hadoopsummit

Its not just a chess board, but a 3D chess board
- Peter Goldmacher @petergoldmacher f/ #Aerospike on #theCUBE
#Hadoop

What I find so exciting, is that these technologies are enabling new new things
- Peter Goldmacher @petergoldmacher f/ #Aerospike on #theCUBE
#Hadoop

When you look at a lot of the technologies, a lot are lower cost alternatives for proprietary software. A lot of it is reaction to what has been 20 years of abuse of the community by vendors - #Aerospike @petergoldmacher on #theCUBE #Hadoop.

@JeffFrick Thats a great quote! @PeterGoldmacher

Companies like MongoDB and Couch – their primary message is we’re an Oracle alternative. The new technology is often better for the applications users want to build - #Aerospike @petergoldmacher on #tdheCUBE.

One thing I like about being at Aerospike is rather than we can help you do old things better, most of the applications are in new areas like Ad Tech - #Aeerospike @petergoldmacher on #theCUBE

One of things I like best about being at @aerospikedb is that we can help you do old things better. For Example adtech -@PeterGoldmacher

you beat me to it ;)

Speed at Scale
- Peter Goldmacher @petergoldmacher f/ #Aerospike on #theCUBE
#Hadoop

. @BertLatamore commodity mkt speed is the advantage

It's not clear to me that other than Red Hat the Open Source model is commercially viable for more than a very small number of companies. #Aerospike @petergoldmacher on #theCUBE.

@petergoldmacher speaking the truth - most of these open source co's contribute 99% of the code to "their" open source project

"If i build it...they will come" - or will they?

Most of the big raises turn out that most of the dough goes into sales and marketing

Goldmacher says "Cloudera / Hortonworks is 'tomAto tomahto' - not sure which one wins...they are both interesting" - I agree

.@furrier correctly points out the issue with in-memory ram vs flash is persistence

@petergoldmacher stressing that Aerospike is optimizing for flash/ssd

What Spark does because it runs in memory you can put more compute behind it & do what you normally do in Hadoop faster. #Aerospike @petergolmacher on #theCUBE

What we do is we have a different kind of database in SSD. Not doing a lot of traditional database operations. Instead of 20%-30% performance improvement we talk about 100% improvement. #Aerospike @petergoldmacher on #theCUBE

We are not an Oracle alternative -- #Aerospike @petergoldmacher on #theCUBE

@petergoldmacher says there are generally two types of co's: 1/ great product, shitty marketing or 2/ shitty product great marketing - he sees Aerospike as the commercial version of websalers

@dvellante we should talk about that ram vs flash thing :)

The more systems you have to support, the more complexity, and complexity kills
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

One exciting thing now is the power of consolidation. This only opened up recently. The systems for ticker reporting is different from the systems for historical analytics. They don’t have to be separate any more - Bloombert's Matthew Hunt on #theCUBE

We're seeing more opportunity to consolidate on fewer systems and still deliver breadth of functionality
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

The increasing maturity of frameworks for distributed systems allow more commercial and opensource movement up the stack
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

Spark is a distributed computing platform that unifies a number of different functionalities
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

Bloomberg has nearly 4,000 engineers. We want to abstract as much of the infrastructure complexities away from the developers as possible. -- Matthew Hunt @instantmattheew on #theCUBE

You still need storage and a database and a way to do computation
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

When you have 10,000 machines & they are changing all the time, you need a way to express infrastructure in a repeatable way. Bloomberg's Matthew Hunt @instantmatthew on #theCUBE

All these things are classic enterprise software hardening
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

It's about making things work in production, in reality.
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE

You can go through a @Docker tutorial in a day.
- Matthew Hunt @instantmatthew - Bloomberg on #theCUBE
#Hadoop

Where all this is heading & how it will come together is much clearer to most people than it was a year ago. Bloomberg's Matthew Hunt @instantmatthew on #theCUBE

"We are seeing an era of innovation here that is like the railroad," John @Furrier on #tueCUBE

Hadoop is clearly growing, Hadoop will clearly be standard., John @Furrier on #theCUBE

@ggilbert41 talking about the lack of solutions. Still opportunities for maturity.

Ultimately we will get to the point where Hadoop is a utility like railroads, like electricity. We need to rationalize all the Apache projects so practitioners can have one platform, not a collection of projects. Wikibon's George Gilbert @ggilbert on #theC

One caution: Learn from others and avoid the severe fragmentation of solutions.

Wandisco has this data fabric. It could be in Hadoop, an RDBMS, EMC storage, & we will make it so you don’t care where it is. Wikibon's George Gilbert @ggilbert on #theCUBE

The relationships are huge. We are seeing people who left early adopters and now are in new startups. It takes a while for these things to go - theCUBE Host Jeff Frick

Too many focus on Amplification vs. Quality in the conversation. #socialmedia

A big question two or three Hadoop Summits ago was when will we see the applications. It feels like we are getting close to that - #theCUBE Host Jeff Frick

Amara's Law:
We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.

Enterprise needs a balance of choice: Not 1-size-fits-all, and not fragmented.

At the end of the day, tech is just an enabler. It's about business value.

Yahoo uses it, Google uses it, it works. Let's get past the question of does Hadoop work. - SiliconAngle's Jeff Frick on #theCUBE.

The relational databases may go away, but SQL will still be here. We need full SQL support. Infoobjects CEO Reshi Yadev on #theCUBE.

Big Data projects are multi-year because more and more old stuff is moved to Hadoop and more new analysis are added and insights come from it - Infoobjects CEO Reshi Yadev on #theCUBE.

Data Driven Solutions from@meditativesoul

@JeffFrick Question: Is Spark HANA for the rest of us?

@jefffrick Question: What are some of Infoobjects customers? Is it working with large enterprises?