#hadoopworld - CrowdChat

HadoopWorld Chat

We will be having a pre-event chat with thought leaders prior to #hadoopworld for #BigDataNYC event

#hadoopworld @furrier @dvellante @jeffreyfkelly

LeaderBoard

Stream Ended

Establishing a secure connection

Let's talk #BigData ROI - @Wikibon analysis shows enterprise early adopters only getting 50+ cents on the dollar in return so far - are these findings in line with what you're seeing in the market and what do you think the reasons are?

[+] Show Hidden Comments

yes, reason = no apps! ;-)

Jeff Kelly definitely - but related what about enterprises investing in #hadoop with no real killer use case, just loading up the cluster with lots of data with nothing to do.

I think these #'s are reasonable and encourage you to do more of this research. One reason, Hadoop related, Business Unit rolls out Hadoop cluster and sees performance gain, then says, "what else?" see, apps......

re: loading up the cluster, yes, it's performance related today but then the BU asks, "Where are the dashboards, where are the reports, I need to be able to share this with non-techies." Inflection point that will effect ROI over time.

What's the distribution of ROI though? Some people are seeing a significant return while others are still in the investment phase?

Jeff Kelly yes most enterprises still in PoC phase so ROI difficult to calculate, but those enterprise that have scaled-out deployments challenge is keeping clusters performing even when a killer app is identified

Michael Hiskey5

Inflection point: the balance of $$$ (spend) going into #BigData market today far exceeds $$ Revenue coming out

[+] Show Hidden Comments

true - both for enterprises looking to exploit data and mega-vendors looking to lay groundwork for lucrative #bigdata practices

Real-time recommendations for businesses using agorithms, and design is a hot topic..what is the state of the art..thoughts

[+] Show Hidden Comments

#Rstats is where the movement is, this spells trouble for #SAS, they have been a long time on the list of "vendors I am most loyalty to that I can't wait to stop paying' list - for decades!

Jeff Kelly is SAS not adapting its business model to the open source #bigdata world fast enough or are open source tools like R just that good?

State of the art in recommendations is applying algorithms in real-time, users interact and the underlying indicators change.

John Furrier is wibidata doing anything specific in verticals?

considering app landscape is still pretty barren, what verticals r in best position to capitalize on #BigData in the short-term - or is it not a question of verticals? Maybe the early winners are those enterprises with killer #BigData internal talent?

[+] Show Hidden Comments

i think verticals have to be in the equation as if we determine winner by enterprises with killer #BigData talent only, that's a really small sample size and makes the ROI discussion even tougher....

transportation, oil and gas, media and entertainment, financial services, telco, healthcare not in any particular order

Are there any open-source schema management projects for #HBase #Hadoop? Any users running them at scale speaking at #hadoopworld2013?

[+] Show Hidden Comments

bumping it to @otrajman http://www.kiji.org/ from #wibidata looks like something which does schema management.

Schema? we don't need no stinking schema! ;-)

Take a look at the open source @kijiproject at http://www.kiji.org/ with not only schema management but also a powerful language for modeling data and real-time model scoring.

#Facebook's new Real-Time ETL Morse is getting introduced at #hadoopworld. Why does the ecosystem need yet another #ETL pipeline for #Hadoop? There is Flume, Kafka, Scribe etc. How does Morse differ?

[+] Show Hidden Comments

think it says there is still a need - when one begins to dominate it will become the standard.

real-time ingestion is critical to #BigData apps executing on decisions at the right time (aka"now")

Bill Peterson agree, what does Chris Lynch call it, "in-time" i've stolen that one ;-)

Brad Bokal could not agree more...

When will the Hadoop Skills shortage begin to shrink?
Can it catch up to the adoption growth?

[+] Show Hidden Comments

going to take time - university #datascience programs a good start but in the shorter term need to train existing biz analysts and IT, plus make tools easier to use and abstract away complexity

Speed of adoption is a function of availability of skills

John Furrier agree adoption drives value ..value drives profit ..drives demand for skills.. I love this #bigdatamarket #BigDataNYC

Michael Hiskey2

As a #BigData business HQ'd in #NYC, We are VERY excited about #BigDataNYC and #HadoopWorld fanfare this month!

[+] Show Hidden Comments

Anticipate any big surprises for the last week of Oct in NYC?

[+] Show Hidden Comments

the Giants will win a game? (Oh, couldn't resist, #Patriots fan)

Michael Hiskey Didn't the Pats have like 18 wins one season recently and one GIANT loss?

Jeff Frick Ouch (Michael's reply)

should see some big announcements but really looking forward to talking to #BigData practitioners on #theCUBE at #BigDataNYC at Warwick Hotel

I think that big data bus vs community will be a big conversation

Jeff Kelly biz people want solutions to their problems, don't care about the underlying tech - open source community and biz community need to get better aligned

Put another way, why aren't we seeing more application providers building solutions on top of #hadoop?

[+] Show Hidden Comments

evolution of the market? Hadoop connectors first because it was easy, real API's next, dashboards and visualization and then real apps? thinking out loud

apparently we haven't caught up recently enough