January 28th, 2016

B23 Data Platform is a new and innovative software solution that allows end-users to launch complex “Big Data” distributed processing clusters with a few clicks of the mouse. We call these distributed processing systems Stacks. Examples of Stacks include Hadoop, Kafka, Storm, Spark, Zeppelin, Metron, Elasticsearch, H2O, etc. Using B23 EasyIngest capability, these Stacks can be automatically populated with user-selected data sets during the provisioning process, eliminating the need to manually build a data pipeline. The entire process from start to finish is a few minutes long. B23 Data Platform can launch one or many different stacks simultaneously, all ingesting similar or different sets of data in a secure Space in the Cloud. Terminating Stacks is just as easy. During a recent sales meeting demonstrating B23 Data Platform, a potential enterprise customer told us that he was having an emotional experience and that he had “dreamed of something like this for a long time.” Well, we’ve been dreaming about this for a while too! This is one of many positive reviews since we launched our private beta only a few weeks ago. We are so excited to launch B23 Data Platform to a broader audience to share what makes this software solution so special.

Along our journey over the past several years almost every one of our customers expressed a common desire which was “We hired really talented analysts/data scientists/quants to analyze our data to derive strategic insight, and too often they are stuck in the Hadoop cluster fixing some configuration or system setting. How do we fix this?” B23 Data Platform was designed and built by analysts for analysts. Ironically, very few of our “developers” are traditional computer scientists although we all share a passion for software development and data.

Customers have been asking for a repeatable deployment methodology, consistent “look and feel”, and a user interface that “boiled down” a seemingly endless number of configuration parameters to choices that intuitively made sense to analysts. Most important, it had to be intuitive and easy — better yet, free. Before B23 Data Platform, even the most progressive user interfaces were cluttered with buttons, widgets, dials, wizards, and text boxes. Hopefully, if you were lucky, you could avoid JSON or YAML along the way. If you’re not a systems engineer or developer with years of experience in Big Data tooling, you were facing an uphill battle.

We call the secure place in the Cloud a customer’s Space. A Space contains the user’s Stack and ingested data. A Space is a secure, bounded, virtual private cloud with enhanced security features that embodies years of knowledge migrating hundreds of applications to the Cloud. We know that securing the Cloud is very different than securing traditional on-premise networks and infrastructure. When done right, the Cloud can be more secure. When done wrong, it can be a huge disaster and incur significant financial and reputational repercussions. This knowledge of building secure Cloud infrastructures cannot be conveyed in a one-hour seminar in Las Vegas, but from thousands of hours of implementing Cloud security parameters that have stood the test of time.

The Cloud, specifically its elastic nature, has made it a great pairing for operating large-scale distributed systems. Static, non-elastic clusters are typically built and designed for a specific capacity, so they are almost always underutilized or at peak usage. B23 Data Platform supports Stacks that are elastic by default. Ironically, outside of B23 Data Platform, most Cloud-based distributed systems are not very secure or elastic, which makes us scratch our head why they exist in the Cloud in the first place.

Ingesting data into a system is hard. Building data pipelines is a full-time business for a lot of people and organizations. That’s why we have included EasyIngest in B23 Data Platform. EasyIngest allows B23’s customers to build repeatable data pipelines to automatically ingest data as part of the Stack provisioning process, or later after a Stack is running. EasyIngest doesn’t solve every data acquisition requirement, but it certainly helps our customers overcome the most common ones.

B23 Data Platform provisions Stacks and Spaces via secured interactions in your accounts — not ours. Those interactions are logged in your accounts, not ours. Dashboards are provisioned in your Space and applied to your data with your access controls. Very few managed Big Data platforms can make a similar claim as ours.
We don’t believe in the “just trust us we can analyze your data better than you” approach. B23 Data Platform applies your analytics to your data. Proprietary machine learning algorithms require customers relinquish control of their data to external parties. This approach places too much value on the vendor’s analytics, and not nearly enough on protecting customers’ data. Algorithms are being commoditized so rapidly that we believe it’s the customers’ data that is more valuable, and should be protected at all costs.

What started out as a convenience is now a clear strategy for customer lock-in and vendor monetization. Packaged distributions are so complex these days that its difficult for the vendors themselves to keep track of specific components and versions, much less customers. Gaps and overlaps exist between all of these vendor distributions. So how do customers know which one to pick, and does it even matter which one they pick? Then there’s the one-off’s that everyone wants (Elasticsearch) that are nowhere to be found in these distributions. Customers ultimately want choice, and not get locked into a one vendor’s distribution and their choices. B23 Data Platform allows you to choose your own Stacks and make your own distributions.

We look forward to more customers joining us on our journey and helping shape our future roadmap. What Stacks would you like to see next in B23 Data Platform? We encourage everyone to try B23 Data Platform for free by signing up at https://platform.b23.io/.