Architecting for Fast Data Applications

Introduction 2
Fast Data: The New Big Data 3
Fast Data Applications in Action 8
A Reference Architecture for Fast Data Applications 10
1. High availability with no single point of failure 11
2. Elastic scaling 12
3. Storage management 12
4. Infrastructure and application-level monitoring & metrics 13
5. Security and access control 14
6. Ability to build and run applications on any infrastructure 15
Fast Data Applications Require New Platform Services 17
1. Delivering real-time data 19
2. Storing distributed data 21
3. Processing fast data 22
4. Acting on data 23
Key Challenges Implementing Fast Data Services 24
1. Deploying each data service is time consuming 24
2. Operating data services is manual and error-prone 25
3. Infrastructure silos with low utilization 26
Public Cloud - The Solution? 28
Mesosphere DC/OS: Simplifying the Development and Operations of Fast Data Applications 30
1. On-demand provisioning 32
2. Simplied operations 34
3. Elastic data infrastructure 36
Case Studies: Fast Data Done Well 37
Verizon Adopts New Strategic Technologies to Serve Millions of Subscribers in Real-
Time 37
Esri Builds Real-Time Mapping Service With Kafka, Spark, and More 39
Wellframe Expands its Healthcare Management Platform 41

1 Mesosphere, Inc.
Architecting for Fast Data Applications

INTRODUCTION
In todays always-connected economy, businesses need to provide real-
time services to customers that utilize vast amounts of data. Examples
aboundreal-time decision-making in nance and insurance, to enabling
the connected home, to powering autonomous cars. While innovators
such as Twitter, Uber and Netix were at the forefront of creating
personalized, real-time services for their customers, companies of all
shapes and sizes in industries including telecom, nancial services,
healthcare, retail, and many more now need to respond or face risk of
disruption.

To serve customers at scale and process and store the huge amount of
data they produce and consume, successful businesses are changing how
they build applications. Modern enterprise applications are shifting from
monolithic architectures to cloud native architectures: distributed systems
of microservices, containers, and data services. Modern applications built
on cloud native platform services are always-on, scalable, and efcient,
while taking advantage of huge volumes of real-time data.

However, building and maintaining the infrastructure and platform services

(for example, container orchestration, databases and analytics engines)
for these modern distributed applications is complex and time-consuming.
For immediate access, many companies leverage cloud-based
technologies, but risk lock-in as they build their applications on a specic
cloud providers APIs.

This eBook details the vital shift from big data to fast data, describes the
changing requirements for applications utilizing real-time data, and
presents a reference architecture for fast data infrastructure.

2 Mesosphere, Inc.
Architecting for Fast Data Applications

FAST DATA: THE NEW BIG DATA

Data is growing at a rate faster than ever before. Every day, 2.5 quintillion
bytes of data are created1 - equivalent to more than 8 iPads per person.2
The average American household has 13 connected devices, and
enterprise data is growing by 40% annually. While the volume of data is
massive, the benets of this data will be lost if the information is not
processed and acted on quickly enough.

One of the key drivers of the sheer increase in the volume of data is the
growth of unstructured data, which now makes up approximately 80% of
enterprise data. Structured data is information, usually text les, displayed
in titled columns and rows which can easily be analyzed. Historically,
structured data was the norm because of limited processing capability,
inadequate memory and high costs of storage. In contrast, unstructured
data has no identiable internal structure; examples include emails, video,
audio and social media. Unstructured data has skyrocketed due to the
increased availability of storage and the number of complex data sources.

1 Vouchercloud Big Data infographic

2 Based on 32 GB iPad

3 Mesosphere, Inc.
Architecting for Fast Data Applications

Block Based (CAGR = 34.0%) File Based (CAGR = 45.6%)

160

120

80

40

0
2008 2009 2010 2011 2012 2013 2014 2015

Worldwide File-Based Versus Block-Based Storage Capacity Shipments, 2008-2015

Source: IDC Worldwide File-Based Storage 2011-2015 Forecast, December 2011

The term big data was popularized in the early- to mid-2000s, when many
companies started to focus on obtaining business insights from the vast
amounts of data being generated. Hadoop was created in 2006 to handle
the explosion of data from the web.

While most large enterprises have put forth efforts to build data
warehouses, the challenge is in seeing real business impact
organizations leave the vast amount of unstructured data unused. Despite
substantial hype and reported successes for early adopters, over half of
the respondents to a Gartner survey reported no plans to invest in Hadoop
as of 2015.3 The key big data adoption inhibitors include:

3 Survey Analysis: Hadoop Adoption Drivers and Challenges, Gartner, May 2015

4 Mesosphere, Inc.
Architecting for Fast Data Applications

1. Skills gaps (57% of respondents): Large, distributed systems are

complex, and most companies do not want to staff an entire team on a
Hadoop distribution.

2. Unclear how to get value from Hadoop (49% of respondents): Most

companies have heard they need Hadoop, but cannot always think of
applications for it.

1990s 2013+
Online customer Real-time & predictive
engagement customer engagement

1980s 2000s
Electronic customer Customer analytics
records

Industry Transitions

Over the past two to three years, companies have started transitioning
from big data, where analytics are processed after-the-fact in batch mode,
to fast data, where data analysis is done in real-time to provide immediate
insights. For example, in the past, retail stores such as Macys analyzed
historical purchases by store to determine which products to add to stores
in the next year. In comparison, Amazon drives personalized
recommendations based on hundreds of individual characteristics about
you, including what products you viewed in the last ve minutes.

Big data is collected from many sources in real-time, but is processed

after collection in batches to provide information about the past. The
benets of data are lost if real-time streaming data is dumped into a
database because of the inability to act on data as it is collected.

Modern applications need to respond to events happening now, to provide

insights in real time. To do this they use fast data, which is processed as it
is collected to provide real-time insights. Whereas big data provided

5 Mesosphere, Inc.
Architecting for Fast Data Applications

insights into user segmentation and seasonal trending using descriptive

(what happened) and predictive analytics (what will likely happen), fast
data allows for real-time recommendations and alerting using prescriptive
analytics (what should you do about it).

VERTICAL BIG DATA FAST DATA

Automotive Automakers analyze large Connected cars provide real-

sets of crash and car-based time trafc information and
sensor data to improve safety alerts for predictive
features maintenance.

Healthcare Doctors provide care Doctors provide insightful

suggestions based on care recommendations based
historical analysis of large on predictive models and in-
datasets the-moment patient data.

Retail Stores determine which Online retailers provide

products to stock based on personalized
analysis of previous quarters recommendations based on
purchase data. hundreds of individual
characteristics, including
products you viewed in last
ve minutes.

Financial Credit card companies create Credit card companies alert

Services models for credit risk based customers of potential fraud
on demographic data. in real-time.

Manufacturing Manufacturing plants Manufacturing plants detect

improve efciency based on product quality issues before
throughput analysis. they even occur.

Big Data Vs. Fast Data Examples

6 Mesosphere, Inc.
Architecting for Fast Data Applications

Businesses are realizing they can leverage multiple streams of real-time

data to make in-the-moment decisions. But more importantly, fast data
powers business critical applications, allowing companies to create new
business opportunities and serve their customers in new ways. Over 92%
of companies plan to increase their investment in streaming data in the
next year4 , and those who dont face risk of disruption.

80%

68%
60%

40%

20%

14%
10%
5% 1%
0%
Reduce batch, Increase investment Eliminate batch, Reduce stream, Eliminate stream,
increase stream in both shift to stream increase batch shift to batch

How Will Usage of Batch and Streaming Shift in Your Company in the Next One Year?
Source: 2016 State of Fast Data and Streaming Applications, OpsClarity

4 OpsClarity Fast Data Survey, 2016

7 Mesosphere, Inc.
Architecting for Fast Data Applications

FAST DATA APPLICATIONS IN ACTION

GE is an example of an organization that is already using fast data both to

improve existing revenue streams and create new ones. GE is harnessing
the data generated by its equipment to improve performance and
customer experience through preventive maintenance. Additional benets
include reduced unplanned downtime, increased productivity, lowered fuel
costs, and reduced emissions. The platform will also be able to offer new
services such as remote monitoring and customer behavior analysis that
will represent new revenue streams.5

Ubers ride sharing service depends on fast datathe ability to take a

request from anywhere in the world, map that request to available drivers,
calculate the route cost, and link all that information back to the customer.
This requirement may seem simple, but it is actually a complex problem to
solveresponding within just a few seconds is necessary in order to
differentiate Uber from the wider market. Uber is also using their fast data
platform to generate new revenue streams, including food delivery.

5 Big & Fast Data, CapGemini, 2015

8 Mesosphere, Inc.
Architecting for Fast Data Applications

At Capital One, analytics are not just used for pricing and fraud detection,
but also for predictive sales, driving customer retention, and reducing the
cost of customer acquisition. Machine learning algorithms play a critical
role at Capital One. Every time a Capital One card gets swiped, we capture
that data and are running modeling on it, Capital One data scientist
Brendan Herger says. The results of the fast data analytics have made
their way into new offerings, such as the Mobile Deals app that sends
coupon offers to customers based on their spending habits. It has also
enabled predictive capabilities in the call center, which CapGemini says
can determine the topic of a customers call within 100 milliseconds with
70 percent accuracy.6

6 How Credit Card Companies Are Evolving with Big Data, Datanami, May 2016

9 Mesosphere, Inc.
Architecting for Fast Data Applications

A REFERENCE ARCHITECTURE FOR FAST DATA

APPLICATIONS
Real-time data-rich applications are inherently complicated. Data is
constantly in motion, moving through distributed systems including
message queues, stream processors, and distributed databases.

To handle massive amounts of data in real-time, successful businesses

are changing how they build applications. This shift primarily entails
moving from monolithic architectures to distributed systems composed of
microservices deployed in containers, and platform services such as
message queues, distributed databases, and analytics engines.

Architectural Shift From Monolithic Architectures to Distributed Systems

The key reasons enterprises are moving to a distributed computing

approach include:

10 Mesosphere, Inc.
Architecting for Fast Data Applications

1. The large volume of data created today cannot be processed on any

single computer. The data pipeline needs to scale as data volumes
increase in size and complexity.

2. Having a potential single point of failure is unacceptable in an era when

decisions are made in real time, loss of access to business data is a
disaster, and end users don't tolerate downtime.

To successfully build and operate fast data applications on distributed

architectures, there are 6 critical requirements for the underlying
infrastructure.

1. High availability with no single point

of failure
Always-on streaming data pipelines require a new architecture to retain
high availability while simultaneously scaling to meet the demands of
users. This is in contrast to batch jobs that are run ofineif a three-hour
batch job is unsuccessful, you can rerun it. Streaming applications need to
run consistently with no downtime, with guarantees that every piece of
data is processed and analyzed and that no information gets lost.

Today, applications no longer t on a single server, but instead run across

a number of servers in a datacenter. To ensure each application (e.g.
Apache CassandraTM) has the resources it needs, a common approach is
to create separate clusters for each application. But what happens when a
machine dies in one of these static partitions? Either there is extra
capacity available (in which case the machines have been over-
provisioned, wasting money), or another machine will need to be
provisioned quickly (wasting effort).

The answer lies in datacenter-wide resource scheduling. Machines are the

wrong level of abstraction for building and running distributed
applications. Aggregating machines and deploying distributed

11 Mesosphere, Inc.
Architecting for Fast Data Applications

applications datacenter-wide allows the system to be resilient against

failure of any one component, including servers, hard drives, and network
partitions. If one node crashes, the workloads on that node can be
immediately rescheduled to another node, without downtime.

2. Elastic scaling
Fast data workloads can vary considerably over a month, week, day, or
even hour. In addition, the volume of data continues to multiply. Based on
these two factors, fast data infrastructure must be able to dynamically and
automatically scale horizontally (i.e. changing the number of service
instances), and vertically (i.e. allocating more or less resources to
services), up or down. And so data doesnt get lost, scaling must occur
with no downtime.

Elastic Resource Sharing Example

A shared pool of resources across data services facilitates elastic

scalability, as workloads can burst into available capacity occupied in the
cluster.

3. Storage management
Fast data applications must be able to read and write data from storage in
real time. There are many kinds of storage, such as local le systems,

12 Mesosphere, Inc.
Architecting for Fast Data Applications

volumes, object stores, block devices, and shared, network-attached, or

distributed lesystems, to name a few. Each of these storage systems
have different characteristics and each data service may require or
support a different storage type.

In some cases, the data service by nature is distributed. Most NoSQL

databases subscribe to this model and are optimized for a scale out
architecture. In these cases, each instance has its own dedicated storage
and the application itself has semantics for synchronization of data. For
this use case, local, dedicated, persistent storage optimized for
performance and resource isolation is key. Local persistent storage is
local to the node within the cluster and is usually the storage resident
within the machine. These disks can be partitioned for specic services
and will typically provide the best in terms of performance and data
isolation. The downside to local persistent storage is that it binds the
service or container to a specic node.

In other cases, services that can take advantage of a shared backend

storage system are better suited for external storage which may be
network attached and optimized for sharing between instances. External
storage in that case may be implemented in some form of storage fabric,
distributed or shared lesystem, object store, or other storage service.

4. Infrastructure and application-level

monitoring & metrics
Collecting metrics is a requirement for any application platform, but it is
even more vital for data pipelines and distributed systems because of the
interdependent nature of each pipeline component and the many
processes distributed across a cluster. Metrics allow operators to analyze
issues across the data pipeline, including latency, throughput, and data
loss. In addition, metrics allow organizations to gain visibility on
infrastructure and application resource utilization, so that they can right

13 Mesosphere, Inc.
Architecting for Fast Data Applications

size the application and the underlying infrastructure, ensuring optimum

resource utilization.

Traditional monitoring tools do not address the specic capabilities and

requirements of web scale, fast data environments. With no service-level
metrics, operators cannot troubleshoot or monitor performance and/or
capacity. Traditional monitoring tools can be adapted, but they require
additional operational overhead for distributed applications. If monitoring
tools are custom implemented, they require signicant upfront and
ongoing engineering effort.

To build a robust data pipeline, collect as many metrics as feasible, with

sufcient granularity. And to analyze metrics, aggregate data in a central
location. But beyond per-process logging and metrics monitoring, building
microservices at scale also requires distributed tracing to reconstruct the
elaborate journeys that transactions take as they propagate across a
distributed system. Distributed tracing has historically been a signicant
challenge because of a lack of tracing instrumentation, but new tools such
as OpenTracing7 make it much easier.

5. Security and access control

Without platform-level security, businesses are exposed to tremendous
risks of downtime and malicious attacks. Independent teams can
accidentally alter or destroy services owned by other teams or impact
production services.

Traditionally, teams build and maintain separate clusters for their

applications including dev, staging, and production. As monolithic
applications are rebuilt as microservices and data services, the size and
complexity of these clusters continue to grow, siloed by teams and the
technology being used. Without multitenancy, running modern
applications becomes exponentially complex, because different teams

7 http://opentracing.io

14 Mesosphere, Inc.
Architecting for Fast Data Applications

may be using different versions of data services, each congured with

hardware expecting peak demand. The result is extremely high operations,
infrastructure and cloud costs driven by administration overhead, low
utilization, and multiple snowake implementations (with unique clusters
useable for only one purpose).

To create a multi-tenant environment while providing appropriate platform

and application-level security, it is necessary to:

1. Integrate with enterprise security infrastructure (directory services and

single sign-on).

2. Dene ne-grained authentication and authorization policies to isolate

user access to specic services based on a users role, group
membership, or responsibilities.

6. Ability to build and run applications on

any infrastructure
Fast data pipelines should be architected to exibly run on any on-premise
or cloud infrastructure. For performance and scalability of fast data
workloads, you need to have the choice to deploy infrastructure that meets
the specic needs of your application. For example, the most sensitive
data can be kept on-premises for privacy and compliance reasons, while
the public cloud can be leveraged for dev and test environments. The
cloud can also be used for burst capacity. Wikibon estimates that
worldwide big data revenue in the public cloud was $1.1B in 2015 and will
grow to $21.8B by 2026or from 5% of all big data revenue in 2015 to 24%
of all big data spending by 2026.8

For such hybrid scenarios companies often nd themselves stuck with

two separate data pipeline solutions and no unied view of the data ows.
While the choice of infrastructure is vital, a key requirement is a similar

8 Big Data in the Public Cloud Forecast, Wikibon, 2016

15 Mesosphere, Inc.
Architecting for Fast Data Applications

operating environment and/or single pane of glass, so that workloads can

easily be developed in one cloud and deployed to production in another.

16 Mesosphere, Inc.
Architecting for Fast Data Applications

FAST DATA APPLICATIONS REQUIRE NEW

PLATFORM SERVICES
Weve covered the requirements for the underlying infrastructure for fast
data applications. What about the fast data services deployed on this
infrastructure, such as analytics engines and distributed databases?

Another key shift in the architectural components of fast data systems is

from the use of proprietary closed source systems to data pipelines
stitched together from a variety of open source tools. In a recent survey,
over 90% of respondents leveraged open source distributions for fast data
applications, and almost 50% used open source exclusively.9 Open source
tools are a force multiplier for developers getting started, and also can be
used to avoid lock-in to proprietary solutions.

9 OpsClarity Fast Data Survey, 2016

17 Mesosphere, Inc.
Architecting for Fast Data Applications

Platform Services

Today, most people think of Hadoop or NoSQL databases when they think
of big data. Recently, several open source technologies have emerged to
address the challenges of processing high-volume, real-time data, most
prominently including Apache KafkaTM for data ingestion, Apache SparkTM
for data analysis, Apache Cassandra for distributed storage, and Akka for
building fast data applications.

18 Mesosphere, Inc.
Architecting for Fast Data Applications

Fast Data Pipeline with Kafka, Cassandra, Spark, and Akka

1. Delivering real-time data

When constant streams of data arrive in millions of events per second
from connected sensors and applicationsfrom cars, wearables, buildings
to everything else, the data needs to be ingested in real-time, with no data
loss.

Apache Kafka, originally developed by the engineering team at LinkedIn, is

a high-throughput distributed message queue system for ingesting
streaming data. Kafka is known for its unlimited scalability, distributed
deployments, multitenancy, and strong persistence.

Kafka allows companies to publish and subscribe to streams of records

(i.e. as a message queue), store streams of records in a fault-tolerant way,
and process streams of records as they occur. Kafka makes a great buffer
between downstream tools like Spark and upstream sources of data, in
particular data sources that cannot be queried again if data is lost
downstream.

19 Mesosphere, Inc.
Architecting for Fast Data Applications

Apache Kafka Architecture

While Kafka is the most popular message broker, other popular tools
include Apache FlumeTM and RabbitMQ. Apache Flume is a distributed,
reliable, and available service for efciently collecting, aggregating, and
moving large amounts of data. RabbitMQ, backed by Pivotal, is a popular
open source message broker that gives applications a common platform
to send and receive messages. RabbitMQ is preferred for use-cases
requiring support for Advanced Message Queuing Protocol (AMQP).

80%
86%
60%

40%

20%
22% 21%
11% 11%
0%
Apache Kafka Apache Flume Rabbit MQ Amazon SQS AWS Kinesis

Most Popular Ingestion Queues

Source: OpsClarity Fast Data Survey, 2016

20 Mesosphere, Inc.
Architecting for Fast Data Applications

2. Storing distributed data

Traditional relational databases (RDBMSs) were the primary data stores
for business applications for 20 years, and new databases such as MySQL
were introduced with the rst phase of the web. However, the scaling and
availability needs of modern applications require a new highly durable,
available, and scalable database to store streaming and processed data.

Apache Cassandra is a large-scale NoSQL database designed to handle

large amounts of data across many commodity servers, providing high
availability with no single point of failure. Cassandra supports clusters
spanning multiple datacenters, providing lower latency and keeping data
safe in the case of regional outages.

Some of the largest production deployments include Apple's, with over

75,000 nodes storing over 10 PB of data, Netix (2,500 nodes, 420 TB,
over 1 trillion requests per day), Chinese search engine Easou (270 nodes,
300 TB, over 800 million requests per day), and eBay (over 100 nodes, 250
TB).10

Other popular distributed NoSQL databases include MongoDB, Redis, and

Couchbase:

MongoDB is an open-source, document database designed for ease of

development and scaling.

Redis is an in-memory data structure store, used as a database, cache

and message broker, for high performance operational, analytics or
hybrid use

Couchbase is a NoSQL database that makes it simple to build

adaptable, responsive always-available applications that scale to meet
unpredictable spikes in demand and enable mobile and IoT
applications to work ofine.

10 http://cassandra.apache.org/

21 Mesosphere, Inc.
Architecting for Fast Data Applications

3. Processing fast data

Incoming data from sensors and applications needs to be processed
using batch, streaming, machine learning, or graph compilation to gain
new insights. Hadoop is a well-known tool for batch analysis, but Hadoop
MapReduce has limitations for rapid analysis of smaller datasets.

Apache Spark is an open source processing engine built around speed,

ease of use, and sophisticated analytics. Originally developed at UC
Berkeley in 2009, Spark allows companies to run programs up to 100x
faster than Hadoop MapReduce in memory, or 10x faster on disk.11 It is
easy to use, with the ability to write applications quickly in Java, Scala,
Python, and R. Spark supports SQL, streaming, machine learning, and
graph computation.

Apache Spark is seeing exponential growth in adoption and awareness.

Today, Spark remains the most active open source project in big data, with
over 1,000 contributors. In a recent Spark survey by Taneja group, nearly
one half of all respondents were already using Spark, and of those, 64
percent say that its proving invaluable and that they intend to increase
usage of Spark within the next 12 months.12

Other popular analytics tools include MapReduce, Apache StormTM and

Apache FlinkTM:

MapReduce is a programming model and an associated

implementation for processing and generating big data sets with a
parallel, distributed algorithm on a cluster. Spark runs in memory, while
MapReduce is batch-based and is restricted to writing to and from
disk. MapReduce is part of Hadoop and thus requires HDFS as its
primary storage layer. MapReduce is included in Hadoop distributions,
and powers approximately half of processing environments.

11 http://spark.apache.org/
12 Apache Spark Market Survey, Taneja Group, November 2016

22 Mesosphere, Inc.
Architecting for Fast Data Applications

Apache Storm, an open source distributed realtime computation

system, is another popular stream processing platform. Storm makes
it easy to reliably process unbounded streams of data, doing for
realtime processing what Hadoop did for batch processing.

Apache Flink is an open source distributed data stream processor;

Flink provides efcient, fast, consistent, and robust handling of
massive streams of events, as well as batch processing as a special
case of stream processing.

80%

70%
60%

50%
40%

20%
27%

0%
Apache Spark MapReduce Apache Storm

Most Popular Data Processing Technologies

Source: OpsClarity Fast Data Survey, 2016

4. Acting on data
Once real-time data is analyzed, insights need to be presented to a human
or trigger actions in connected devices or applications. Akka is a popular
toolkit and runtime to simplify development of data centric applications.
Akka was designed to enable developers to easily build reactive
applications using a high level of abstraction, and the technology makes
building highly concurrent, distributed, and resilient message-driven
applications on the JVM a much simpler process.

23 Mesosphere, Inc.
Architecting for Fast Data Applications

KEY CHALLENGES IMPLEMENTING FAST DATA

SERVICES
While the use of open source tools such as Kafka and Spark has grown
immensely, the distributed nature of these fast data technologies can
make them difcult to deploy and operate. For example, Braintree, a
mobile and web payment systems company, built their own data pipeline
using open source software and lots of custom code. Braintree devoted a
team of four full-time engineers for six months to get their data
infrastructure off the ground, and even after the initial launch, a two-
person team is required to maintain and extend the project.13

But just why is building and maintaining data infrastructure so taxing?

1. Deploying each data service is time

consuming
First, deploying each data service is time consuming. Installing a
production-grade platform service such as Kafka or Cassandra requires
specialized knowledge for operators; even for an expert, deployment is
time consuming and often requires signicant engineering effort. For
example, when the engineering team at AirBnB rst attempted to deploy
Kafka, it took multiple months, and ultimately ran into many issues.
Additionally, if fast data services are not architected correctly, the result is
snowake implementations that are dependent on key personnel to
maintain.

In the past ve to ten years, an explosion of datastores and analytics

engines have emerged, many open source. Beyond the signicant cost of

13 Stitchdata blog, Why you shouldnt build your own data pipeline

24 Mesosphere, Inc.
Architecting for Fast Data Applications

resources to deploy each new data service, a primary downside of

deployment challenges is that developers and data engineers cannot
easily experiment with these new technologies.

Explosion of Datastores and Analytics Engines

2. Operating data services is manual

and error-prone
The second major challenge with implementing fast data technologies is
the ongoing operations of these technologies; common tasks such as
upgrading software, deploying updates, rolling back in failure scenarios,
monitoring health, and managing storage resources are often manual and
error-prone. This time-consuming ongoing maintenance requires
additional headcount and can be a drag on new innovation.

Applications teams and operators need the latest capabilities in data

services in order to x critical bugs, add valuable features, and reduce
operational overhead. Software upgrades and updates are time
consuming because operators and developers manually upgrade the
services or are required to build conguration management and
orchestration software to automate the upgrade. In addition, updates and

25 Mesosphere, Inc.
Architecting for Fast Data Applications

rollback need to be thoroughly validated on a testing environment before

they are ready for production.

3. Infrastructure silos with low utilization

The third key challenge with operating fast data infrastructure is
maintaining enough infrastructure to handle data-processing peaks in a
cost-effective way. Average datacenter utilization is approximately 6-12%14
and remained static from 2006-201215, driven by companies desire to
maintain high service quality through peak load periods.

Traditional Approach: Silos with Low Utilization

The increased utilization benet of virtual machines does not apply to

distributed systems. Operators typically create separate clusters for Spark,
Kafka, Cassandra, and so on. This is because these services are
distributed systems with their own scheduling logic to grow and shrink
their footprint, so operators run them in different clusters to avoid
resource conicts. For example, the Spark data processing service may
want all the capacity it can get to nish a job, the Kafka message queue
may need resourcing based on the volume of data passing through, while
the Cassandra distributed database may need steady resourcing to persist
data. The result is extremely low utilization, and waste of infrastructure
resources.

14 The Sorry State of Server Utilization and the Impending Post Hypervisor Era, Gigaom,
November 2013
15 NRDC Data Center Efciency Assessment, August 2014

26 Mesosphere, Inc.
Architecting for Fast Data Applications

Rather than running workloads in silos based on the application,

datacenter-wide resource sharing drives increased utilization as all
workloads can access a shared pool of resources.

27 Mesosphere, Inc.
Architecting for Fast Data Applications

PUBLIC CLOUD - THE SOLUTION?

Many companies are shifting fast data workloads to the cloud or hybrid
environments. The two key reasons for moving the cloud are:

1. To get immediate access to platform services such as analytics tools

and databases. Cloud providers such as Amazon Web Services and
Microsoft Azure package data services and make them easy to install
and operate.

2. To enable scalability and elasticity, an imperative for bursting fast data

workloads.

In a recent survey, less than a third of companies hosted their data

pipeline on-premises, while more than two thirds hosted in the cloud or
hybrid environments. 16

While public cloud provides clear advantages for fast data workloads, the
major downside is the risk of lock-in. Applications that are developed
using public cloud platforms are tied to a specic cloud providers APIs,
and moving workloads after the fact is near impossible without rewriting
them.

One recent story highlights the risk of cloud lock-in, that of Snap (of
Snapchat). In the S1 Registration Statement issued by Snap17 in February
2017, it came to light that Snap had handcuffed itself to Google Cloud. Of
the annual loss of over $500 million, 80% was attributed to contractually
obligated spend with Google. In the same ling, Snap states that they
wrote their application to use some Google services which do not have an
alternative in the market. Google now has them in handcuffs, and there is

16 OpsClarity Fast Data Survey, 2016

17 Snap S1 Registration, February 2017

28 Mesosphere, Inc.
Architecting for Fast Data Applications

little Snap can do to change that without having to invest a tremendous

amount of money to free themselves.

As companies continue to rush to the cloud to quickly build new services,

the costly risk of lock-in continues to mount. But what is the alternative?

29 Mesosphere, Inc.
Architecting for Fast Data Applications

MESOSPHERE DC/OS: SIMPLIFYING THE

DEVELOPMENT AND OPERATIONS OF FAST DATA
APPLICATIONS
Mesosphere delivers a platform for building and running highly scalable
data pipelines, in any cloud or datacenter. Mesosphere DC/OS accelerates
deployment and simplies operations for a broad set of data services
including databases, message queues, analytics engines, and more.
Mesosphere enables experimentation with new data services and provides
a future-proof platform that is highly available and scales to meet the
demands of users.

30 Mesosphere, Inc.
Architecting for Fast Data Applications

Mesosphere DC/OS: Cloud Native Platform Services on Any Infrastructure

The core of DC/OS is the Apache MesosTM distributed systems kernel. Its
power comes from the two-level scheduling that enables distributed
systems to be pooled and share datacenter resources. Mesos provides
the core primitives for distributed systems, such as resource allocation,
isolation, and quota management. DC/OS provides a highly-available
infrastructure for fast data workloadsworkloads are automatically
restarted when a server fails. Pooling resources across a datacenter or
cloud also enables elastic scaling, where workloads can scale up or down
based on demand.

31 Mesosphere, Inc.
Architecting for Fast Data Applications

Distributed System A Distributed System B Mesosphere DC/OS Approach

Scheduler Scheduler Scheduler

Resource Task
Offer Launch
Task Task
Launch Launch

Task Task
Status Status
Task
Status

Executor
Executor Executor Executor
Executor Executor Executor
Executor Executor Executor
Executor

Distributed Systems
A+B+C+

Apache Mesos Two-Level Scheduling

Two-level scheduling in DC/OS provides key differentiators versus simply

running services in containers on Kubernetes or Docker Swarm, and
simplies the three key challenges with operating data services
deploying services, operating them, and overcoming silos with low
utilization.

1. On-demand provisioning
Mesosphere DC/OS enables single-command install of data services such
as Spark, Cassandra, Kafka and Elasticsearch, among many others. Where
deployment of these services used to be incredibly time-consuming and
error prone, data services can be up and running across an entire cluster in
a matter of minutes with Mesosphere DC/OS.

32 Mesosphere, Inc.
Architecting for Fast Data Applications

Data Service Installation with Mesosphere DC/OS

The DC/OS Universe is an open source ecosystem where anyone can

publish services to be installed on DC/OS. The DC/OS Universe includes
both open source and partner-supported data services, including products
from DataStax, Conuent, Elastic, and Alluxio, among many others. There
is a growing community of users and partners creating new DC/OS
packages.

In addition, an open source SDK provides a high-level interface for building

new stateful services on DC/OS. Developers can write a stateful service
complete with persistent volumes, fault tolerance, and conguration
management in about 100 lines of code. This SDK is the product of
Mesosphere's experience writing production stateful services for DC/OS
such as Kafka, Cassandra and HDFS.

33 Mesosphere, Inc.
Architecting for Fast Data Applications

Mesosphere DC/OS Universe with Over 100 Services

Faster deployment of data services allows companies to avoid nding and

paying for specialized talent, enables faster time to market for new fast
data applications, and allows data scientists to experiment with the broad
swath of new data services, and build on a composable architecture.

Mesosphere DC/OS also dramatically simplies resizing instances of a

data service, as well as adding more instances. With Mesosphere DC/OS,
operators can easily scale up and scale out data services, with no
downtime.

2. Simplied operations
Mesosphere DC/OS dramatically reduces the time and effort involved with
operating data services through simple runtime software upgrades and

34 Mesosphere, Inc.
Architecting for Fast Data Applications

updates, application-level monitoring and metrics, and managed

persistent storage volumes.

A key challenge with operating data services is upgrading and updating

cong settings in the data services themselvesthe risk is downtime and
wasted operator time. Mesosphere DC/OS includes built-in runtime
software upgrade capabilities, with rollback in case of failure. Application
setting updates can also be performed during runtime using the same
mechanism. Runtime upgrades and updates with the ability to rollback
minimizes maintenance windows and the risk due to unsuccessful cong
updates, and saves on operator and engineering time.

Mesosphere DC/OS Data Service Upgrades

Mesosphere DC/OS provides out-of-the-box application-level monitoring

and troubleshooting, so operators can easily troubleshoot or monitor
performance and capacity. DC/OS services send metrics to the customers
provided statsd metrics service, and commercial monitoring tools
integrate to provide aggregate metrics and logs per node, app and
container.

Mesosphere DC/OS also simplies operations by managing storage

volumes, like CPU, memory and network resources, so operators do not
need to manually keep track of resources. As a result, administration time
and overhead for provisioning required resources for data services at
scale are signicantly reduced.

35 Mesosphere, Inc.
Architecting for Fast Data Applications

3. Elastic data infrastructure

Mesosphere DC/OS enables multiple data services, containerized
applications and traditional applications to all run on the same
infrastructure, dramatically increasing utilization. Some Mesos and DC/OS
users have approached average utilization rates of over 90%, and reduced
hardware and cloud costs by over 60%. Operators can easily bring new
services such as Kafka online, or add more instances of the same service
to a cluster with already available resources, increasing efciency. In
addition, app teams can use the specic version of software they prefer,
without the need to create separate silos.

Traditional Approach Mesosphere DC/OS Approach

Server Server Server

Container Apps

Container Big Data Analytics Stateful Services

Big Data Stateful
Analytics Service
PaaS
PaaS

Server Server Server

Mesosphere DC/OS
Container
Big Data Stateful
Analytics Service
PaaS Any Infrastructure (On-Prem, Cloud)

Elastic Data Infrastructure with Mesosphere DC/OS

To increase utilization further, multiple users and teams can share a

DC/OS cluster. Built-in security features such as ne grained access
control lists, secrets management, and integration with directory services
(LDAP) and single sign-on solutions (SAML, OpenID connect) enable
companies to isolate access to specic services based on a users role,
group membership, or responsibilities.

36 Mesosphere, Inc.
Architecting for Fast Data Applications

CASE STUDIES: FAST DATA DONE WELL

Verizon Adopts New Strategic

Technologies to Serve Millions of
Subscribers in Real-Time
Verizon Communications is the #1 wireless phone service in the US. The
company's core mobile business, Verizon Wireless, serves about 113
million retail connections.

Verizon needed to easily deploy and manage thousands of Docker

containers and needed to be able to quickly adopt new strategic
technologies such as Apache Spark. Verizons research identied Apache
Mesos as the best option given Verizons scale. Verizon chose
Mesosphere DC/OS because it let the company more easily adopt Mesos
and the necessary components for proper microservices architectures,
and helped Verizon adopt new strategic technologies such as Apache
Spark for data-processing and analytics.

Verizon uses Mesosphere DC/OS across thousands of nodes across

multiple datacenters, and is continuing to expand deployment. Verizon
uses Spark, Kafka and Cassandra, among other data services. Thanks to
strong collaboration between Verizon and Mesosphere, Mesosphere
DC/OS is now the computing backbone of next-generation productson-
demand TV on Fios, smartphone streaming via Go90, and smarter devices
via the Internet of Thingsserving tens of millions of consumers.

37 Mesosphere, Inc.
Architecting for Fast Data Applications

Mesosphere DC/OS enabled:

Shared data infrastructure: The ability to run Spark, Hadoop, Kafka,
Cassandra and more on a single Mesosphere DC/OS cluster lets
Verizon build data-driven applications without moving data across
clusters or managing multiple environments.

Faster time to market: No dedicated hardware means applications

come online in a fraction of the time compared with historical
approaches. Verizon deployed a Mesosphere DC/OS datacenter
environment in the time it used to take to stand up application
infrastructure.

Acting fast on new trends: From streaming video services like Go90 to
drone video analysis, Mesosphere DC/OS lets Verizon act quickly on
the types of applications its consumer and enterprise customers
demand.

Mesosphere DC/OS gives Verizon far-reaching benets to quickly launch new products
and services while reducing the IT requirements in our data centers

- Kumar Vishwanathan, VP & Chief Technologist, Verizon

38 Mesosphere, Inc.
Architecting for Fast Data Applications

Esri Builds Real-Time Mapping Service

With Kafka, Spark, and More
Esri is a world leader in geographic information system (GIS) technology.
Esris clients wanted to deploy new IoT and data-driven applications that
were pushing the boundaries of Esris current on-premises Real-Time GIS
software solution. The performance and scalability demands of these new
applications required Esri to adopt a new technology platform to help
customers take their applications to the next level.

Esri has developed a new managed servicebuilt on Mesosphere DC/OS

that lets users achieve real-time, predictive mapping. It combines Esris
ArcGIS platform with real-time and big data analytic capabilities to
process and analyze up to millions of events per second from sources
such as:

Sensors on moving objects such as vehicles, vessels and people

Stationary sensors on electric, water and gas utility networks

Feeds from social media, weather and environmental sensors

Esri utilizes Apache Kafka, Apache Spark (and Spark Streaming),

Elasticsearch, and the Lightbend Reactive Platform (which includes Akka)
to bring their service to market.

Mesosphere DC/OS enabled:

Increased performance and scalability: Esris previous on-premises
software was able to process thousands of events per second. On the
Mesosphere DC/OS platform, Esris new managed service can process

39 Mesosphere, Inc.
Architecting for Fast Data Applications

millions of events per second, easily meeting the expanding needs of

its clients.

New class of customer: Esri is now getting requests from a new class
of customer with more sophisticated and large-scale applications.
Before DC/OS, Esri would shy away from those opportunities because
it was outside the scale of what its technology could handle.

Faster time to market: The time to get a customers environment up

and running has been drastically reduced. Depending on the
deployment, it could take anywhere from hours to days to get the
environment established and sometimes required sending staff to the
customer site if the customer requested. With the new cloud-based
platform, Esri can deploy a new cluster in minutes.

A foundation for innovation: The Mesosphere DC/OS platform provides

an innovative delivery model and is also a foundation to build higher-
level business applications. Having the investment in place enables
Esri to move from real-time GIS to predictive GIS, where they can make
predictions and recommendations rather than only providing alerts
about what has already happened.

With the Mesosphere DC/OS platform, we can serve a new set of customers with entirely
new capabilities in terms of the performance and intelligence of their map and analytic
applications. And with our cloud-based platform, we can get these solutions up and
running in minutes. This gives Esri and our clients a level of innovation and business
agility weve never had before.

- Adam Mollenkopf, Real-Time & Big Data GIS Capability Lead, Esri

40 Mesosphere, Inc.
Architecting for Fast Data Applications

Wellframe Expands its Healthcare

Management Platform
Wellframes intelligent care-management platform allows health plans and
care-delivery organizations to better manage large populations of patients
with complex medical conditions such as diabetes, heart disease and
transplants. To meet the diverse and ever changing needs of these
patients, Wellframe provides personalized and adaptive care programs
that dynamically adjust based on a combination of data from the
healthcare system and the Wellframe platform.

Wellframe faced a twofold challenge as its healthcare management

platform matured: increased complexity and limited resources. As
Wellframes business and product offering became more sophisticated,
the company needed to add new services. Each service leveraged new and
different technologies (including Apache Spark, Apache Kafka and Apache
Cassandra) and required additional hardware and engineering resources
to manage. With its small team and limited resources, Wellframe soon
realized it needed a different approach.

Wellframe had experience managing infrastructure with Apache Mesos,

and it chose DC/OS as the long-term foundation for its personalized
healthcare offering because DC/OS offered a more complete and much
simpler platform experience. Wellframe also was interested in the growing
number of services that DC/OS supportsincluding Spark, Cassandra,
Kafka and more being added all the time. The company liked the unied
user experience of being able to manage all of these systems through a
single interface.

41 Mesosphere, Inc.
Architecting for Fast Data Applications

Mesosphere DC/OS allows Wellframe to spend more time building

technology that directly affects the core business of the company, and
less time on the operational management that comes with running a large
production deployment. DC/OS allows Wellframe to add services and
technologies with very few people to manage the infrastructure.

DC/OS has allowed us to take on, manage and open up wide areas of business that we
couldnt address before. We are able to expand our business to new geographies and
deliver services that address the complexity of serving patients managing chronic health
conditions.

- Gopal Ramachandran, CTO, Wellframe

42 Mesosphere, Inc.
Architecting for Fast Data Applications

ABOUT MESOSPHERE
Mesosphere is leading the enterprise transformation toward distributed computing and
hybrid cloud. We combine the rich capability you get from public cloud providers with
the freedom and control of choosing your own infrastructure. Mesosphere DC/OS is the
premier platform for building, deploying, and elastically scaling modern applications and
big data services. DC/OS makes running containers, data services, and microservices
easy across your own hardware and cloud instances. Mesosphere was founded in 2013
by the architects of hyperscale infrastructures at Airbnb and Twitter and the co-creator
of Apache Mesos. Mesosphere is headquartered in San Francisco with additional
ofces in New York and Hamburg, Germany. Mesospheres investors include
Andreessen Horowitz, Hewlett Packard Enterprise, Khosla Ventures, Kleiner Perkins
Caueld & Byers, and Microsoft.

43 Mesosphere, Inc.