Cloudera will one day become a public company, but unusually won't do so with the primary aim of finding working capital, according to co-founder and chief strategy officer Mike Olson.
Speaking to The Reg in Sydney today, Olson said the US$740m Intel invested in the company means it has all the cash it needs to get where it wants to go. The company still plans to one day go public, but for now is content to be able to work on long-term projects rather than conform to public investors' insistence on quarterly financial data dumps.
So why go public at all? Olson feels there will one day be advantages to doing so, perhaps including giving early investors a way to solidify profits.
For now Cloudera is enjoying its close partnership with Intel, as access to Chipzilla's roadmaps and technologies lets the big data company improve its offerings. Encryption of all data, a chore made possible by some on-Silicon features in Intel chipery, was recently added to Cloudera's version of Hadoop with an overhead of just one or two per cent of processor resources. Doing so delivers on pledges made by Hadoop Daddy Doug Cutting to improve Hadoop security.
Another thing Cloudera's Intel-derived cash stash enables is acquisitions and Cloudera made one last week, scooping database log files analytics outfit Xplain.io. Olson said the company is comprised of “database geniuses” from IBM, Oracle and others and that its wares are already widely used by Cloudera customers, who drew it to the company's attention.
Cloudera plans to build on Xplain.io's self-serve analytics offering to help users manage its Hadoop distribution.
Olson also discussed trends in Hadoop adoption, which he said now splits along lines of data's origin. The likes of VMware keep making plays for Hadoop users, arguing the it's better to spin up a temporary cluster, or scale a cluster, with virtual machines rather than buying hardware dedicated to a single purpose. the
Cloudera offers just such hardware in the form of Handoop-optimised, and Intel-equipped, "pizza box" servers. Olson acknowledged that not all on-premises users want dedicated clusters, but said a more marked split can be observed by looking at the place data was created.
Applications that generate data on-premises, he said, spawn on-premises big data infrastructure. Those running cloudy applications tend to do their analytics in the cloud.
Users just don't move data between the two environments, he said, leading Cloudera to treat conventional and cloudy users with equal affection. ®