Order and Product Updates
7 Job Queue Mechanism
8 PHP Interpreters
8 Database Improvements
8 Multiple Database Masters
10 MySQL Cluster Support
10 Performance Toolkit

10 Performance Results
10 Test Configuration
10 Software
11 Scenarios
11 Testing Methodology
12 Small Merchant Deployment
12 Throughput Results
13 Server Response Time Results
14 Large Merchant Deployment
14 Throughput Results
15 Server Response Time Results

17 Conclusion
17 Small Merchant Deployment
18 Large Merchant Deployment

19 Appendix
19 Merchant Profile Details
19 Small Merchant Deployment (4 web node cores)
20 Larger Merchant Deployment (20 web node cores)

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS

Introduction
Magento 2.0 was designed for the next era in eCommerce—to make it
faster, easier and more cost effective than ever before to create shopping
experiences that are tailored precisely to your business needs.
It all starts with a modern architecture based on popular frameworks and coding patterns that gives you
full flexibility to meet virtually any business need with speed and agility. Magento 2.0’s modular design
reduces application complexity and enables you to easily extend or customize core functionality with
off-the-shelf extensions or your own code so you can bring innovative ideas to market fast. A built-in
automated testing framework also speeds development and you can easily integrate Magento 2.0 with
3rd party solutions thanks to enhanced and efficient APIs.

Magento 2.0 also offers new functionality to help you create powerful shopping experiences. An all-
new, streamlined checkout and enhanced responsive designs provide you with an improved toolset
for differentiating your site and growing sales across devices. Redesigned administrative controls and
merchandising tools enable more efficient business operations and faster onboarding of new team
members so you can deliver better service to your customers. And, with easier upgrades enabled by
Magento 2.0’s modular architecture, you can immediately take advantage of new functionality as it
becomes available.

On top of this flexible and feature-rich new platform, Magento 2.0 also includes comprehensive changes
that set a new standard for performance and scalability. Out-of-the-box, Magento 2.0 can now:

• Process significantly more orders per hour on the same hardware

• Offer near-instant server response times for catalog browsing
• Deliver double-digit decreases in response times for cart and checkout pages
• Better handle peak order volume, extra-large catalogs, and outsized customer lists
• Support significantly more simultaneous administrative users on the backend

This paper explores the Magento 2.0 performance and scalability enhancements in depth to help you
understand how the new platform can benefit your business and support your long-term growth plans. It
also provides configuration recommendations and benchmark testing results that you can use to optimize
your Magento 2.0 site performance.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 1

Recommended Magento Configurations
Magento provides a wide range of deployment and configuration options to support different use cases.
To optimize performance, many of the settings within Magento must be correctly configured in order to take
advantage of the performance and scalability enhancements that are included in Magento 2.0. There are
also specific recommended environment configurations that are beyond the scope of this white paper.

IMPORTANT NOTE: All tests referenced in this paper are based on recommended and supported
configurations for Magento platforms. Deviation from these configurations may results in varied results,
which may not accurately reflect the nature in which the product was designed and architected.

Core Configuration
Magento 2.0 is optimized to run using PHP7 and Varnish. Although it is possible to run with PHP5.6 and
without Varnish, it is not recommended. It is advised that sites use both PHP7 and Varnish to support
the operation of their business and to achieve optimal performance results.

Static Asset Settings

The static assets setting controls how assets such as CSS, JavaScript, HTML, and images are prepared
for use by the website.

• Stores->Configuration->Advanced->Developer:
- Grid Settings: Asynchronous indexing: Enable
- CSS Settings: Minify CSS Files: Yes
- JS Settings: Minify JS Files: Yes
- Merge JavaScript Files: No
- Template Settings: Minify HTML: Yes
- JavaScript Settings: Enable JavaScript Bundling: Yes

Page Cache
The page cache setting controls which version of the page cache is used. For the best performance,
enabling the Varnish cache is strongly recommended. Note that this requires the deployment and
configuration of Varnish.

• Stores->Configuration->Advanced->System->Full page cache

- Caching Application->Varnish Caching

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 2

Email Settings
Email settings control when emails are sent relative to when they are generated.

• Stores->Configuration->Sales->Sales Emails:
- General settings: Asynchronous sending: Enable

Index Settings
• System->Index Management:
- All indexers should be in “Update on schedule” mode

Application Cache
• Configure the platform to use Redis in env.php
- Requires deployment and configuration of Redis

Production Mode
Production mode can either be enabled using a single command:

php bin/magento deploy:mode:set production

or can be broken into multiple steps if desired to give developers additional control over the process:

php bin/magento setup:di:compile

php bin/magento setup:static-content:deploy

set MODE production manually in web server config

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 3

Layered Architecture
The Magento 2.0 platform is a layered architecture with four tiers that are optimized for performance and
scalability. These are: the client, the page cache, the application, and the database tiers.

This standard web application structure allows for different components of the architecture to be scaled
independently in order to meet the requirements of different business use cases and different loads on
the system.

ADVANCED TOPOLOGY

CLIENT TIER
Client Workstation

Web Browser Browser Cache

PAGE CACHE
Varnish Reverse Proxy
and Full Page Cache

Load Balancer

APPLICATION TIER
Magento Memcache
(session + markup
Apache/Nginx + PHP + data cache)

Load Balancer

DATABASE TIER

MySQL MySQL
Search
Master R/O Replicas

FIGURE 1: Magento Layered Architecture

The following section describes the performance and scalability improvements made in each of these tiers
in Magento 2.0.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 4

Performance and Scalability Improvements
Magento 2.0 improves how the application interacts with the browser as well as reduces the amount of
data that must be returned from the server to render web pages. These changes significantly improve page
response and load times, giving you a much better starting point for developing a fast site.

Client-side Improvements
Page load time is a critical factor for creating a successful eCommerce site and has been shown to
directly impact consumer engagement and conversion rates. One key factor in providing a fast page load
and good customer experience is reducing the overall page weight or the amount of data that must be
loaded to render the page. The amount of data required to render a web page is both a function of the
visual design of the page and the technologies and web development strategies used.

Magento 2.0 provides a number of optimizations to reduce page weight and improve response times:

• Magento 2.0 optimizes the assets required to render the page by using minification.
Minification is a process of removing whitespace, comments, and other unnecessary characters
in order to reduce the size of files served to the browser. With Magento 2.0, all HTML, CSS, and
JavaScript served by the application are minified to reduce their size. This minification is carried
out in advance of requests and is cached so the minified files are available for all website requests.
Magento uses industry-standard minification engines such as jShrink for Javascript and cssmin for
CSS, as well as an internally-developed engine for HTML. These engines were selected to maximize
compression, but merchants can choose to use others if they desire. To further reduce page
weight, image files used by Magento websites are compressed using the PHP gd2 library to provide
high quality images in the smallest file size.
• Magento 2.0 maximizes the usage of the browser cache for storing assets. All JavaScript
and CSS are loaded to the browser and cached on the initial web page visited. This minimizes
the amount of content that needs to be downloaded to render any subsequent page. To further
improve the ability of the cache to service page requests and allow the browser to begin rendering
the page, private data (such as consumers’ names) are no longer provided as HTML blocks. Now
they are returned to the browser as JSON and are inserted into the rendered page when available.
• Magento 2.0 improves response time by using asynchronous processing during checkout
and cart operations. Items are added to the cart asynchronously to improve response times for
the browsing experience. Many checkout operations are also done asynchronously to improve
response time, such as address validation and cart updates.

Collectively, these optimizations reduce the amount of data that must be passed to the browser in order
to render the webpage. These approaches help ensure that nearly any page design can be efficiently
rendered for speed.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 5

Page Caching
Another key to delivering fast page loads is improving the server response time. Magento 2.0 makes
extensive use of caching of page content and static assets on the server to accelerate response times.

With Magento 2.0, the application now directly integrates with Varnish page caching out-of-the-box.
Varnish is a reverse-HTTP proxy or web accelerator. The Varnish application stores HTML and other files or
fragments so that they can be returned extremely quickly.

Varnish provides a very fast and efficient mechanism for serving content that is highly scalable. The
requests that are served by the Varnish cache never need to reach the Magento application servers,
which reduces the load on the web nodes while dramatically improving the response time. One or more
Varnish servers can be used in front of the Magento server(s), an approach that provides a much faster
and more efficient system.

Cached content can be composed of different elements with different lifespans or times-to-live using
Edge Side Includes (ESI). Edge Side Includes allow for different cached elements to be combined
dynamically before being served from the cache. Page elements that cannot be cached, such as cart
contents, customer name, or other private data, are provided separately. These page elements are passed
to the browser as JSON that can be rendered asynchronously in the browser rather than on the server.
This approach allows the page to begin rendering as quickly as possible with the other content added in
when it is available, creating a superior user experience. Magento 2.0 handles invalidation of content in
the page cache to ensure the right content is served to the website.

Magento 2.0 opens the benefits of Varnish caching to a much broader set of merchants by fully
supporting and providing configuration files (.vcl files) for Varnish 3.X and 4.X. These configurations must
be updated to reflect the deployment topology, but provide a starting point that makes it significantly
easier and less expensive to implement Varnish to improve performance and scalability.

There can be a number of different ways to architect Varnish in the application. The simplest option is to
deploy Varnish on the same server that is running Magento. More sophisticated deployment architectures
can be used to scale out multiple Varnish servers using a load balancer to distribute traffic between
multiple Varnish instances on multiple servers. These deployments can improve scalability and provide
redundancy as part of a high availability system.

To use Varnish with secure pages or sites (i.e., served using HTTPS), another application (such as Nginx)
must be placed in front of the Varnish server(s) to handle SSL termination.

Magento 2.0 does provide an alternative Page Cache in addition to Varnish. This cache is intended
for use in development or in single web node situations where Varnish cannot be used. However, the
Page Cache uses a PHP implementation that requires the Magento application to process the caching
requests, which causes it to be less efficient and to not yield the same performance benefits. Due to
the superior performance and scalability characteristics of Varnish, it is strongly recommended for
use in production deployments.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 6

Application Enhancements
Magento 2.0 is architected to meet the scalability and stability needs of large and growing merchants
and enterprises. The following server-side improvements are implemented in Magento 2.0 to eliminate
blocking operations or gridlocks and to improve efficiency of backend business operations:

Asynchronous Order and Product Updates

Large enterprises with lots of orders and product data tend to have multiple Admin users working
concurrently on the backend. Magento 2.0 introduces optional asynchronous updates for order
management and product data to make concurrent operations efficient and to eliminate gridlock or
blocking operations during updates.

For example, when a server is operating at full capacity, an Admin making updates to product descriptions
or pick, pack and ship orders is still able to quickly interact with Magento, as the actual processing is queued
for later. This means backend Application performance is dramatically improved for situations where there
are over 50+ Admin users simultaneously making order updates and over 25+ Admin users simultaneously
making product updates.

Job Queue Mechanism

To further advance scalability and responsiveness, additional enhancements were made in our commercial
product, Magento Enterprise Edition 2.0. One such Enterprise Edition-only scalability improvement is a
job queue based on Rabbit MQ, which allows for asynchronous processing of jobs. Queue workers pull
jobs that are placed on the queue when they have capacity. Queue workers can operate using separate
server resources from the Magento application servers. This allows the two environments to be optimized
separately for their respective loads.

WEB SERVERS

QUEUE SERVER QUEUE SERVER RabbitMQ

JOB SERVER JOB SERVER JOB SERVER

FIGURE 2: Job Queue Architecture

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 7

The job queue is provided as a Magento Enterprise Edition 2.0 platform feature that can be customized
and extended by the Magento ecosystem for tasks that require highly scalable processing. In subsequent
releases, more native features will make use of the job queue for operations such as sending emails,
indexing, and asynchronous order insertion.

An example of the job queue’s ability to improve efficiency and throughput in operations is deferred stock
updates, an optional configuration in Magento Enterprise Edition 2.0. Deferred stock updates allow for
inventory levels to update asynchronously as orders are placed to increases checkout throughput. It helps
ensure that orders will be captured when checkout transactions are high. This functionality is best suited for
high inventory items or products that can easily be backordered. Deferred stock updates can be enabled on
either a per-product or per-website scope, giving merchants the flexibility to use where ideal.

PHP Interpreters
Beyond the updates made in the Magento application, there are significant performance advancements
that have come from the larger PHP community. One such improvement is PHP7, a radically enhanced PHP
interpreter recently released by Zend that has been demonstrated to boost the performance of nearly all
PHP applications. Magento recommends and supports the use of PHP7 to achieve optimal performance
with the Magento 2.0 platform. Earlier versions of PHP may be used, but may not yield the same
performance benefits and are not recommended.

Database Improvements
To further improve scalability in our commercial product, Magento Enterprise Edition 2.0, we made a
number of improvements to the database tier of the application. These options allow for tuning and
optimization of the databases to better handle high traffic and transaction volumes.

Multiple Database Masters

One major enhancement is the ability to use different databases for different sub-systems or areas of
the application. This approach supports up to three different master databases for checkout, orders, and
product data that can be broken out in separate database instances. This separation of database instances
for different sub-systems effectively shards the Magento database by business entity. Each master database
can have multiple slave databases to further scale database read operations.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 8

 Desktop Browser Native Apps
Mobile Browser Enterprise Applications

FRAMEWORK

PRESENTATION API (REST & SOAP)

DIGITAL COMMERCE

IMPORT / EXPORT DATABASE PERSISTENCE WEB SERVICES MESSAGE BUS

MASTER DB MASTER DB MASTER DB

(MAIN) (CHECKOUT) (OMS)

SLAVE DB SLAVE DB SLAVE DB SLAVE DB SLAVE DB

(MAIN) (CHECKOUT) (OMS) (CATALOG) (WISHLIST)

FIGURE 3: Magento Database Tier with Multiple Master Databases

The division helps ensure that the load from merchandising and order management activities can be
isolated from users browsing and purchasing on the website. This separation allows different functional
areas to be scaled independently depending on the system load and unique needs of the business, such
as high order volume in checkout or a very high SKU count.

Magento uses the Command Query Responsibility Segregation (CQRS) database pattern to seamlessly
support the routing of queries to the appropriate database. Because of this, developers do not need
to know which database configuration is being used, which simplifies customization. CQRS assumes
a Create Read Update Delete (CRUD) database access pattern is in use. Our framework uses that
assumption to route the generated queries to read (slave) or write (master) databases. Developers
customizing Magento do not need to incorporate code to support different database configurations
as the framework will handle this automatically.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 9

MySQL Cluster Support
For situations where additional database scalability is needed, Magento Enterprise Edition 2.0 also supports
MySQL Cluster for checkout and order management databases. MySQL Cluster is a third-party offering
and provides a high-availability, multiple-master solution to scale MySQL. MySQL Cluster manages the
sharding of data across multiple database instances. Multiple master databases can be used in each of the
application domains as a result. This approach improves the write scalability of the application and enables
merchants to use MySQL Cluster to ease management of multiple master databases.

These database improvements represent a major advancement for Magento Enterprise Edition and provide
considerable flexibility in how the database tier can be scaled. Now each area of the application can be
tuned independently for its expected load. In addition, one can isolate the critical customer facing database
interactions from lower priority administrative tasks to achieve comprehensive scale for large and growing
enterprise customers.

Performance Toolkit
The Magento Performance Toolkit is a script and a set of JMeter tests that allow for consistent and
repeatable performance testing for Magento applications. The script allows the generation of four distinct
customer profiles that are intended to represent different eCommerce business sizes. The data from the
script populates a Magento instance. The Performance Toolkit is bundled with the Magento distribution
(in the setup/performance-toolkit directory).

Performance Results
To show the impact of Magento 2.0’s more flexible architecture and performance enhancements, a series
of benchmark tests were performed comparing the recommended configurations of Magento Enterprise
Edition 2.0 and Magento Enterprise Edition 1.14.2. These tests revealed that Magento Enterprise Edition
2.0, with its performance enhancements, tight integration with Varnish, and support for PHP7,
delivers higher throughput (orders/hour) and faster server response times across the board for
both small and large merchant deployments.

Test Configuration

Software
We tested the recommended and fully-supported out-of-the-box configurations for both products:

• Magento Enterprise Edition 2.0 with PHP7 and Varnish caching

• Magento Enterprise Edition 1.14.2 with PHP5.6 and Full-Page Caching

Testing of other scenarios is not valid as they do not provide data on recommended or supported
configurations. Additional software details are available in the Appendix.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 10

Scenarios
Two scenarios were investigated:

• Small merchant with $1-$5M in online sales deployed on a single 4-core web node and a single
database node. 25 simultaneous JMeter threads (representing 25 concurrent requests) were used to
show site performance during a period of peak traffic, such as a sale.
• Large merchant with $50-$100M in online sales deployed on five 4-core web nodes and a single
database node. 25 to 100 simultaneous JMeter threads (representing 25 to 100 concurrent requests)
were used to show site performance under increasing traffic loads.

Both scenarios assumed the following Magento store profile:

Websites / SKUs (Simple/ Categories / Catalog / Cart Customers Orders

Store Views Configurable) Nesting Rules (in Database) (in Database)

2/2 16,000 / 1,000 300 / 3 20 / 20 200 1600

Additional details on the scenarios are provided in the Appendix.

Testing Methodology
The tests emulated typical eCommerce site usage scenarios, such as:

• Catalog browsing, including visiting the home page, a catalog page, a configurable product page,
and a simple product page
• Adding both a simple and configurable product to the cart
• Completing the full checkout process (all steps) as a guest and a registered customer

We assumed the following site traffic patterns for user sessions:

• 62% browsing only

• 30% browsing and adding products to the cart, but abandoning
• 4% completing checkout as a guest
• 4% completing checkout as a registered customer

All tests were conducted using the Magento Performance Toolkit and JMeter was used to measure and
report server response times in milliseconds.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 11

Small Merchant Deployment

The small merchant test scenario shows that Magento Enterprise Edition 2.0 successfully runs on a small
4-core server, even when handling high loads associated with 25 concurrent users during a sales event.
Magento Enterprise Edition 2.0 delivers better performance than Magento Enterprise Edition 1.14.2 across all
use cases:

• Processes 28% more order per hour, reaching 597 orders per hour
• Delivers nearly instant response times for catalog pages
• Enables up to 66% faster add-to-cart server response times
• Provides 48% faster guest checkout response times and 36% faster customer checkout response
times when all checkout steps are combined

Throughput Results

Orders Per Hour

597
600

500 467
Orders per Hour

400

300

200

100

Magento EE 1.14.2 Magento EE 2.0

FIGURE 4: Throughput results for 25 JMeter threads show

that Magento Enterprise Edition 2.0 can process 597 orders
per hour, which is 28% more orders per hour than Magento
Enterprise Edition 1.14.2.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 12

Server Response Time Results

Catalog Browsing

50
Median Server Response Time (ms.)

41
40 37
35 35

30

20

10

1 1 1 1
0
Home Category Configurable Simple
Page Page Product Product

Magento EE 1.14.2 Magento EE 2.0

FIGURE 5: Catalog browsing server response times show the

Magento Enterprise Edition 2.0 integration with Varnish provides
nearly instant responses for cached content. It also generates
new pages up to 50% faster than Magento Enterprise Edition
1.14.2 for situation where there is a cache miss or cold cache.

Add to Cart Customer Checkout Operations (All Steps Combined)

2000 10000

8677
Median Server Response Time (ms.) - All Steps

1662
1533
Median Server Response Time (ms.)

8000
1500

6325
6000 5558

1000

656 4000
3303
523
500
2000

0 0
Simple Product Configurable Product Guest Checkout Customer Checkout

Magento EE 1.14.2 Magento EE 2.0 Magento EE 1.14.2 Magento EE 2.0

FIGURE 6: Server response times for adding products to FIGURE 7: Consolidated server response times for all steps of
the shopping cart are sub-700 milliseconds and up to 66% the checkout process show that Guest Checkouts are 48% faster
faster for Magento Enterprise Edition 2.0 when compared to and Customer Checkouts are 36% faster on Magento Enterprise
Magento Enterprise Edition 1.14.2. Edition 2.0 when compared to Magento Enterprise Edition 1.14.2.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 13

Large Merchant Deployment
Test results show that Magento Enterprise Edition 2.0 also delivers strong performance for large, high-
volume sites, and once again outperforms Magento Enterprise Edition 1.14.2 across all use cases, and under
increasing load. Results show that it:

• Processes 39% more order per hour, reaching up to 2,558 orders per hour
• Delivers nearly instant response times for catalog pages
• Enables up to 66% faster add-to-cart server response times that are under 500 milliseconds
• Provides 51% faster guest checkout response times and 36% faster customer checkout response
times for all checkout steps combined

Throughput Results

Orders Per Hour Orders Per Hour - Load Profile

3000 3000

2558
2500 2500

2000 1838 2000

Orders / Hour

Orders / Hour

1500 1500

1000 1000

500 500

0 0
25 Threads 50 Threads 75 Threads 100 Threads
Magento EE 1.14.2 Magento EE 2.0

Magento EE 1.14.2 Magento EE 2.0

FIGURE 8: Throughput results for 100 JMeter threads show FIGURE 9: Throughput data for a range of JMeter thread
that Magento Enterprise Edition 2.0 can process 2,558 counts show Magento Enterprise Edition 2.0 consistently
orders per hour, or 39% more orders per hour than Magento processes more orders per hour than Magento Enterprise
Enterprise Edition 1.14.2. Edition 1.14.2 across different load profiles, and is particularly
strong at the highest load level.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 14

Server Response Time Results

Catalog Browsing

Home Page Category Page

40 38 50

35 44
35 41
Median Server Response Time (ms.)

Median Server Response Time (ms.)

40
30 28 34

25 22 30
26
20

20
15

10
10

5
1 1 1 1 1 1 1 1
0 0
25 Threads 50 Threads 75 Threads 100 Threads 25 Threads 50 Threads 75 Threads 100 Threads

Magento EE 1.14.2 Magento EE 2.0 Magento EE 1.14.2 Magento EE 2.0

Configurable Product Simple Product

35 34 35 33
31 31
30 30
Median Server Response Time (ms.)

Median Server Response Time (ms.)

26 26
25 25
21
20
20 20

15 15

10 10

5 5

1 1 1 1 1 1 1 1
0 0
25 Threads 50 Threads 75 Threads 100 Threads 25 Threads 50 Threads 75 Threads 100 Threads

Magento EE 1.14.2 Magento EE 2.0 Magento EE 1.14.2 Magento EE 2.0

FIGURE 10: Benchmark data for catalog browsing pages show Magento Enterprise Edition 2.0 outperforms Magento Enterprise Edition
1.14.2 and delivers nearly instant server response times across all load profiles.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 15

Add to Cart

Simple Product Configurable Product

1189
1200 1137 1500
1380
994 1236
Median Server Response Time (ms.)

Median Server Response Time (ms.)

1000
1200
844 1076
800
896
900

600

600
397 402 486 497
400 345 418
303 366
300
200

0 0
25 Threads 50 Threads 75 Threads 100 Threads 25 Threads 50 Threads 75 Threads 100 Threads

Magento EE 1.14.2 Magento EE 2.0 Magento EE 1.14.2 Magento EE 2.0

FIGURE 11: Server response times for adding products to the cart on Magento Enterprise Edition 2.0 are up to 66% faster when compared
to Magento Enterprise Edition 1.14.2.

Customer Checkout Operations (All Steps)

Guest Checkout Customer Checkout

4979
5000 4805 8000
7086
4253 7000
6432
Median Server Response Time (ms.)

Median Server Response Time (ms.)

4000
3651 5933
6000
5141
5000 4696
3000 4516
2384 2434 3833
4000
2068 3336
2000 1850
3000

2000
1000

1000

0 0
25 Threads 50 Threads 75 Threads 100 Threads 25 Threads 50 Threads 75 Threads 100 Threads

Magento EE 1.14.2 Magento EE 2.0 Magento EE 1.14.2 Magento EE 2.0

FIGURE 12: Consolidated server response times for all steps of the checkout process show that Guest Checkouts are up to 51% faster and
Customer Checkouts are up to 35% faster on Magento Enterprise Edition 2.0 when compared to Magento Enterprise Edition 1.14.2.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 16

Conclusion
Magento 2.0 is a highly efficient eCommerce platform that performs better at scale compared to previous
Magento versions. Its top-tier performance and scalability, combined with new functionality, powerful
business user tools, and unprecedented flexibility to create tailored shopping experiences make it the next-
generation solution that is right for your business.

Powering Magento 2.0’s performance and scalability gains are a comprehensive set of enhancements that
optimize the client side to reduce page weight, tightly integrate Varnish caching to accelerate page load
times, and optimize the application itself to support more efficient job processing and additional concurrent
Admin users. It also incorporates new technologies, like PHP7, to boost performance and includes a re-
architected database tier with support for multiple master databases to offer a better ability to scale and
manage unique system load and business requirements.

Together, these enhancements result in a solution that sets a new standard for performance and scalability.
Benchmark tests show that Magento Enterprise Edition 2.0 delivers across-the-board improvements in
server response times for both small and large merchants. Catalog browsing, add-to-cart operations, and
checkouts are significantly faster and Magento Enterprise Edition 2.0 can process up to 39% more orders
per hour, enabling you to get more from your hardware investment. The chart below summarizes all key
benchmark test results so you can see for yourself the power of the Magento 2.0 platform.

Small Merchant Deployment

Throughput – orders per hour

Magento Enterprise Edition 1.14.2 Magento Enterprise Edition 2.0

467 597

Server Response Time - milliseconds

Magento Enterprise Edition 1.14.2 Magento Enterprise Edition 2.0

Catalog Browsing
Home Page 37 1
Category Page 41 1
Configurable Product 35 1
Simple Product 35 1

Add to Cart
Simple Product 1,533 523
Configurable Product 1,662 656

Checkout (All Steps)

Guest Checkout 6,325 3,303
Customer Checkout 8,677 5,558

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 17

Large Merchant Deployment

Throughput – orders per hour

Magento Enterprise Edition 1.14.2 Magento Enterprise Edition 2.0

25 Threads 50 Threads 75 Threads 100 Threads 25 Threads 50 Threads 75 Threads 100 Threads

919 1,531 1,930 1,838 1,179 1,930 2,351 2,558

Server Response Time - milliseconds

Magento Enterprise Edition 1.14.2 Magento Enterprise Edition 2.0

25 50 75 100 25 50 75 100
Threads Threads Threads Threads Threads Threads Threads Threads

Catalog Browsing
Home Page 22 28 35 38 1 1 1 1
Category Page 26 34 41 44 1 1 1 1
Configurable Product 21 26 31 34 1 1 1 1
Simple Product 20 26 31 33 1 1 1 1

Add to Cart
Simple Product 844 994 1,137 1,189 303 345 397 402
Configurable Product 896 1,076 1,236 1,380 366 418 486 497

Checkout (All Steps)

Guest Checkout 3,651 4,253 4,805 4,979 1,820 2,068 2,384 2,434
Customer Checkout 5,141 5,933 6,432 7,086 3,336 3,833 4,516 4,696

Learn More about Magento 2.0

If you are an existing Magento customer, feel free to reach out to your account manager to learn more
about the Magento 2.0 platform. Prospective merchants can speak with a Magento Specialist by calling
1-877-574-5093 or visiting magento.com/explore/contact-sales.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 18

Appendix
Merchant Profile Details

Websites / SKUs (Simple/ Categories / Catalog / Cart Customers Orders

Store Views Configurable) Nesting Rules (in Database) (in Database)

1/1 16,000 / 1,000 300 / 3 20 / 20 200 1600

Definitions

• Websites: The number of websites on the store.

• Store views: The number of store views in the system total.
• Categories: The number of categories in the store.
• Nesting: How many nested layers of categories are in the store.
• Catalog / Cart Rules: The number of rules in each category in the system.
• Customers (in DB): The number of registered customers in the system before the test starts.
• Orders (in DB): The number of order existing in the system before the test starts.

Small Merchant Deployment (4 web node cores)

• Software
- PhP 5.6.13 with Zend Opcache v7.0.6-dev (for Magento Enterprise Edition 1.14.2)
- PhP 7.0.3-1 with Opcache (for Magento Enterprise Edition 2.0)
- Nginx 1.6.2
- MySQL 5.6.28
- Varnish 4.0.2 (for Magento Enterprise Edition 2.0)
- Redis 3.0.5
- CentOS 6.4, CentOS 7.0 and Debian (on core 3.16.7)

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 19

 • 1 Web Node
- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks
- Running Software
- Nginx
- Varnish (for Magento Enterprise Edition 2.0)
- php-fpm

• 1 Database Node
- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks
- Running Software
- MySQL
- Redis (FPC Cache for Magento Enterprise Edition 1.14.2, Session Storage)

• 1 JMeter node (to drive the tests)

- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks

• All performance tests leverage real cores. Virtual can bring unexpectedly high deviation.

Larger Merchant Deployment (20 web node cores)

• Software
- PhP 5.6.13 with Zend Opcache v7.0.6-dev (for Magento Enterprise Edition 1.14.2)
- PhP 7.0.3-1 with Opcache (for Magento Enterprise Edition 2.0)
- Nginx 1.6.2
- MySQL 5.6.28
- Varnish 4.0.2 (for Magento Enterprise Edition 2.0)
- Redis 3.0.5
- Memcache 1.4.21
- CentOS 6.4, CentOS 7.0 and Debian (on core 3.16.7)

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 20

 • 5 Web Nodes
- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks
- Running Software
- Web Node #1:
- Nginx
- Varnish (for Magento Enterprise Edition 2.0)
- php-fpm
- Web Node #2 - #5:
- php-fpm

• 1 Database Node
- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks
- Running Software
- MySQL
- Redis (Session Storage, FPC storage for Magento Enterprise Edition 1.14.2)

• JMeter node (to drive the tests)

- 4 CPUs Core i7 with hyper threading
- 8GB of memory
- 500 GB 7200 RPM SSHD and RAID 1, 5, or 10 disks

• All performance tests leverage real cores. Virtual can bring unexpectedly high deviation.

MAGENTO 2.0 SITE PERFORMANCE AND SCALABILITY OPTIMIZATIONS | PAGE 21