Friday, April 12, 2013

Machine data and the Internet of things

It was interesting to see a recent announcement where two companies have teamed up to provide machine data analytics solution to the market. They should thank me for free press by posting this link here :-) However, I want to thank them for increasing the awareness in the market for such an incredible solution, an area that Glassbeam team knows really well for last few years since the day we launched our solution to the market (in 2009).  The only catch, and a big one, is that approach mentioned in this new announcement is largely professional services driven. In contrast, Glassbeam is cloud based product driven solution and was founded on the premise that any product company will have a huge challenge in assimilating intelligence from its machines by looking at the log data.  

Our core IP of Semiotic Parsing Language (SPL) came about from this very need to handle the complexity of multiple log formats that keep pouring in at increasing frequency and need to be parsed almost at near- real time to make information accessible across different business units (product marketing, management, engineering, sales, services and support).  As a result, SPL innovation from Glassbeam today provides almost 10x process efficiency in creating and maintaining an on-going machine data analytics project inside a product organization.  Proof is in the pudding where in last few years, we have successfully deployed this cloud solution with some of our blue chip accounts such as IBM, EMC, HDS, Aruba networks, Polycom, Fusion IO, etc.  See this earlier blogpost where I talked about the use cases from one of the leading companies that has been an avid user of our cloud based applications 

Another key point to highlight here is the fact that “building” such a solution in-house, even with help from leading technology and service companies, is a monumental undertaking for many product companies whose core competency is to build their core products and not build internal business applications as adhoc IT projects for business needs.  And I consider a machine data analytics as a new class of Big Data business applications whose core input is log data from a product manufacturers’ installed base.  It is quite easy for these companies to underestimate the complexity, size and on-going investments needed to sustain efforts in collecting, storing, parsing, analyzing, and reporting this intelligent output from machines into their support, engineering, marketing, sales, and services organizations.  See this video on our take on “build vs buy” for just this solution of machine data analytics. What you will see in this video is really a true story that we watched unfold at a leading storage company over a number of years.  This story is not uncommon and therefore a gentle reminder to anyone who is thinking of building this solution on-house.  Here is the link... enjoy!


No comments:

Post a Comment