Download the magic number 5

THE MAGIC NUMBER 5
Why Cloud Applications Require an
Always-On Data Platform
The Magic Number 5
Welcome to the Cloud
It’s 2017; the iPhone is ten years old, as is Amazon Web Services. The past
decade has been a blur of constant and rapid technology evolution fueled by a
dizzying array of applications designed to make our lives more convenient.
Whether it’s for buying groceries or selling software, we expect these applications
to perform a certain way all the time, and when they don’t, we get frustrated.
Welcome to the era of cloud applications.
In this eBook, we’ll take a look at how we got here, why data plays such an
important role in making cloud applications work correctly, and why DataStax is
the ideal solution for powerful applications that are contextual, always on, real
time, distributed, and scalable.
How Cloud Applications Were Born
Like any “revolutionary” technology, the advent of cloud applications really comes
down to a series of prior inventions:
•
•
•
The invention of computers in the 1940s
The invention of the personal computer by IBM in the 1980s
The invention of the world wide web by English scientist Tim Berners-Lee
in 1989
The invention of “Software as a Service” by Salesforce in 1999
The invention of Amazon’s Elastic Compute Cloud (EC2) web service in 2006
And the invention of the iPhone in 2007
•
•
•
About 35 years ago PCs exploded onto the scene and shortly thereafter started to
connect to the newly invented world wide web, making advances like web-based
email possible.
These internet-based applications were the precursors of today’s cloud
applications, and their architecture used the “client-server” model:
1. Client – usually a personal computer
2. Server – usually a high-powered
network computer
3. Network – whatever is needed to connect
the client to the server
2
The Magic Number 5
In 1999, Salesforce introduced the first true “Software as a Service” (SaaS) in the form of an enterprise CRM, which allowed
companies to use software on a subscription basis rather than having to pay costly software licenses and annual
maintenance fees.
In 2006 Amazon introduced the Elastic Compute Cloud (EC2) – a service designed to make web-scale cloud computing
easy for developers. With EC2, developers could use Amazon’s infrastructure to build applications, thus creating the first
“public” cloud.
Just one year later Apple gave us a new and ubiquitous way to connect to applications: the iPhone. Smartphones evolved
quickly and soon enough you could use them to take, upload and share photos, watch movies, shoot movies, stream music,
make music, do your expenses, and access your CRM, among many other things, all while sitting on a beach.
Users could access their applications from anywhere and were using these applications for everything.
And thus, cloud applications were born.
The Data Layer and Why It’s Important for Cloud Applications
During the client-server era, as with now, people (and machines) needed a way to store and manage data. Take any customer
relationship management (CRM) application – if it can’t store and retrieve customer-related data such as account name holder,
company name, address, and activity, it’s useless.
Relational database management systems were perfect for managing data from these types of applications because the data
was highly structured and static, and the query language relational database management systems used to manage the data,
SQL (structured query language), was easy to learn.
For legacy applications, data can reside on a single server and scale up simply by adding more servers, and that’s exactly the
way it was done for a long time.
But when the cloud era began, architects and developers had to start doing un-natural things with the relational model to be
able to distribute the data.
They had to break the model to partition the data into smaller chunks and shard them across servers to spread the load. They
also found that there were new types of data coming in – unstructured data from social media data which did not fit into the
relational model.
Thus, a new type of database was required for these new cloud applications and their new distributed data demands.
These new “NoSQL” (not only SQL) databases started springing up everywhere and used a variety of data models, including
key/value, document, graph, and columns.
By 2012 the competition began to shake out into NoSQL winners and losers, and Apache Cassandra established itself as the
go to open-source database for scale-out, distributed applications. It would go on to be shepherded and developed for
enterprises by DataStax.
3
The Magic Number 5
The 5 Characteristics of Cloud Applications
Cloud applications need to be able to operate at a global
scale. Wherever they are in the world, users expect the
application to be available and responsive at all times as
long as they have a functional web connection.
This puts a lot of strain on the application, but even more
strain on the database layer supporting the application,
which is responsible for reading and writing data
instantaneously and scaling to millions of globally
distributed users.
That’s why cloud applications are responsible for the
advent of NoSQL databases and the need for powerful
database management solutions. To perform well, they
need to be:
Users expect a certain level of performance
from the applications they use, and if they
don’t operate at real-time speeds, the effect
on the user is similar to the application being
unavailable. Frustration levels rise, and
customers begin to look elsewhere.
4. Distributed
The application is distributed across many
servers in multiple locations, allowing its users
to access it from any geographic location and
ensuring very fast response times.
1. Contextual
5. Scalable
To provide the best user experience possible,
applications need to provide contextually
relevant data. This can include things like a
nearest location, a previous order at a
restaurant, insight into customer buying
patterns, recommendations for a movie to
watch, and on and on – the possibilities are
endless.
Today, user volumes can expand almost
instantly. If a program goes viral, the
company, which may have planned for one
hundred thousand customers, suddenly finds
it has over a million. While this may be an
enviable position to be in, it can also be
extremely dangerous to companies unable to
scale their applications to support all the new
users. In this scenario, the application and
data management layer must be able to
scale instantly.
2. Always On
It wasn’t that long ago that companies could
schedule downtime to update applications.
But nowadays, if an application isn’t available
when a user wants to access it, or if it is down
completely, then the company is losing
customers and risking its entire brand and
reputation. In this era of cloud applications,
downtime simply is not an option. The
applications must be continually updated,
which also means rolling updates to the data
management layer with no planned downtime.
4
3. Real-Time
These are the 5 characteristics of highperforming cloud applications. These are the
traits that make up the apps you probably use
on almost a daily basis – apps that make your
life easier in so many ways.
But without a powerful data layer, cloud
applications couldn’t perform the way they do.
The Magic Number 5
DataStax Enterprise: Managing Data for Cloud Applications
As noted above, cloud applications rely on data to make them contextually relevant, and they must also be available and
responsive, which means that the data management platform must be an enabler of these characteristics.
DataStax Enterprise, built on Apache Cassandra, is architected to provide contextual data, and to be always-on, real time,
distributed, and scalable. It has an elegant flexible architecture which makes it easy to scale across data centers to distribute
the data, and place it close to a cloud application’s end users.
Cloud applications consume data via a variety of different access patterns, and this has historically required using several
different databases, one for each access method. The complexity created by stitching together multiple systems makes it
difficult and costly to maintain or develop new functionality.
That’s why DataStax Enterprise exists, and here’s what it brings to the table.
Zero Downtime
Every month there’s a new headline about a datacenter going down.
The fact is that hardware failures can and will occur, so architects need to ensure their data management layer has a built-in
failsafe, a distributed architecture with no single points of failure and redundancy of both functionality and data.
To support the distributed and always-on nature of cloud applications, DataStax Enterprise uses advanced replication, which is
specifically designed to support microservice deployments such as local retail analytics and to tolerate sporadic connectivity
that can occur in constrained environments, such oil-and-gas remote sites and cruise ships.
In these environments, customers need the ability to flexibly deploy a multi-directional replication system that allows them to
strategically manage and prioritize data in order to make the best use of resources and limited bandwidth. Additionally, when
connectivity is lost, data must be stored at the edge so that when connectivity is restored, replication resumes and no data
is lost.
With data replication, if one or more database servers or ‘nodes’ goes down, the other nodes in the system are able to
continue with operations without data loss, thereby providing true fault tolerance. In this way, the data management layer can
provide continuous availability whether in single locations, across data centers, or in the cloud.
Multi Modes
DataStax Enterprise employs a powerful, multi-modal platform with support for key-value, tabular, document (JSON) and
graph. This capability allows architects to write data to a single solution and access it using a variety of methods based on the
needs of the application. The addition of graph capabilities is particularly powerful as it can be used for managing and querying
data that is complex and highly connected. It makes it effortless to find commonalities or abnormalities and unlock the value in
data relationships.
5
The Magic Number 5
Security
Enterprises increasingly expect mature security functionality in any platform that they are going to widely deploy in the
enterprise to power cloud applications. DataStax Enterprise provides the tools necessary for stringent HIPAA, PCI and SOX
compliance requirements. This functionality includes support for industry-standard authentication mechanisms, role-based
authentication, user activity auditing, and end-to-end encryption.
Search
Most cloud applications need to provide a way to search the
data, and DataStax Enterprise is the only commercial search
solution that can scale for the needs of enterprises today. DSE
Search intelligently partitions data across the cluster to avoid hot
spots, evenly distribute load, and route queries to provide the
responsiveness that a cloud application requires.
Analytics
As noted earlier, cloud applications need to be contextual to be
effective, and to do this their data platforms need to use best-ofbreed data analytics. Fast analysis of transactional data is
another big advantage of DataStax Enterprise. DSE Analytics
makes it easy to generate ad-hoc reports, target customers with
advanced personalization, and process real-time streams of
data to create intelligent cloud applications.
OpsCenter
Devops teams want to get their cloud applications to market as quickly as possible and don’t want to have to spend time
managing the data platform. DataStax OpsCenter is a browser-based, visual management and monitoring solution for
DataStax Enterprise. OpsCenter provides architects, database administrators, and operations staff with the tools necessary to
intelligently and proactively ensure their databases are running well and that administrative tasks are simplified so IT staff can
concentrate on things other than babysitting their database systems.
All of these added features set DataStax Enterprise apart from the Cassandra base and make it a powerful tool for enterprises.
But no technology survives without adapting, and DataStax is constantly evolving to match the evolution of data itself.
6
The Magic Number 5
Goodbye, Cloud. Hello… What?
In client-server era not many people could have imagined the cloud, and now, in the cloud era, it’s hard to see exactly what’s
coming next. But with all devices becoming “smart”, via sensors and network connections, it’s clear we’re headed for
connectivity at a massive global scale – not just between people but things.
In a way, the Internet of Things is the new cloud. Applications are becoming far more powerful with their connections to
objects, and this is creating a huge opportunity when it comes to data and understanding the behaviors and preferences of
B2B and B2C customers.
Enterprises have a big challenge ahead of them:
corralling this avalanche of data into useful,
actionable insights they can use to understand their
customers at a granular level, create targeted
marketing promotions and product introductions or
enhancements, and prevent fraud and identity theft.
DataStax Enterprise, offering the best distribution of
Apache Cassandra, is building out in the direction of
customer experience – using its powerful platform to
capture and capitalize on every moment and allow
companies to provide personalized experiences via
their customer-facing cloud applications. The data
layer is the key. The PC is still here, but data
management has evolved with it, and the time to
invest in powerful data management is now, before
the next revolution leaves you in the dust.
DataStax is the best distribution of Apache Cassandra. But what does that mean, exactly? Click here to learn more.
About DataStax
It starts with a human desire, and when a universe of technology, devices and data aligns, it ends in a moment of fulfillment and insight.
Billions of these moments occur each second around the globe. They are moments that can define an era, launch an innovation, and
forever alter for the better how we relate to our environment. DataStax is the power behind the moment. Built on the unique architecture of
Apache Cassandra™, DataStax Enterprise is the always-on data platform and has been battle-tested for the world’s most innovative, global
applications.
With more than 500 customers in over 50 countries, DataStax provide data management to the world’s most innovative companies, such
as Netflix, Safeway, ING, Adobe, Intuit, Target and eBay. Based in Santa Clara, Calif., DataStax is backed by industry-leading investors
including Comcast Ventures, Crosslink Capital, Lightspeed Venture Partners, Kleiner Perkins Caufield & Byers, Meritech Capital, Premji
Invest and Scale Venture Partners. For more information, visit DataStax.com/customers or follow us on @DataStax.
© 2017 DataStax, All Rights Reserved. DataStax is a registered trademark of DataStax, Inc. and its subsidiaries in the United States and/or
other countries. Apache Cassandra is a trademark of the Apache Software Foundation or its subsidiaries in Canada, the United States
and/or other countries.