When companies use new technologies to rethink their business processes and serve their customers better, there may be some hurdles along the way. It can be difficult to build new applications that access the data they need quickly enough to serve business needs.
Many of these problems can be solved through the benefits of Hadoop 3.0, the latest version of the open-source distributed computing and data storage system. Here’s how.
The lifeblood of digital transformation is data. The analytics capabilities that fuel digital transformation rely on ubiquitous access to data. You can’t mine multiple data sources for new insights unless you can aggregate and reference those sources together.
This data costs businesses huge amounts to acquire, so when they have it, they want to utilize those digital assets as much as possible. Today, your data might be used to build a 360-degree view of your customer interactions. Tomorrow it could be used to build a machine-learning model that analyzes buying signals to deliver real-time personalized promotional offers for customers whether they visit an online or traditional shop.
The more that your business can use acquired data across multiple business use cases, the more valuable it becomes. From a technical point of view, the key to unlocking this value lies in making the data available for as many workloads as possible.
Unfortunately, enterprise IT architectures haven’t always supported this reuse. Individual IT projects have traditionally evolved independently of each other, on a piecemeal basis. This has created pockets of siloed data, locked up in best-of-breed applications across companies. One of the benefits of Hadoop 3.0 is that it can help with this experience—thanks in part to its new data storage approach.
Hadoop 3.0 will allow companies to store and analyze data more efficiently, squeezing up to twice as much business information into the same capacity as before. This is significant because it enables companies to keep their valuable corporate data on hand in extremely fast storage media, which in turn boosts application performance.
Instead of having to restore archived data from slow storage media, companies can keep large quantities of data in a single data lake, making it quickly available to any applications that need it.
Instead of having to create a new, specially configured data set each time you develop and deploy an application, Hadoop 3.0 enables you to go back to the same data watering hole with different workloads—and the workloads themselves will be quick and easy to deploy. Of course, this is only useful if the applications themselves are easy to deploy, and that’s where another of the benefits of Hadoop 3.0 comes into play.
Traditionally, running multiple applications and integrating them to work together has been difficult for enterprise users. Applications have their own dependencies and requirements, making it hard to configure and set them all up.
Containers change all that. They take applications and effectively put them in a software-defined box that takes care of those quirks, presenting the outside world with a single, easily accessible interface to use that application.
These containers can be bolted together like Legos to create services that make sense to business users, such as machine learning for analytics. By introducing combinable containers into Hadoop 3.0, Hortonworks makes it possible to build applications quickly. A container-based service can often be rolled out in minutes, whereas a business service may have taken weeks to create using a traditional application infrastructure.
Not only can you quickly build new business services to exploit your corporate data, you can run multiple versions of the same application. This enables you to rapidly iterate and create new features by developing and testing new versions of business services without disrupting the old ones. Because everything’s in a container, applications need not interfere with each other.
There are many other benefits of Hadoop 3.0, but one in particular will appeal to those using machine learning or other artificial intelligence techniques: the ability to specify which type of computer processor a business service runs on.
In artificial intelligence, specialist computer chips called graphical processing units (GPUs) and field programmable gate arrays (FPGAs) can provide far higher performance than traditional central processing units (CPUs). In Hadoop 3.0, for the first time, administrators will be able to allocate specific jobs to these processes, making far better use of their computing infrastructure for machine learning analytics workloads.
Not only will companies be able to make better use of their data, they will be able to make more efficient use of their computing infrastructure. Hadoop 3.0’s benefits span both hardware and software, and it promises to supercharge your digital transformation in 2018.
Now is the time to find out more about the benefits of open source technologies.