IBM’s Big Data Platform is fusing Hadoop with the Relational Database approach (among other things…)
Well, we might have just encountered a significant leap forward with IBM’s latest barrage of BIG data solutions. Instead of attacking the big problems of BIG data with somewhat segregated solutions, IBM has instead opted to combine several approaches at once with their new offerings. IBM’s BIG data platform now includes not only a Hadoop approach, but also other interesting analytical tools (each having their own specific uses, of course).
Under the infosphere department heading on the IBM site there is something called “Hive”. Basically, Hive is a Hadoop-laden approach which is designed for those who understand SQL. What makes Hive different however is that it is able to conduct traditional relational database tasks as well as those problems more suited toward Hadoop. Large scale data quandaries tend to gravitate toward Hadoop you see, while other types of relational data (or, data which features parts which are interrelated in various ways) are best organized with something like an RDMS.
One of the most interesting recent developments is “InfoSphere Streams”, which is not only a high-volume data analysis tool, but also one that is supposed to work in real-time. This is going to be a dynamite product for companies who are heavily invested in IT technology and perhaps work in fields where ever-increasing speeds are a top priority (think financial firms and algorithmic trading).
IBM’s BigInsights platform is not only utilizing Hadoop in a more complete way, it is also bringing a coterie of additional tools to the forefront. As previously mentioned, SQL-based relational database technology is prominently featured with BigInsights, but the fun doesn’t end there. Several new analytical tools (accelerators) are making their way into the platform; for example, “BigSheets”, which is a data-exploration interface is being leveraged to apply analysis to various types of social media (Facebook and Twitter mostly). The of course there’s “Data Explorer”, which is poised to offer nearly automated means of scanning multiple types/sources of data.
In what might be the most exciting news emerging from IBM with regards to BIG data, “SmartCloud” (IBM’s premier cloud service) is adding “analytic answers” to its list of tools. What this allows businesses to do is leverage the power of cloud computing along with some very refined analysis methods (which are more-or-less prepackaged). Likewise, because we’re talking about cloud services here, the costs associated with conducting in-depth analysis through various channels will undoubtedly be much more affordable than certain alternatives. But we’re not simply talking about post-analysis here; no, analytic answers is being applied to predictive analysis as well, which adds yet another layer of possibilities to the mix, doesn’t it?
Then of course we have “Unica”, which IBM recently acquired. In a nutshell, Unica can be either a cloud service or a software deployment, and it’s designed to assist in data analysis and aggregation. But what’s really interesting is how Unica is being used in IBM’s “Big Data Solution for CMO’s”. These new types of offerings are bringing the idea of a more readily useable form of BIG data analysis service to the forefront as opposed to those which require heavy configuration.
It seems that most large organizations which are investigating and researching BIG data problems are branching out in various ways right now. IBM it seems is one of the current leaders when it comes to fusing BIG data solutions and approaches. It’s difficult to see anything negative stemming from these recent developments, after all, combining various approaches and refining the existing ones while searching for newer solutions are the only things that can really be done. However, it is extremely refreshing to see that several areas of BIG data analysis, which have long been thought to be somewhat incompatible, coming together.
Likewise, it would be an enormous mistake to not utilize cloud computing in BIG data analysis, especially considering its many positive attributes. On the immediate side of things though, it’s great that someone is finding ways to bring together relational database approaches with that of Hadoop, given that they’re (arguably) the most popular/valued frameworks for dealing with BIG data. If IBM continues along this development path we might see plenty of additional packaged solutions for BIG data emerging in the very near future.
BIG data certification can be yours for a small price, but it offers very large long-term rewards…