June 23, 2013
Hadoop news and rumors, June 23, 2013
Cloudera
- Cloudera changed CEOs last week. Tom Reilly, late of ArcSight, is the new guy (I don’t know him), while Mike Olson’s titles become Chairman and Chief Strategy Officer. Mike told me Friday that Reilly had secretly been working with him for months.
- Mike shared good-sounding numbers with me. But little is for public disclosure except the stat >400 employees.
- There are always rumors of infighting at Cloudera, perhaps because from earliest days Cloudera was a place where tempers are worn on sleeves. That said, Mike denied stories of problems between him and COO Kirk Dunn, and greatly praised Kirk’s successes at large-account sales.
- Cloudera now self-identifies pretty clearly as an analytic data management company. The vision is multiple execution engines – MapReduce, Impala, something more memory-centric, etc. – talking to any of a variety of HDFS file formats. While some formats may be optimized for specific engines – e.g. Parquet for Impala – anything can work with more or less anything.*
- Mike told me that Cloudera didn’t have any YARN users in production, but thought there would be some by year-end. Even so, he thinks it’s fair to say that Cloudera users have substantial portions of Hadoop 2 in production, for example NameNode failover and HDFS (Hadoop Distributed File System) performance enhancements. Ditto HCatalog.
*Of course, there will always be exceptions. E.g., some formats can be updated on a short-request basis, while others can only be written to via batch conversions.
Everybody else
- There’s a widespread belief that Hortonworks is being shopped. Numerous folks – including me — believe the rumor of an Intel offer for $700 million. Higher figures and alternate buyers aren’t as widely believed.
- Views of MapR market traction, never high, are again on the downswing.
- IBM Big Insights seems to have some traction.
- In case there was any remaining doubt — DBMS vendors are pretty unanimous in agreeing that it makes sense to have Hadoop too. To my knowledge SAP hasn’t been as clear about showing a markitecture incorporating Hadoop as most of the others have … but then, SAP’s markitecture is generally less clear than other vendors’.
- Folks I talk with are generally wondering where and why Datameer lost its way. That still leaves Datameer ahead of other first-generation Hadoop add-on vendors (Karmasphere, Zettaset, et al.), in that I rarely hear them mentioned at all.
- I visited with my client Platfora. Things seem to be going very well.
- My former client Revelytix seems to have racked up some nice partnerships. (I had something to do with that. :))
Categories: Cloudera, Data warehousing, Datameer, Hadoop, Hortonworks, IBM and DB2, Intel, MapR, Market share and customer counts, Platfora, SAP AG, Zettaset
Subscribe to our complete feed!
Comments
11 Responses to “Hadoop news and rumors, June 23, 2013”
Leave a Reply
Curt, SAP is actually pretty clear on their integration with Hadoop. Think of HANA as a data virtualization engine, which can join, cache, etc any data sources, including HDFS. Cheers!
Thanks, Boris.
So how does Sybase IQ fit into the picture?
Hi Kurt,
I feel it’s worth pointing out that Parquet is not custom to Impala — in fact, at the moment Parquet is in use at Twitter and Impala is not (we have a lot of nested data structures, which Impala doesn’t support yet). Impala certainly works better on it, but so does Pig, Cascading, etc. There is nothing Impala-specific in Parquet.
Intel, really? I would’ve expected MSFT.
Tempers on our sleeves, Curt? I always thought that was openness and transparency ;-).
We started shipping the Hadoop 2 code last year, despite the fact that YARN was then (and arguably is now) still too fragile for production workloads. We went to a lot of trouble to back-port MapReduce 1 to the release so that our customers could get the HA namenode, performance improvements and other goodness in the 2.0 trunk release, and still have production-grade MapReduce available. We frankly expected more experimentation with the YARN (aka MR2) code than we’ve seen, but I am betting that trial deployments and test workloads will start in the wild in 2H13.
[…] Jun 24, 2013 Hadoop news and rumors, June 23, 2013 by Curt Monash Go to Source (Comments & likes on the source site, thank you) […]
Dmitry,
Microsoft and Teradata are currently Hortonworks’ 2 best friends. An ideal deal would preserve both those relationships.
—————————-
Meanwhile, a subsequent rumor has, in its most popular form, Hortonworks raising $50 million at a $500 million valuation. If that’s exactly correct, then a $700 million near-term buy-out price is too low.
[…] for a few hundred million dollars a while back. The most recent rumor is new Hadoop entrant Intel offering $700 million for Hortonworks. Think about that: two companies who’ve faced antitrust lawsuits in recent memory trying to […]
[…] same goes for my recent Hadoop […]
[…] June, the company announced a $50 million series C round of venture capital, which, it is rumored (according to database industry analyst Curt Monash) valued the company at $500 million. That was the first time Hortonworks had publicly announced how […]
[…] company announced a $50 million Series C round of venture capital, which, it is rumored (according to database industry analyst Curt Monash) valued the company at $500 million. That was the first time Hortonworks had publicly […]
[…] Hortonworks denies advanced acquisition discussions with either Microsoft and Intel. Of course, that doesn’t exactly contradict the widespread story of Intel having made an acquisition offer. […]