Comments on: What it means if Yahoo Hadoop spinoff doesn’t do distribution

By: Waqas Khan

Waqas Khan — Sun, 24 Jul 2011 10:37:23 +0000

Hi Andy — Thank you for your comment. The principle remains the same for organizations of any size. My suggestion would be to start by listing your existing channels of communication with your audience. Then consider which would be the most practical to test the addition of an opt-in request. Ind Tv Forum

By: Anand Babu Periasamy

Anand Babu Periasamy — Sat, 23 Jul 2011 07:09:04 +0000

Hartonworks is certainly a bad news for MapR and Cloudera. Others have to substantially differentiate with their proprietary addons or get acquired soon. There is no place for 3rd or 4th best.

By: Curt

Curt — Sat, 23 Jul 2011 06:06:42 +0000

I’m not sure I agree. Cloudera does provide some value by building RPMs and Debian packages. They are a services company, not a technology/product company, but that’s a legitimate business. We used CDH at our company for the last year. After the Hadoop Summit we tried MapR, and I’ll admit I can’t imagine going back to CDH (I’m not sure why the article mentions vendor lock-in, since it’s trivial to move your apps from CDH to MapR and vice versa). We just finished migrating our production from CDH to MapR (the M3 edition for now). Having full read/write NFS access is a game-changer for us. IMHO, the folks at MapR seem to be years ahead of anything else that’s out there.

By: Bill

Bill — Fri, 22 Jul 2011 23:26:16 +0000

Not sure about HortonWorks but Cloudera is a joke of a company. Greatest scam ever pulled on the investor community in recent time.

By: Jack Norris

Jack Norris — Fri, 22 Jul 2011 23:24:16 +0000

The fact that Hortonworks will focus on improving Apache Hadoop and provide services is good for the entire industry. This is not bad news for MapR. In an earlier article you pointed out that MapR understands it’s not just the sum of its parts — the project is, in fact, all about parts. We’ve worked hard to improve many of these parts and include these in a complete distribution for Apache Hadoop. These innovations provide tremendous customer value in the form of greater performance, ease of use and dependability. Our model is not that different from that of other commercial vendors that provide a distribution. Although instead of focusing solely on management tools that are “proprietary”, MapR has provided many more underlying innovations. Providing these innovations does not automatically create vendor lock-in. In fact, one of our innovations is the ability to mount a Hadoop cluster via NFS. We’ve had customers tell us specifically that it makes it far easier to get data into and out of MapR clusters. Customers are looking for innovations with a focus on API compatibility so that MapReduce programs run without any required changes.

Jack Norris VP Marketing, MapR Technologies