So what are the top 10 trends we expect to see this year around Big Data? That's what we will get into now. So, the first trend that we will go through is, Big Data becomes fast and approachable. Options expand to speed up Hadoop, and I'll hand things over to Larry to first give us some commentary on this one.
Larry Chiang: Thanks Abhishek, Hi everyone, so this is essentially a two part trend. The first, it is all about why Big Data has become fast and approachable right, and it all has to do with, how users have been perceiving Big Data. To give you some context, when Hadoop was invented by the folks at Yahoo the original goal was to index the internet.
It was designed to be used as a batch processing engine, but you know as it evolved, people realized the power of scaled out storage, the power of processing and clustered workload management, and they forget, you know, it did make a great data analysis platform. So they started adding components that allow data to be stored and queried using common SQL tools in response to a lot of that demand.
 |
View a 2-minute demonstration of InetSoft's easy, agile, and robust BI software. |
Expanded Data Source Connectivity
Here at InetSoft we've continued to expand our connectivity with all of our SQL drivers, but the early SQL drivers were still based on those that produce an algorithm and, which you know, are never great for real-time analytics queries. So SQL Hadoop happened which was still not fast enough which led to continued innovations and optimizations to the execution engines in a Hive on Tez and HIve on Spark.
These have significantly improved speed and as all of this continues to take place, we are seeing that users are increasingly demanding that Hadoop be used by workloads. This is something we hear from our customers. We see this being reflected in a lot of the market trends and research studies.
One that I particularly like is a BI, Hadoop Maturity Survey done by our partners and friends at AtScale, and it's all driven by this movement to expand or expose Hadoop to the business users and not just data scientist. Abhishek or Holly, any comments on that?
Abhishek: Yeah, certainly buddy, I mean, I think from my perspective there's a bit of a data physics I guess you could say, at play here, but to your point around the origins of Hadoop, speed of analytical query was not what the platform was designed for, and the improvements have been made through things like Impala and Spark SQL and Hive on Tez.
And as you mentioned a number of other things have been pretty incredible and it has continued to go really quickly, but it's also been an interesting case were we've seen things that are less typical in InetSoft deployments like OLAP and pre-aggregations and things like that, becoming really very popular in the Hadoop space and that's where the technology we see our customers adopting to essentially index or aggregate the query set to deliver that level of performance.
So they are sort of taking advantage of what Hadoop was best designed to do, originally designed to do and then layering on some other capabilities that really make it work well for interactive real or near real time- query responses. Holly anything, anything you're seeing, you are seeing here from your workout with customers.