TABLE OF CONTENTS

3 Executive Summary

4 Retail and eCommerce companies embrace NoSQL to improve customer experience,

efficiency and agility

4 Database performance and availability are key to delivering great customer experiences

5 NoSQL distributed databases provide better performance and availability than relational

6 NoSQL delivers faster, easier, more affordable scalability

6 Single-node type simplifies scaling

6 Bidirectional replication improves availability and data locality

7 Couchbase combines single node type and bidirectional replication

7 NoSQL drives increased operational efficiency: lower costs and faster time to market

7 Elastic scaling optimizes hardware usage

8 Data flexibility speeds innovation and time to market

8 NoSQL can support numerous retail and eCommerce use cases

9 NoSQL is well suited to many data types

10 Use Case Spotlight: Product Catalog

10 Product catalogs are challenging for relational databases

10 Product catalog data modeling in a NoSQL database

12 Retail and eCommerce data modeling in a NoSQL database

13 How to introduce a NoSQL database into a relational environment

14 NoSQL in a microservices architecture

14 Caching with NoSQL

15 As retail and eCommerce go mobile, NoSQL can deliver a better customer experience, online
and in-store

15 Why Couchbase is a great NoSQL solution for retail and eCommerce

16 Conclusion: NoSQL is the right choice for many retail and eCommerce use cases
 NoSQL for Retail and eCommerce

Executive Summary

Digital Economy companies – including retail and eCommerce leaders like Walmart, Tesco, eBay,
Fanatics, StubHub, Staples, cars.com and hundreds more – are embracing NoSQL database tech-
nology to build and run modern web and mobile applications. Why? Because NoSQL is better than
relational technology in meeting the performance, scalability, availability, agility, and affordability
requirements of these applications.

Rising customer Rising customer expectations and competitive pressures are driving the need for NoSQL. Com-
expectations and panies face increasing demands to deliver great customer experiences – i.e., fast, personalized,
competitive pressures context- and location-aware. At the same time, they have to manage growing volumes of users and
are driving the need data, while reducing costs and time to market. These pressures create a new set of requirements
for NoSQL. for the operational database. It must:

nn Service data requests with submillisecond latency

nn Scale to meet peak demand (e.g. Black Friday), easily and affordably
nn Provide 24x7x365 availability
nn Easily accommodate evolving data types and queries
nn Ingest, aggregate, and store multi-structured data from many sources
nn Support multiple channels and devices
nn Replicate data across data centers globally
nn Integrate with other big data tools like Hadoop, Spark, Kafka and others
nn Accelerate and simplify development

The operational database must also be versatile to support many use cases – e.g., customer profile
management; shopping cart; session store; product and pricing catalog; 360-degree customer
view and loyalty program management, to name just a few.

This paper provides an overview for CIOs, CTOs, architects, developers, and operations engineers
interested in building and supporting modern applications on NoSQL:
nn Why relational database technology (Oracle, SQL Server, MySQL, PostgreSQL, etc.) cannot
meet the new requirements for web and mobile applications
nn How NoSQL – specifically Couchbase, an open source document database – provides a
better solution
nn Examples of retail and eCommerce use cases supported by NoSQL, showing how data can be
modeled and queried with Couchbase.

We conclude with a section on how to introduce NoSQL into a relational environment,

emphasizing two approaches:

nn NoSQL in a microservices architecture

nn NoSQL for high-performance caching
 Retail and eCommerce companies embrace NoSQL to improve customer
experience, efficiency and agility

In 2016, 89% of In the Digital Economy, retail and eCommerce companies compete based on customer experience
companies expect to and operational efficiency – customers expect personalized engagements, relevant information,
compete based on and instant gratification, but their expectations have to be met while reducing operational costs
customer experience. and time to market. In the Digital Economy, retail and eCommerce companies win by creating
loyal customers who spend more, but cost less.
(Source: Gartner Survey)
Companies not only face competition from within their industry – whether it’s apparel, automo-
tive, electronics, office supplies, or other products – but also from Internet companies like Ama-
zon and eBay. These digital-born, tech-savvy companies are renowned for how they continuously
improve their eCommerce applications, and for the speed and agility with which they release new
features.

A recent CIO article For architects, developers, and operations engineers, the pressure to deliver great customer
on ways to improve experiences, reduce costs, and accelerate innovation has never been greater. As a result, across
the customer all retail and eCommerce categories, from cars to clothes to office supplies and more, innovative
experience for online companies – companies like cars.com, Fanatics, StubHub, Best Buy, Foot Locker, Nike, Staples,
shoppers highlights Tesco, Walmart, and hundreds more – are embracing NoSQL to improve customer experience,
the importance of operational efficiency, and agility.
site performance,
customer reviews, stock NoSQL is a modern database technology developed about a decade ago by leading Internet com-
availability, live chat, panies including Google, Facebook, Amazon and LinkedIn. It was created to overcome the limits of
the checkout process, relational databases in meeting the requirements of modern web, mobile and IoT applications.
shopping carts,
and more.
Database performance and availability are key to delivering great
(Source: CIO) customer experiences
When it comes to customer experience, performance and availability are critical. Whether custom-
ers know it or not, they’re interacting with a database. They’re accessing product data, customer
data, engagement data – and if the data is not readily available, the customer’s experience suffers.

Shoppers expect every request, whether it’s finding and viewing a product, adding it to the
Walmart saw a sharp shopping cart, or proceeding to checkout, to be handled immediately – not within seconds, but
decline in their milliseconds. It doesn’t matter what time it is, or what time zone they’re in. It doesn’t matter if it’s
conversion rate Black Friday, Cyber Monday, or Super Sunday – shoppers expect eCommerce applications, web and
when page load time mobile, to be available 24 hours a day, 365 days a year, anywhere in the world, and on any device.
increased from one to
four seconds. However, Historically, relational databases have been the bottleneck, as DBAs and developers try to squeeze
for every one second out every last drop of performance and work around the clock to maintain availability. However,
of improvement, with an ever-growing number of customers and rising expectations, they’re fighting an uphill
their conversion rate battle.
increased by 2%.
It’s one reason why retail and eCommerce companies are embracing NoSQL: They require a da-
(Source: Walmart Labs) tabase that’s engineered for a higher level of performance and availability – capable of meeting
the expectations of an online, 24/7, global customer base.

4
 NoSQL distributed databases provide better performance and availability
than relational

Applications with Typically, relational databases run on a single server, so the resources available to the database – pro-
thousands, if not cessing, storage, and memory – are not only limited, they’re fixed. This was not a problem when
millions, of web and relational databases powered internal-facing applications with a limited number of users and predict-
mobile customers can able workloads. However, applications with thousands, if not millions, of web and mobile customers
overwhelm relational can overwhelm relational databases with increasing, and more importantly, volatile workloads.
databases with
increasing, and more
importantly, volatile
workloads.

Figure 1: Relational databases suffer from under/over-provisioning, leading to performance and cost issues.

The only way to maintain performance as the number of customers and products grows, especially
during peak periods, is to add more resources – often on demand. With relational databases, this is
both difficult and costly, because you have to replace the RDBMS server with a bigger, more expensive
server. But with many NoSQL databases, it’s simply a matter of adding more servers to scale out. That’s
because they often run on a cluster of servers – i.e., they’re distributed databases.
SERVER LOAD

SERVER CAPACITY
POSITIVE USER EXPERIENCE
(Acceptable Server Performance)

INCREASE OF USERS OVER TIME – COUCHBASE

Figure 2: NoSQL databases can increase capacity in small increments for more efficient use of hardware.

5
 The same is true for availability. If a relational database is running on a single server, and that server
fails, the data becomes unavailable. NoSQL databases, when running on a cluster of servers, not only
tolerate hardware failure, they expect it. That’s why the data is replicated to multiple nodes. If one fails,
the data remains available on the others.

Figure 3: Relational databases suffer from a single point of failure, but NoSQL databases do not.

NoSQL delivers faster, easier, more affordable scalability

For a NoSQL database to effectively scale, it requires two things: a flat topology and bidirectional
replication across data centers.

Single-node type simplifies scaling

It’s not enough for a NoSQL database to be able to scale – it also has to scale on demand and with ease.
This can be difficult for NoSQL databases that rely on complex topologies with multiple moving parts,
like master-slave or other specialized node type topologies, for example. Such complexity introduces
barriers to easy scaling, because it requires careful configuration of different nodes with different roles
and different relationships.

Rather, every node should be capable of performing the same roles – a single node type. A flat topology
with a single node type is easy to scale because it requires little more than adding a node to the cluster.

Figure 4: It’s easy to scale with a single-node architecture - start more nodes, add them to the cluster.

Bidirectional replication improves availability and data locality

In order to provide Bidirectional replication between data centers is also essential, because customers are often not
the highest level confined to a single geography; they can be in different states, countries, or regions. In order to provide
of performance the highest level of performance and availability to a geographically distributed customer base, the
and availability to database must also be geographically distributed.
a geographically
distributed customer However, being geographically distributed isn’t enough: the database must be deployed in an
base, the database must active-active configuration using bidirectional replication. This enables eCommerce applications to
also be geographically read and write to the data center closest to the customer – providing local data access for the best
distributed. possible performance. For example, the shopping carts of U.S. customers should be updated on a North
American data center while the carts of UK customers should be updated on an EMEA data center. It
would not make sense for all customer carts to be updated on a single, possibly geographically distant,
higher latency, data center.
6
 In addition, with active-active configurations, an entire
data center can fail without interrupting availability.
That’s because applications can simply write to a
different data center, or requests can be automatically
rerouted by routers and load balancers. For example,
if the North American data center failed, U.S. custom-
ers would access their online shopping carts from the
EMEA data center – and they wouldn’t even know.

Providing this kind of seamless failover without

interrupting the customer experience is a challenge
for NoSQL databases that rely on unidirectional Figure 5: Deploy to multiple data centers in an active/
replication (master-slave) between data centers. They active configuration with bidirectional replication.
can provide local access to mostly read-only data, like
a product catalog, but they can’t do it for things like shopping carts, customer profiles, and inventory –
i.e., data services that may be location specific. In addition, because writes can only be handled by the
data center with the primary node, applications have to wait for a new primary to be elected should the
original fail – resulting in temporary loss of availability, potentially on a global scale.

Couchbase combines single node type and bidirectional replication

Among leading NoSQL databases, Couchbase stands out as the only one that combines a single-node
architecture (i.e., flat topology) and bidirectional replication between data centers, to deliver fast and
easy scalability, along with high performance and availability.

NoSQL drives increased operational efficiency: lower costs and faster time to market

In addition to improving the customer experience with greater performance and availability, NoSQL
databases improve operational efficiency by reducing costs and enabling a faster time to market.

Elastic scaling optimizes hardware usage

Database workloads for retail and eCommerce companies can be highly volatile and unpredictable –
spikes can occur due to events like Black Friday and Cyber Monday, as well as special promotions like
flash sales or when an item goes viral. As a result, it can be challenging to maintain operational efficien-
cy. When a relational database server is configured for peak workloads, and it’s not operating under
one, then it’s over-provisioned.

Capacity can be adjusted Hardware, like any other resource, must be elastic – it must be possible to use more or less of it de-
simply by increasing or pending on demand. NoSQL databases can both scale out and scale in. Capacity can be adjusted simply
decreasing the number by increasing or decreasing the number of nodes in a NoSQL database cluster, depending on workload.
of nodes in a NoSQL With a NoSQL database, it’s possible to always use only as much hardware as needed – never too
database cluster, much, never too little. Because hardware is always optimized for the current workload, NoSQL is more
depending on workload. cost efficient than relational.

7 Figure 6: Relational databases with fixed capacities cannot support by volatile workloads with high peaks.
 Data flexibility speeds innovation and time to market

Innovation is key to competitive success – for retail and eCommerce companies, that means a steady
flow of new products, new offers, new programs, and more. And when it comes to innovation, time to
market is critical. Relational databases, and wide-column stores, limit the pace of development due to
their static schemas. Adding a new feature often requires changing the database schema, which in turn
requires planning and coordination – both time-consuming tasks.

However, NoSQL databases – in particular key-value stores and document databases – leverage
dynamic schemas. Rather than defining a static schema, they defer to the application. As a result,
developers can create new features without having to wait for a schema change. When the database
enables this kind of agile development instead of hindering it, new features can released far more
quickly, resulting in faster time to market for retail and eCommerce companies.

NoSQL can support numerous retail and eCommerce use cases

Innovative retail and eCommerce companies are successfully leveraging NoSQL for use cases that involve
products, customers, and interactions – for everything from product and pricing catalogs, to content
management, to personalization to customer profiles, and more. All of these use cases benefit from one
or more of the advantages of NoSQL: dynamic schema, low latency throughput, scalability and availability.

For example, when stored in a relational database, a product catalog requires a complex schema across
multiple tables and complex queries to access it. However, when stored in a NoSQL document data-
base, all of the information for a single product can be stored together in a single JSON document –
making it much easier and faster to access.

Ingesting clickstream When managing millions of customers, storing customer profiles and all of their interactions requires
and session data a database that can easily and affordably scale. It’s a matter of volume, as well as velocity. Ingesting
to power real-time clickstream and session data to power real-time recommendation engines and create personalized
recommendation experiences requires a database capable of sustaining high throughput while maintaining low latency.
engines and create
personalized experiences Below are some of the common retail and eCommerce application use cases where companies are
requires a database leveraging NoSQL for greater performance, availability, scalability, and agility:
capable of sustaining
high throughput while
maintaining low latency. Unified Customer Profile Management
Store and aggregate customer profile data to enable multiple services (e.g., authentication, personal-
ization). Customer profiles include data that’s local to the application/customer (demographic profile,
behavior, history, location, preferences, etc.) as well as data aggregated from multiple internal and
external sources and channels (financial profile, credit history, purchasing history, etc.), ultimately
providing a 360-degree view of the customer.

Omnichannel Shopping Services

Provide customer-centric, personalized, low latency, scalable data services to support seamless,
always-on shopping experiences for in-store, kiosk, mobile and web-based shopping. The shopping
service must provide a personalized experience and be able to scale to handle peak internet and mobile
device usage of the service.

Shopping Cart Session Management

Reliably manage real-time online shopping cart session data in memory for low latency performance.
Allow customers to view and modify the contents of their online shopping cart, as well as finalize their
purchases (ie, checkout) securely, quickly and easily.

Order & Fulfillment Management

Capture orders and generate appropriate fulfillment data in order to accelerate order processing, tracking
and delivery to the customer. Applications typically include flexible business rules that manage how
orders are fulfilled in order to minimize cost and enhance the customer experience. Amazon Prime and
8 other premium fulfillment services are great examples of this customer and cost-driven fulfillment model.
 Product Catalog Management
Scalably manage a unified repository of all product data, including product SKUs, descriptions (in
multiple languages), images, ratings, reviews, etc. The product repository may also include product
relationships or product “baskets” for products which are often purchased together, as well as com-
petitive product information. Schema flexibility, rich data modeling abilities and performance are key
requirements to provide enhanced product-customer interaction.

Price Catalog Management

Provide a low-latency repository of pricing data per product, including static as well as dynamic pricing
rules (e.g., by geography, customer status, date & time, promotions, discounts, etc.) which need to be
applied in order to generate a real-time, customer specific price. In retail and eCommerce applications
pricing is never static – complex business rules drive optimized pricing. Customer expectations require
that dynamic pricing applications return results in milliseconds.

Inventory & Availability Management

Manage and maintain inventory data to enable real-time inventory checks and updates. Inventory
must be tracked both locally within a store, regionally, as well as globally, in order to support services
such as in-store order fulfillment and lowest-cost, shortest-time delivery options. When availability
falls below business-rule specified levels, the application should trigger actions in downstream legacy
systems, such as Supply Chain Management (SCM) for example to trigger automatic re-stocking.

Personalization
Utilize market segmentation data, business rules and other relevant data in combination with the
customer profile and product catalog repositories to feed a personalization / recommendation engine.
Pre-generated as well as real-time recommendations are often combined in order to provide a seam-
less, highly responsive, personalized customer experience.

Loyalty Program Management

Aggregate and store relevant customer purchase history, loyalty status, and rewards activity (e.g.,
points, redemptions, discounts, promotions), and maintain associated flexible loyalty program business
rules and policies. Design flexibility is a key requirement in this use case because each Loyalty Program
is different. How loyal customers are rewarded is a key apect of long term customer retention for retail
and eCommerce companies.

NoSQL is well suited to many data types

NoSQL document NoSQL document databases are well suited to many types of data – making them very effective as
databases are well a general-purpose database for retail and eCommerce applications. They are ideal for product, click-
suited to many types of stream, session, order, and user / member data – all part of today’s customer experience, whether
data – making them very online or in-store.
effective as a general-
purpose database for
retail and eCommerce
applications.

Figure 7: The types of uses cases and data that retail / eCommerce companies use NoSQL databases for.

9
 Use Case Spotlight: Product Catalog

As a concrete example of how NoSQL can be a better fit for retail and eCommerce applications than
relational databases, let’s consider one of the most common use cases: Product Catalog.

Product catalogs are challenging for relational databases

Products can be very complex. In addition to a name and base price, a product can have significantly
more data associated with it. Products can belong to multiple categories. Product data can include nu-
merous features, specifications, descriptions and images. Data can include reviews, ratings, and other
third-party generated content.

There are several ways to model

product catalog data with a relational
database. However, they’re simply
workarounds to get past limitations of
relational modeling. No one approach
is ideal. They all introduce new prob-
lems and challenges for developers.
(see table, right)

Product catalogs can easily be modeled with a NoSQL database

It’s much easier to model It’s much easier to model product catalogs with a NoSQL document database, because all data for
product catalogs with a single product can be stored together. It can all be stored in a single document instead of multiple
a NoSQL document rows, often in multiple tables. Not only is it easier to model the data, it’s simpler and faster to access –
database, because all there’s no need to perform a query with multiple joins or to pivot the results.
data for a single product
can be stored together.

10
Uniquely with Couchbase, Uniquely with Couchbase, product data can be queried via a SQL-based query language, N1QL
product data can be (“nickel”), or via simple key-value operations. In the following examples, the results are displayed
queried via a SQL-based as tables for simplicity. However, within applications, the results can be JSON documents,
query language, N1QL providing much more flexibility – especially when querying multiple product types.
(“nickel”), or via simple
key-value operations.
Example One: List all products by title and price

Example Two: Find all movies under $10

Example Three: Find all laserdisc movies

Example Four: Find all science fiction books

11
 All sorts of retail and eCommerce data can easily be modeled in a NoSQL database

It’s easy to model other types of data with NoSQL databases, too – shopping carts, product reviews,
inventory, and more. As with product catalogs, this data is easier and faster to access with a NoSQL
database.

Document databases In addition, document databases provide a great deal of flexibility by supporting nested elements like
provide a great deal of arrays and objects. For example, product reviews – they could be embedded within the product doc-
flexibility by supporting ument, stored as individual documents, or stored as a group of documents. In the example below, ten
nested elements like product reviews are stored per document using an array – making it easy and fast to paginate through
arrays and objects. them, ten at a time.

The following examples highlight the subdocument API in Couchbase Server. It enables applications
to read or write specific fields within a document for maximum performance. This includes updating
fields, incrementing counters, adding elements to arrays, and more.

Example One: Add two copies of Chasm City to the cart instead of one

Example Two: Add a copy of Neuromancer to the cart

Example Three: Add a review to Chasm City

12
 Example Four: Get the first ten reviews of 2001: A Space Odyssey on Blu-ray

Modeling inventory data

In addition to product catalogs, modeling inventory data with a document database is easy
and flexible, too. Below are two examples. The first models inventory by product, the second by
instance. For example, when there is a fixed number of instances (e.g. event tickets), you can store
each instance as a separate document.

Example One: Update the inventory, 100 copies of Chasm City (softcover) have arrived

Example Two: Change the inventory status, set 12-102 for the Super Bowl has been purchased

How to introduce a NoSQL database into a relational environment

Most innovative retail and NoSQL document databases are excellent general-purpose databases for retail and eCommerce
eCommerce companies applications. However, that doesn’t mean there isn’t a place for relational databases. In particular,
have embraced NoSQL by relational databases are well suited to legacy business management applications, such as enter-
successfully introducing prise resource planning (ERP) and supply chain management (SCM). Whereas these are employ-
it into their relational ee-facing applications, NoSQL is better suited to meet the performance, scalability, availability,
environments. and agility requirements of interactive customer-facing applications.

In fact, the most innovative retail and eCommerce companies have embraced NoSQL by suc-
cessfully introducing it into their relational environments. It’s a matter of architecture, and there
are two popular approaches to adding NoSQL into an existing relational database environment:
microservices and caching.
13
 NoSQL in a microservices architecture

Most successful retail and eCommerce applications are deployed using a microservices
architecture. In a microservices architecture, applications are deployed as a set of independent,
full-stack services.

An important concept of microservices architectures is that of decentralized data management.

Where a monolithic application would store all data in a single database, a set of microservices
can and should store their own data in their own databases. This often leads to polyglot per-
sistence – different services using different types of databases. While some services may continue
to use a relational database, others may use NoSQL databases – some may use a document data-
base, others may use a graph database.

With a microservices
architecture, eCommerce
companies can choose
when and where to
introduce a NoSQL Figure 8: A microservices architecture supports different services using different types of databases.
database, enabling them
to migrate to NoSQL one With a microservices architecture, eCommerce companies can choose when and where to introduce a
service at a time. NoSQL database, enabling them to migrate to NoSQL one service at a time.

Caching with NoSQL

In some environments, it may be better to augment an

existing relational database with a NoSQL database. In
this architecture, data is loaded into a NoSQL database
that is deployed between the application and the rela-
tional database. In a sense, the NoSQL database func-
tions like a cache. For example, the product catalog can
be exported from the relational database to the NoSQL
database. When the data changes in the relational data-
base – for example, the addition of new products – the
NoSQL database is updated.

However, the data doesn’t have to be read-only – inven-

tory data, for example. The relational database may be
the primary source of record, but retail and eCommerce
applications will interact with the NoSQL database for
faster performance. While it’s possible for the inventory
to become out of date for a brief time – until the relation- Figure 9: Deploy a NoSQL database on top
al and NoSQL databases are synchronized – the benefits of a relational database to improve
of better performance, and thus a better customer performance and reduce costs.
experience, far outweigh the drawbacks.

Applications can leverage Database Change Protocol (DCP) in Couchbase Server to tap into a change
stream, and apply these changes to the underlying relational database. In the reverse direction, with
Oracle Database for example, applications can use Oracle GoldenGate to capture changes and apply
them to Couchbase Server. Finally, applications can use the Couchbase Kakfa Connector to stream
changes between Couchbase Server and a relational database in both directions.

In addition to providing better performance and availability, adding a NoSQL database can reduce the
costs associated with maintaining relational databases. For example, a NoSQL database can be de-
ployed to lessen the load on a mainframe – reducing MIPS and thus costs – or to avoid having to deploy
the relational database on bigger hardware, and thereby avoid higher licensing and maintenance costs.
14
 As retail and eCommerce go mobile, NoSQL can deliver a better customer
experience, online and in-store
Every year, more and more customers are shopping online via mobile devices. According to a BI
Intelligence report, 20% of all eCommerce will be mobile in 2016, 45% by 2020 (up to $284 billion
in sales). However, many retail and eCommerce companies struggle with the mobile experience,
leading to lower conversion rates.

To create a mobile experience that is on par, or exceeds, the web experience requires a native mobile
solution. A key component of that solution is an embedded database for mobile devices, providing
mobile apps with local data access. It not only improves performance, because the data is local – it
also improves availability, because an embedded database ensures the app will always work whether
or not there’s an active network connection. Furthermore, the solution must provide built-in syn-
chronization – enabling the database to push data to the consumer, and vice-versa.

A native mobile solution can be used to improve both online and in-store experiences. For example,
relevant coupons and rewards can be pushed to customers’ mobile phones while they’re shopping
– even using beacons to detect what department they’re in to increase relevancy and thus conver-
sion. Customers can add items to a virtual shopping cart or browse their wishlist – perhaps receive
a notification when something is in stock or on sale. By leveraging the full power of a native mobile
solution, there are many possibilities for improving the customer experience, online or in-store.

Couchbase Mobile is the Couchbase Mobile is the only native, NoSQL-based solution for mobile apps. It is comprised of
only native, NoSQL-based Couchbase Lite and Couchbase Sync Gateway, and when paired with Couchbase Server, it’s a com-
solution for mobile apps. plete platform for mobile apps. In addition, it’s a cross-platform solution that is available for iOS,
Android, .NET/Java, and Mac OS X.

Why Couchbase is a great NoSQL solution for retail and eCommerce

Couchbase is a powerful NoSQL solution for a broad range of use cases, delivering the high performance,
scalability, availability and agility that today’s retail and eCommerce applications require. Numerous
retail and eCommerce companies – including Walmart, eBay, Office Depot, Tesco, Fanatics, Nu Skin, and
many more – have chosen Couchbase for several key advantages.

Memory-centric architecture
Couchbase takes full advantage of all available memory to give your application the sub-millisecond
responsiveness that today’s shoppers expect.

Integrated cache
While other NoSQL databases like MongoDB require a third-party cache – adding to both cost and
complexity – Couchbase has a fully integrated cache that delivers blazing performance. No need for a
separate product to install and manage.

Powerful, SQL-based query language

Unique among all NoSQL document databases, Couchbase provides N1QL (“nickel”) – a powerful query
language that lets developers easily query JSON data using familiar, SQL-like expressions.

Built-in high availability and disaster recovery

Couchbase comes with high availability within a cluster and provides market-leading cross datacenter
replication (XDCR) capabilities to support DR and data locality requirements. No need for complicat-
ed third-party systems. You have full control over the topology – unidirectional, bidirectional or any
configuration you need.

15
 Complete, GUI-based admin console
Among NoSQL document databases, only Couchbase provides a fully integrated GUI-based manage-
ment console, complete with hundreds of pre-built metrics and easy to use tools like push-button
scaling, rebalancing, and memory tuning.

Always-on mobile support

Couchbase Mobile is a complete NoSQL solution for mobile application support. It includes Couchbase
Lite – an embedded JSON database for devices – and Sync Gateway, a pre-built solution that syncs the
device with the cloud. Couchbase Mobile lets you easily support use cases such as in-store personal-
ized apps, point of sales systems, and mobile-optimized digital catalogs.

Conclusion: NoSQL is the right choice for many retail and eCommerce use cases

If you’re running into If you’re running into scalability, performance, or availability challenges with your relational database
scalability, performance, – and you’re looking to speed development and innovation – it’s probably time to consider NoSQL
or availability challenges technology.
with your relational
database – and you’re Hundreds of leading retail and eCommerce companies have already made the move, deploying NoSQL to
looking to speed support a growing number of use cases from customer profile management personalization, to product
development and catalogs, shopping carts, 360-degree customer view, and many more.
innovation – it’s probably
time to consider NoSQL
technology. Getting started is easy

Download the software and install it in a non-production corner of your IT environment. Couch-
base Server is available for Microsoft Windows, Apple OS X, and leading Linux platforms such as
Red Hat, Debian, and CentOS.

Explore our sample travel reservation app. A ready-made testing environment is waiting for your
input, or your can install it on your own machines and get intimate with the app’s source code.

Talk to a Couchbase Solutions Engineer about your retail or eCommerce use case. We have ded-
icated sales and support offices in North America, Europe, and Asia Pacific, and trusted partners
all over the globe.

About Couchbase

Couchbase delivers the world’s highest performing NoSQL distributed database platform. Developers around the world use the
Couchbase platform to build enterprise web, mobile, and IoT applications that support massive data volumes in real time. The
2440 West El Camino Real | Ste 600 Couchbase platform includes Couchbase Server, Couchbase Lite - the first mobile NoSQL database, and Couchbase Sync Gate-
Mountain View, California 94040 way. Couchbase is designed for global deployments, with configurable cross data center replication to increase data locality and
availability. All Couchbase products are open source projects. Couchbase customers include industry leaders like AOL, AT&T,
1-650-417-7500 Bally’s, Beats Music, BSkyB, Cisco, Comcast, Concur, Disney, eBay, KDDI, Nordstorm, Neiman Marcus, Orbitz, PayPal, Rakuten /
Viber, Tencent, Verizon, Wells Fargo, Willis Group, as well as hundreds of other household names. Couchbase investors include
www.couchbase.com Accel Partners, Adams Street Partners, Ignition Partners, Mayfield Fund, North Bridge Venture Partners, and West Summit.