Microsoft SQL Server 2012 with Hadoop: Getting SQL Server talking to Hadoop is a smooth process when you follow this tutorial. Learn all the tools and techniques you need integrate the data and then extract powerful business insights from the merged result.

There's a lot of talk about Big Data—estimates are that the total amount of digital information in the world is increasing ten times every five years, with 85 percent of this data coming from new data types for example, sensors, RFIDs, web logs, and so on. This presents a huge opportunity for businesses that tap into this new data to identify new opportunity and areas for innovation.

However, having a platform that supports the data trend is only a part of today's challenge; you need to also make it easier for people to access so that they can gain insight and make better decisions. If you think about the user experience, with everything we are able to do on the Web, our experiences through social media sites, how we're discovering, sharing, and collaborating in new ways, user expectations of their business, and productivity applications are changing as well.

One of the first questions we should set out to answer is a simple definitional one: how is Big Data different from traditional large data warehouses? International Data Corporation has the most broadly accepted theory of classifying Big Data as the three Vs:

Today, every organization finds it difficult to manage and track the right dataset within itself, the challenge is even greater when they need to look out for data which is external to the system. A typical analyst spends too much time searching for the right data from thousands of sources, which adversely impacts productivity. We will move from a world of search to one of discovery, where information is brought to the user based on who you are, and what you are working on. There has never been such an abundance of externally available and useful information as there is today. The challenge is how do you discover what is available and how do you connect to it?

To answer today's types of question, you need new ways to discover and explore data. By this we mean, data that may reside in a number of different domains such as:

You could derive much deeper business insight and trends by combining the data you need across personal, corporate, community, and world data. You can connect and combine data from hundreds of trusted data providers—data includes demographic data, environment data, financial data, retail and sports data, social data such as twitter and facebook as well as data cleansing services. You can combine this data with your personal data through self-service tools, for example, PowerPivot, you can use reference data for cleansing your corporate data with SQL Server 2012, or you can use it in your custom applications.

Existing RDBMS solutions as SQL Server are good in managing challenging volumes of data, but it falls short when the data is unstructured or semi-structured with variable attributes such as the ones discussed previously. The current world seems almost obsessed with social media sentiments, tweets, devices, and so on; without the right tools, your company is adrift in a sea of data. You need the ability to unleash the wave of new value made possible by Big Data. It's all and every bit of data that you should be able to easily monitor and manage regardless of type or structure. That's why organizations are trending to build an end-to-end data platform for nearly all data and easy-to-use tools to analyze it. Regardless of data type, location (on-premises or in the cloud), or size, you have the power of familiar tools coupled with high-performance technologies to serve your business needs from data storage, processing, and all the way to visualization. The benefits of Big Data are not limited only to Business Intelligence (BI) experts or data scientists. Nearly everyone in your organization can analyze and make more informed decisions with the right tools.

In a traditional business environment, the data to power your reporting mechanism will usually come from tables in a database. However, it's increasingly necessary to supplement this with data obtained from outside your organization. This may be commercially available datasets, such as those available from Windows Data Market and elsewhere, or it may be data from less structured sources such as feeds, e-mails, logfiles, and more. You will, in most cases, need to cleanse, validate, and transform this data before loading it into an existing database. Extract, Transform, and Load (ETL) operations can use Big Data solutions to perform pattern matching, data categorization, deduplication, and summary operations on unstructured or semi-structured data to generate data in the familiar rows and columns format that can be imported into a database table. The following figure will give you a conceptual view of Big Data:

Big Data requires some level of machine learning or complex statistical processing to produce insights. If you have to use non-standard techniques to process and host it; it's probably Big Data.

The data store in a Big Data implementation is usually referred to as a NoSQL store, although this is not technically accurate because some implementations do support a SQL-like query language. NoSQL storage is typically much cheaper than relational storage, and usually supports a write-once capability that allows only for data to be appended. To update data in these stores you must drop and recreate the relevant file. This limitation maximizes performance; Big Data storage implementations are usually measured by throughput rather than capacity because this is usually the most significant factor for both storage and query efficiency. This approach also provides better performance and maintains the history of changes to the data.

Hadoop is an open source software framework that supports data-intensive distributed applications available through the Apache Open Source community. It consists of a distributed file system HDFS, the Hadoop Distributed File System and an approach to distributed processing of analysis called MapReduce. It is written in Java and based on the Linux/Unix platform.

It's used (extensively now) in the processing of streams of data that go well beyond even the largest enterprise datasets in size. Whether it's sensor, clickstream, social media, location-based, or other data that is generated and collected in large gobs, Hadoop is often on the scene in the service of processing and analyzing it. The real magic of Hadoop is its ability to move the processing or computing logic to the data where it resides as opposed to traditional systems, which focus on a scaled-up single server, move the data to that central processing unit and process the data there. This model does not work on the volume, velocity, and variety of data that present day industry is looking to mine for business intelligence. Hence, Hadoop with its powerful fault tolerant and reliable file system and highly optimized distributed computing model, is one of the leaders in the Big Data world.

The core of Hadoop is its storage system and its distributed computing model:

MapReduce

MapReduce is a programming model for processing large datasets using distributed computing on clusters of computers. MapReduce consists of two phases: dividing the data across a large number of separate processing units (called Map), and then combining the results produced by these individual processes into a unified result set (called Reduce). Between Map and Reduce, shuffle and sort occur. Hadoop cluster, once successfully configured on a system, has the following basic components:

NameNode

This is also called the Head Node/Master Node of the cluster. Primarily, it holds the metadata for HDFS during processing of data which is distributed across the nodes; it keeps track of each HDFS data block in the nodes.

Note

The NameNode is the single point of failure in a Hadoop cluster.

Secondary NameNode

This is an optional node that you can have in your cluster to back up the NameNode if it goes down. If a secondary NameNode is configured, it keeps a periodic snapshot of the NameNode configuration to serve as a backup when needed. However, there is no automated way for failing over to the secondary NameNode; if the primary NameNode goes down, a manual intervention is needed. This essentially means that there would be an obvious down time in your cluster in case the NameNode goes down.

DataNode

These are the systems across the cluster which store the actual HDFS data blocks. The data blocks are replicated on multiple nodes to provide fault tolerant and high availability solutions.

JobTracker

This is a service running on the NameNode, which manages MapReduce jobs and distributes individual tasks.

TaskTracker

This is a service running on the DataNodes, which instantiates and monitors individual Map and Reduce tasks that are submitted.

The following figure shows you the core components of the Apache Hadoop framework:

Additionally, there are a number of supporting projects for Hadoop, each having their unique purpose for example, to feed input data to Hadoop system, a data warehousing system for ad hoc queries on top of Hadoop, and many more. The following are a few worth mentioning:

Mahout

Mahout is a machine-learning library that contains algorithms for clustering and classification. One major focus of machine-learning research is to automatically learn to recognize complex patterns and make intelligent decisions based on data.

The following figure gives you a 1000 feet view of the Apache Hadoop and the various supporting projects that form this amazing ecosystem:

We will be exploring some of these components in the subsequent chapters of this book, but for a complete reference, please visit the Apache website http://hadoop.apache.org/.

Setting up this ecosystem along with the required supporting projects could be really non-trivial. In fact the only drawback this implementation has, is the effort needed to set up and administer a Hadoop cluster. This is basically the reason that many vendors are coming up with their own distribution of Hadoop bundled and distributed as a data processing platform. Using these distributions, enterprises would be able to set up Hadoop clusters in minutes through simplified and user-friendly cluster deployment wizards and also use the various dashboards for monitoring and instrumentation purposes. Some of the present day distributions are CH4 from Cloudera, Hortonworks Data Platform, and Microsoft HDInsight, which are quickly gaining popularity. These distributions are outside the scope of this book and won't be covered; please visit the respective websites for detailed information about these distributions.

In this chapter, we went through what Big Data is and why it is one of the compelling needs of the industry. The diversity of data that needs to be processed has taken Information Technology to heights that were never imagined before. Organizations that are able to take advantage of Big Data to parse any and every data will be able to more effectively differentiate and derive new value for the business, whether it is in the form of revenue growth, cost savings, or creating entirely new business models. For example, financial firms using machine learning to build better fraud detection algorithms, go beyond the simple business rules involving charge frequency and location to also include an individual's customized buying patterns ultimately leading to a better customer experience.

When it comes to Big Data implementations, these new requirements challenge traditional data management technologies and call for a new approach to enable organizations to effectively manage, enrich, and gain insights from any data. Apache Hadoop is one of the undoubted leaders in the Big Data industry. The entire ecosystem, along with its supporting projects provides the users a highly reliable, fault tolerant framework that can be used for massively parallel distributed processing of unstructured and semi-structured data.

In the next chapter, you will see how to use the Sqoop connector to move Hadoop data to SQL Server 2012 and vice versa. Sqoop is another open source project, which is designed for bi-directional import/export of data from Hadoop from/to any Relational Database Management System; we will see its usage as a first step of data integration between Hadoop and SQL Server 2012.