Blogs

Recap of IBM Twitterchat: Mobile Data - Taking Your Big Data on the Road

Post Comment
Big Data Evangelist, IBM

Smartphones and other mobile gadgets have become integral to every aspect of modern life. So it’s no surprise that enterprises everywhere are starting to tap into them as a rich source of data for deep analysis in Hadoop, NoSQL and other big-data platforms. By the same token, mobile devices are increasingly being used to access online services powered by big data.

On Wednesday, March 13, IBM and Wikibon/SiliconANGLE sponsored a Twitterchat focusing on the growing intersection of mobility with big data. Using hashtag #bigdatamgmt, we dissected this growing trend from many significant angles. And I was on the chat, of course, tweeting, fittingly enough, from an extremely mobile commercial airline 35,000 up in the troposphere.

What follows are the key points raised by the various tweeters, airborne and otherwise, in response to the key questions. I’ve removed some hashtags and corrected a few obvious typos, but I’ve left it all in its native Twitterese for authenticity:

Which enterprise initiatives require mobile access to big data?

Mobile access is important for many enterprise applications, said the tweeters, with a clear focus on customer-facing and other front-office initiatives.

Jeff Kelly (@jeffreyfkelly) of Wikibon/SiliconANGLE said “marketing for sure - consumers are mobile and so must marketing campaigns,” adding that “failure to take mobile data into acct in marketing, you’re missing possibly most important data related to consumer .” He also noted that “social & mobile closely linked - consumer creating data via social apps on mobile devices.”

IBM-er Susan Visser (@susvis) said “customer support seems like a must have on #mobile.” Another IBM-er, Natasha Bishop (@Natasha_D_G) said “Sales & service depts. need customer insight at right time, at right place which may be on road.”

Several third-party influencers noted that mobile big-data access is important in many industries and business initiatives. Craig Mullins (@craigmullins) stated that “marketing and sales initiatives are ripe areas for #Mobile #bigdata support...but they are not the only ones: retail, health care, finance, manufacturing, etc. etc.” Cristian Molaro (@cristianmolaro) said “almost any enterprise initiative could benefit from #bigdata in a “small” #mobile package.” Alex Philp (@BigDataAlex), added “advertising, marketing, field services, law enforcement, land management, infrastructure management, all require mobility” and “real estate” be added to the list of industries and initiatives where mobile access to big data is a key requirement.

Philp also noted that it’s important to “combine mobility with location analytics - where you are is changing how we buy, how we work and how we play.”

And I tweeted that mobile access is pivotal to “any enterprise decision automation initiative supporting mobile employees with analytic-powered apps #bigdata. It’s also essential to any big-data “app that aggregates & analyzes ‘ambient’ data from smartphones & uses it to optimize mobile experience.”

In essence, analyst Doug Laney of Gartner (@doug_laney) underlined that latter point by stating that “#mobile *access* to #BigData a lot less interesting or valuable than as a source of bigdata.” Laney seemed to be construing “*access*” within the context of this tweet as referring to mobile user access to applications, as opposed to the broader meaning of mobile infrastructure access to the data generated by devices. He stressed that “#mobile = comparatively restricted platform for info delivery, but entirely new source of data collection.”

Which users, developers, stakeholders require mobile access to big data?

Mobile access is a priority for pretty much everybody, the tweeters agreed.

Kelly said it is essential for “any worker in the field making decisions on front line need mobile access to insights (not nec analytics tho).” However, he stated that user “mobile access to #BigData doesn’t need to be full functional analytics but insights based on situation, location.”

Bishop stated that “in an always connected world of instant, we’re quickly moving 2 ALL needing #mobile access to #bigdata.” Mullins emphasized that the priority users who might benefit from mobile access to big data are “folks who travel, work remotely, or require flexibility.”

I noted that mobile big-data access is necessary to any user who “requires mobile access 2 decision support/automation powered by massive multistructured data sets.” Where developers are concerned, I noted that “any data scientist or SME developing, collaborating, or tweaking statistical models from wherever” needs mobile access to big data.

What are the security and privacy risks of providing mobile access to big-data apps?

The security and privacy risks on mobile access to big-data are not different in kind from the risk of providing mobile access to smaller-scale data platforms.

However, as I noted, the “privacy risks can B considerable, considering sheer comprehensiveness of personally identiable info in #bigdata...same risks as with any mobile data access: theft, loss, eavedropping...but MORE data is at risk with #bigdata.”

Kelly concurred, tweeting that this new data-gorged environment puts “more data at risk and potentially more sensitive data....security, privacy risks are similar to traditional data and #mobile, but stakes are higher w #BigData.”

Kelly’s colleague John Furrier (@furrier) noted that “identity & authentication are key to mobile security - same old issue with new architecture & new software models.”

Does mobile big-data require different retention policies?

Mobile devices generate a growing volume of data, a trend that directly impacts enterprise data-retention practices.

Fundamentally, data-retention practices must comply with a wide range of company policies and external mandates. According to Molaro, “retention should be a business variable not a access device matter.”

Nevertheless, as Kelly noted, “mobile data growing exponentially, so how much is retained and for how long gets tricky.”  As tweeter Allan Koivo (@AllanKoivo) said, “[retention] policies always need to be dynamic to reflect new technologies.”

According to Mullins, “#bigdata does not necessarily change retention reqmts but it complicates the issue.... ....You may choose to retain more data for mobile  but be careful because data once retained is discoverable during court trials.” “

“Retention for mobility is a key factor,” tweeted Philp. “Do we really need to remember every location we have been - It links back to privacy....The question is who is doing the retaining of the data.”

I argued that “mobile #bigdata makes imperative 2 have enterprise rights mgmt policies 2 prevent multi-device over-retain/leak.” I also put forward the position that “mobile doesn’t impact retention policies at server – but should @ client. Keep sensitive info on device limited.”

Which type of big-data platform (Hadoop, stream, etc.) is best for mobile clients?

Most, but not all, tweeters agreed there’s no pat answer on which platform is best for mobile access to big data.

Bishop said “should start w/ biz prob” before evaluating alternative big-data platforms to serve mobile clients...Usability, scalability all key.”

Molaro said “any platform that would allow users to get information from #bigdata.... the one able to provide fast access to organization’s insight with easy of operation.”

Mullins said “I think the presentation + interface on the mobile device is more important than Hadoop, etc....I’d look for customizable reports, dashboards, + graphs that users can adapt to their #mobile preferences + needs.....Maybe innovative interaction support (gesture, voice, etc.) too.”

However, Philp (bless his heart) exclaimed “#Streams!” (as in “IBM InfoSphere Streams....no, we didn’t put him up to it). When challenged on that, Philp elaborated: “streaming analytics allows customers to keep up with the volume and interoperate with Hadoop for non-real-time analytics....#Streams means real-time which means millions in the Telco space, esp., in Eurozone.....Real-time at network line speed - 1, 10 and 100 GigE speeds - analytics in process, in flow - compute memory.”

Kelly generally concurred with Philp’s point of view, but kept his comment vendor- and platform-neutral: “definitely streaming, CEP-style tech to allow automated actions - must take action while customer still engaged.”

Your humble IBM big-data evangelist (keeping the discussion on the vendor/platform-neutral plane) offered these thoughts: “The mobile #bigdata environment should have SQL-query-virtualization front-end to simplify access....Need lower-latency in-motion #bigdata platforms closer to mobile client, batch & “data-at-rest” further...in-memory #bigdata clients/servers with back-end streaming best for real-time mobile ....No particular back-end #bigdata platform preferred 4 mobile. Need front-end mobile access infra agnostic 2 all.”

What types of insight are orgs getting from mobile-generated big data?

Here’s where your not-as-humble-as-he-lets-on big-data evangelist cuts to how I personally responded to this from my sky-high vantage point:

  • “vehicle-sourced mobile data insights will help traffic planners dynamically optimize world transportation grids
  • wearable & implanted mobile devices will deliver unparalleled insights into wellness, health, & experience
  • mobile-gen data + data sourced from all other channels = fodder 4 dynamic multi-chann xperience optimization
  • mobile-sourced #bigdata fleshes out the “720-degree customer view” (external behavior + internal experiences)
  • deep machine-data analytics insights on geolocat, sentiment, behavior and other signals sourced from mobiles”

When should you tap into smartphones, mobile clients as sources for big-data apps?

Same applies to this one. Here’s my final tweet on the chat before stowing my mobile device in the overhead compartment and asking for a Diet Coke:

  • “Always. Smartphones becoming the most ubiquitous, valuable source of ambient, geo, sentiment, & experience data”

Not so coincidentally, here’s the blog I posted on this very topic the following day: Big Data On the Move: Everywhere You Need It To Be

Join in!

Now I’ll get humble again. We want to thank everybody for their great contributions to this Twitterchat. I enjoyed re-reading it all again now that I’m down from my lofty perch.

Please look at Jeff Kelly's excellent post, 'The Dual Role of Mobile Devices for Big Data,' which reflects on a theme that surfaced in our Twittertchat. Join us March 27 for another spirited Twitterchat – this one focusing on in-memory data. Just hop onto Twitter at noon ET and join or follow the conversation using #bigdatamgmt.

 

Related information

  • To find out more about managing big data, join IBM for a free Big Data Event
  • To learn about high-level business use cases for big data, listen to this podcast, "Top 5 Big Data Use Cases." Eric Sall, vice president of product marketing at IBM, describes the key use cases that hold high potential value for many organizations.