Sunday, September 13, 2009

Aweekstweets Sept 7-14 2009 @jameskobielus

  1. Contemplating bandwidth (the metaphorical). Ratio of timewidth to taskwidth. Wide time, narrow task = plenty of bandwidth. Opposite: now, me20 minutes ago from TweetDeck
    JK2—Q3 drawing to a close. Deliverable closure activity even more dense, intense—especially when you’re also on the road.
  2. @CurtMonash @SethGrimes @jamet123 I'll be speaking at Teradata Partners. Would like to catch Pred Analytics World. I live in DC area.about 19 hours ago from TweetDeck
    JK2—I’ll be speaking at Teradata Partners on in-database analytics: business drivers, use cases, approaches, standards, maturity, vendor support, futures. It’s the convergence point between my two core Forrester coverage areas: data warehousing and data mining/predictive analytics. Predictive Analytics World is happening the same week, in the same area (Washington DC metro area), where, coincidentally, I reside (not often that tech conferences I want/need to attend are in the DC area—strange that two should happen in the same week).
  3. Pondering the convergence of analytic and transactional database appliances. Evolved DW as the new analytic app server.about 22 hours ago from web
    JK2—This is the thesis of my in-database analytics preso at Teradata Partners, plus the Forrester teleconference I did last week, plus the upcoming Forrester report I’ve drafted. In a nutshell, analytic data repositories—with deep history, multiple subject domains, complex information sets—are getting too massive to move. It makes sense to move predictive-analytic application logic to where the data lives--the enterprise data warehouse (EDW)--than to move the data to physically distinct analytical data marts. The benefits of in-database analytics (aka “no-copy analytics”) include bandwidth saving, faster data prep and scoring, and improved model governance. As more application logic migrates to the EDW and as these apps are wrapped in SOA/REST interfaces, this platform becomes a more mission-critical analytic app server.
  4. rt @markrittman Oracle to announce another "database machine", this time based on Sun hardware and aimed at OTLP :
    JK2—Really? That would make sense: transactional app logic will move to analytic app servers (aka EDWs) such as current Oracle Database Machine with Exadata. Other EDW vendors are moving down similar roadmaps. I’m looking forward to Oracle OpenWorld next month. Seeing as how the European regulators are holding up the Sun acquisition, there may not be any attention to the Oracle/Sun roadmap at that show. For obvious reasons.
  5. "Anti-Tweet Business Retreat" ( jk--Saddle up the ponies, ride the range. No tweeting allowed. Birds: this means u 2!about 1 hour ago from TweetDeck
    JK2—Yeah, yeah, cowpoke. The city slickers are smarter than you give ‘em credit fer. They’ll tweet from their phones, all the time pretending that they’re dialing a particularly long phone number just to call that pretty little filly back in the city and brag about how they’re roughing it.
  6. "How to Calculate Data Warehouse Reliability" ( JK--Very cool, precise, analytical yet actionable assessment metric.about 1 hour ago from TweetDeck
    JK2—I’d like to develop a similar metric on data warehouse scalability.
  7. "Silicon Valley is shrinking" ( jk--So's my Rust Belt hometown. Welcome to the "gee remember those good ol' days" club.about 1 hour ago from TweetDeck
    JK2—Let’s state this in a slightly different way. Silicon Valley has been virtualized to the world. The new “garage” is the social network of familiar strangers hatching new enterprises in the cloud. The earthbound “valley,” no matter how lovely, will be eclipsed by shifting clouds overhead, which know no borders.
  8. "Raytheon To Buy Internet Pioneer BBN" ( jk--I'd forgotten BBN still existed. Pioneered use of "@" in e-mail address.about 1 hour ago from TweetDeck
    JK2—Remember when that symbol, which was on most manual and typewriter keyboards, effectively stood for “each” or “individual”? As in lists that had lines such as “99@ bottles of beer on the wall.”
  9. "Twitter and the New Social Marketing Game" ( jk--"Twitter is a social space, not a selling space"? Don't be naive!about 1 hour ago from TweetDeck
    JK2—That’s what we all said about blogs. Now my blog provider offers a tab labeled “monetize.”
  10. "10 Ways To Complement the Enterprise RDBMS Using Hadoop" ( jk--Yes. Virtualize DW with HDFS i/f. In-cloud analytics.about 1 hour ago from TweetDeck JK2—Here’s part of the confusion. Hadoop refers to an advanced-analytic function development workflow model applicable to complex content, including but not limited to what’s in DBMSs. It also refers to a distributed file system to support that content-rich model. Over the coming 5 years, today’s relational DBMS, the Hadoop Distributed File System (HDFS), and other persistence approaches will be virtualized into a more all-encompassing cloud of spindle and cache, structured and unstructured, at-rest and in-motion analytic content. “Virtualized DW” is too restrictive a term for this coming environment, but it’ll have to do for now.
  11. @jamet123 "Removing Decisioning From the SDLC" ( jk--Service-oriented analytics. Reusable decisioning svcs. Agile dev.about 1 hour ago from TweetDeck
    JK2—I agree with Mr. Taylor. Software development life cycles usually flow in the classic waterfall. But inline decisioning services are evolved in an iterative, dynamic, collaborative, agile fashion. That’s because these services depend on predictive models maintained by analysts—i.e., subject matter experts who themselves work in such a fashion, independent of software developers. Service-oriented analytics is SOA in the predictive modeling realm: enabling maximum reuse of predictive models among teams of analysts leveraging each other’s expertise on business problems that cross multiple domains.
  12. "Oracle finally updates flagship database" ( JK--Oracle DB 11g R2 is very substantial release: scalability, performanceabout 1 hour ago from TweetDeck
    JK2—And high availiabity too. My fave of many new features: Oracle RAC One Node. It supports server consolidation by leveraging RAC technology to enable clustering, load balancing, and failover across server blades in a single node.
  13. "IBM Chief Economist Says Recession Is Over" ( jk--Private-sector economists feel first pulse of fresh vigor. Pay heed!about 2 hours ago from TweetDeck
    JK2—Which brings up academic hairsplitting questions such as “what’s an economy?” and “what’s a recession?” [Note: This tech analyst has a degree in economics from the University of Michigan—Ann Arbor]. There are so many possible economic indicators you could point to (individually or in clusters, from GDP through consumer confidence, lending, hiring, profits, inventories, stock prices, etc) on so many possible levels (micro-economic or macro-economic, across so many sectors, regions, and countries), so many comparative baselines (absolute growth, percentage growth, comparisons with last quarter or year, comparison with worst-case scenario of Great Depression II), and so many authorities voicing their opinions on all this all the time through all possible channels. You accept Ben Bernanke’s word. Others heed Jim Cramer. Some listen to the president of the United States, who’s listening to his advisors. Still others think IBM’s chief economist has a thing or two to say on the matter. We’re all grasping at straws, until the next boom, however we slice, dice, and define it, is here in full force, undeniable.
  14. "SSDs Ready For The Enterprise" ( jk--Flash storage will be all the rage in DW query/mart appliances in 2010 & beyond.about 2 hours ago from TweetDeck
    JK2—By 2015, flash storage will be standard in the query/host tier in all DW appliances.
  15. "From Amazon To IBM, What 12 Cloud Computing Vendors Deliver" ( jk--Great analysis. Check out cloud comparison chart.about 2 hours ago from TweetDeck
    JK2—Some trade papers have excellent analysts on staff. They call them “reporters.” Guess which reporter-analysts all analyst-analysts read.
  16. "Using social media to turn staff into customer svc reps" ( jk--And thru self-service letting customers rep themselves.about 2 hours ago from TweetDeck
    JK2—Social media, coupled with business activity monitoring, will bring this vision to fruition. See Natalie Petouhoff and my preso on this topic at next month’s Forrester Business and Technology Forum in Chicago.
  17. "XML Threatens Big Data? Bull@*#!" ( jk--Usually, XML gets staged, transformed, compressed before DW. Crunched to fit.about 2 hours ago from TweetDeck
    JK2—Apparently, this headline was written by Penn and Teller. I did a piece for the now-defunct Business Communications Review in 2005 on this very topic, from a transactional computing standpoint. With data warehousing, XML is almost never loaded into the production tables, or into cubes. XML is usually a source format that gets transformed at the staging layer to some more storage-efficient format, with the semantics of the original XML content managed in the corresponding ETL/DW/BI metadata layer.
  18. "Madison" ( jk--Wisconsin capital. MSFT DW appliance: scales from Monona to Mendota to Michigan. Deep, fresh data lakesabout 2 hours ago from TweetDeck JK2—Note: This tech analyst also has a journalism degree: from the University of Wisconsin—Madison. The capital city of America’s Dairyland is on a scenic isthmus between two reasonably large inland lakes: Monona (to the south) and Mendota (the larger, to the north, upon whose southern shore sits that university, and in whose waters the plane carrying soul legend Otis Redding crashed in 1967). Madison sits in southern Wisconsin, around 90-something miles to the west of the only Great Lake that sits entirely within the USA’s borders. Seattle’s about 2,000 miles to the west. Aliso Viejo CA, where the bulk of SQL Server “Project Madison” development is taking place, is perhaps 1,500 miles south of Seattle, just north of San Diego. Draw a triangle on the map between Madison, Seattle, and Aliso Viejo, and, smack dab in the middle, you’ll find Yellowstone or Grand Teton. Both of them would make find codenames for this Microsoft project project. Especially Grand Teton: scaling petabyte mountain. I always chuckle remembering what Grand Tetons means in French.
  19. "Surfing in Hawaii" ( jk--KONA important release for Kalido. Even more important for Netezza: DW governance appliance.about 2 hours ago from TweetDeck JK2—Netezza can’t keep margins healthy without move toward solution-based packaging of advanced analytics and data management offerings, such as MDM. Kalido can’t stay in the game without wide range of DW appliance partnerships.
  20. "MySpace, Facebook, Other Social Ntwks Account for 20% of Online Display Ads" ( jk--Display ads. Play ads? Social ads?about 2 hours ago from TweetDeck
    JK2—Ads migrate toward the virtual places where people truly live. Toward social networks, where the ads become an integral component of the relationships that sustain those environments. Not so much ads that display themselves, but ads that figure into people’s social lives. Ads that add to the community, not just suck cash and patience from it.
  21. "11 Apache Technologies That Have Changed Computing" ( jk--Love these substance-packed eWeek slideshows. #15: Hadoop!about 2 hours ago from TweetDeck
    JK2—Apache is as formative to Web 2.0 as IETF was to Internet 1.0.
  22. The ZapForum DC event ( I'll be there. New poem composed for occasion: "SOA DOA? No Way!" ( 5 hours ago from TweetDeck
    JK2—Yeah, I stuck with my tried-n-true triple-haiku. It germinated the night before. It terminated my sleep early on the morning after. It culminated in a quick foxtrot to the laptop with bang and beep and a tweak till it came out just right. For that occasion—in 3 weeks. I never truly intended to write it on the spot as a door prize. Or, rather, I toyed with the idea and then realized I’m not Nipsey Russell.
  23. Social network analysis: look for patterns of background, behavior, preference, interaction of which people are unaware. Make them aware.about 13 hours ago from TweetDeck
    JK2—Social networks aren’t really networks. They’re simply contexts within which people may eventually find themselves and others to be more alike than different—or which they may be found by others to be an affinity cluster capable of being addressed for fun and profit.
  24. Catch and keep a customer. Covet the customer list. Uncover the customer social networks. Social network analysis: next-gen data mining.about 13 hours ago from TweetDeck
    JK2—Relationship mining and predictive relationship analytics. Discover the full extent of the current B2C relationship and/or predict the most likely B2C relationships that may be brought into being—through well-targeted sales, marketing, customer service, and other efforts.
  25. Smiling. Have significantly expanded my knowledge of today's data mining market, and tomorrow's. I'm curiosity's cat, of Cheshire varietyabout 15 hours ago from TweetDeck
    JK2—My Forrester Wave on Predictive Analytics and Data Mining Solutions. Essentially, mining the value prop of the current vendors/solutions and predicting how that is likely to evolve over the next few years. No, there have been no mind-blowing revelations. It’s just good to evaluate them all, their current offerings and roadmaps, side by side by side. The demos have been particularly enlightening.
  26. "Database-less DW": That actually federated data management with robust DQ and MDM. That's also a pipe dream. @markmadsen @Claudia_Imhoff \about 19 hours ago from TweetDeck
    JK2—You need a centralized DW to be the governance focus for ensuring the quality of the “golden record.” Of course, this paradoxical phrase—“database-less data warehouse”--is not technically accurate. Even in federation environments where you have no DW (centralized or hub-and-spoke), you still generally source/OLTP databases that are storing data and responding to queries. If you take this “warehouse” notion for the metaphor that it in fact is, then the “database-less DW” is analogous to a just-in-time supply chain without transshipment depots. Customer orders get routed in real-time to the factories, which respond with production runs and shipments in real-time.
  27. The fact that DW startups are getting new leases on life through fresh funding shows a) economy improving and b) analytics remains red-hot.about 19 hours ago from TweetDeck
    JK2—And c) some DW vendors have hot enough IP and hot enough visions that investors still find impressive enough to bet on.
  28. Enjoyed James McQuivey #forrester discussion of Social TV research ongoing. I'm doing it: Twitter/Facebook over Verizon FIOS.3:28 PM Sep 10th from TweetDeck
    JK2—When I say “I’m doing it,” I mean “I’ve played with it.” I’m not doing it now. I don’t do most of my tweeting or facebooking that way. Mostly, because I don’t have a keyboard that connects with my FIOS TV service. Also, because I don’t watch much TV anymore. And because, when I do watch TV, the only interface I have any tolerance for is the remote—and the only functions that suit me are on/off, volume +/-, and channel +/-.
  29. RT @TonyBaer: RT @jameskobielus & Rob Karel "INFA deal... HP’s last brave attempt....." jk--Not so. We couch it as "case could be made."1:34 PM Sep 10th from TweetDeck
    JK2—In that blogpost, what we said specifically was: “One could even make a case for the Informatica deal representing HP’s last brave attempt to mount a credible data management software strategy.” And then we follow it with a challenge to HP: “If HP is truly serious about becoming a full-fledged data management software provider, now is the time to make strategic acquisitions in the DI, DQ, MDM, BI, PA/DM, and other key solution markets.” Doesn’t sound like we’re sounding the death knell on HP’s data management strategy, does it?
  30. Rather enjoyed the flight up to Boston this eve. Coincidentally @CurtMonash was 1 row up across aisle. Made nice non-tweet industry chitchat8:45 PM Sep 9th from TweetDeck
    JK2—And we gave each other our respective quiet zones. Curt could see me zoning—or, rather, Zuning—with my music and thoughts. I love flying.
  31. Finding folks confusing my upcoming FORR PA/DM Wave & FORR rpt on in-db analytics. These are 2 distinct projects, y'all, but complementary.5:05 PM Sep 9th from TweetDeck
    JK2—The Forrester Wave on Predictive Analytics and Data Mining Solutions will be published by year-end. The Forrester report “In-Database Analytics: Heart of the Predictive Enterprise” is substantially and will be published in the next several weeks. I plan to blog about it soon.
  32. I'm well along with the exec strategy briefings and demos on my Forrester PA/DM Wave. Soon will do ref customer interviews.5:04 PM Sep 9th from TweetDeck
    JK2—The reference customer interviews are where I test the vendors’ claims by speaking to users about their experiences. It’s not perfunctory.
  33. RT @jamet123 @soachief @rschmelzer @DavidLinthicum @jpmorgenthal @TheEbizWizard "reuse decisions made w/them" jk--Reuse best practices too!4:56 PM Sep 9th from TweetDeck
    JK2—Re what I’m calling “service-oriented analytics,” I’ll sum up for you all the things that James Taylor and I said should be reused: best practices, predictive models, predictive modelers, decisioning services, decisions. Here are my thoughts on how they can be reused: best practices (through what Forrester calls “BI Solution Centers,” on which Boris Evelson and I published a report last year); predictive models (through predictive model governance tools such as those offered by SAS, SPSS, KXEN, and others); predictive modelers (by cultivating advanced analytics experts who work across many projects, applications, and business units); decisioning services (through convergence of business rules engines, BPM, and BI; see Forrester report that Boris, John Rymer, and Colin Teubner published last year), and decisions (by making the right decisions, which then serve as proper precedent for future business problems).
  34. True story: Here at RRWashNatlAirpt, me going for meal, overhear guy say "We have best DW in world, it's called [vendor]." No, not telling.4:51 PM Sep 9th from TweetDeck
    JK2—And not agreeing. Studying this industry for a living, I can tell you they all have their strengths and weaknesses. See my Forrester Wave on Enterprise Data Warehousing Platforms, from this past February.
  35. Did great briefing today from MicroStrategy on their BI-integrated predictive analytics and data mining solution, partnership, strategy.2:52 PM Sep 9th from web
    JK2—For me, a “great briefing” is when I get a deep dive on what a vendor’s doing now, and what they plan to do over the coming year or so. Doesn’t necessarily mean I think they’re the best vendor in their niche. But MicroStrategy does indeed have one of the most interesting approaches to BI-integrated PA/DM in the industry. Check the recent Forrester report on convergence of PA/DM into the core BI stack, which I co-authored with Boris and Leslie Owens.
  36. ZapForum DC: Asked Ron to change my descriptor from "SOA chatterbox" to "Champion of Service-Oriented Analytics." Same dude different facets2:50 PM Sep 9th from web
    JK2—I have every right to self-brand.
  37. @soachief @rschmelzer @DavidLinthicum @jpmorgenthal @TheEbizWizard: jk4--Predictive/agile enterprise -> Keep on hitting moving targets.9:41 AM Sep 9th from TweetDeck
    JK2—Business is just a never-ending stream of bets. Hey Socrates, you could say the same about life itself.
  38. @soachief @rschmelzer @DavidLinthicum @jpmorgenthal @TheEbizWizard: jk2--Service-oriented analytics -> comprehensive in-cloud analytics9:33 AM Sep 9th from TweetDeck
    JK2—Check out discussion of inline analytics maturity model, culminating in “comprehensive cloud analytics,” from my recent Forrester teleconference: “In-Database Analytics: Transforming The Data Warehouse Into The New Analytic Application Server.” Same scope as forthcoming Forrester report authored by self.
  39. HP/Informatica alliance sets HP on path to deliver analytic solution appliances with Neoview inside. Needs BI to be complete, though.7:01 AM Sep 9th from TweetDeck
    JK2—See Rob Karel and my recent blogpost: “HP and Informatica’s Expanded Relationship: Portent of Bigger Deals to Come?
  40. Still encountering resistance from Hadoop community to be discussed in "in-db analytics" context, but I discuss in re "in-cloud analytics."6:54 AM Sep 9th from TweetDeck
    JK2—Hadoop won’t come to fruition until it’s integrated into data warehouses everywhere. Data warehouses, in turn, will evolve into clouds
  41. "Pluck to Launch Social Media Application Server" ( jk--New solution segment? Or renaming of collaboration solutions?7:19 AM Sep 8th from TweetDeck
    JK2—Not sure this makes sense as a separate silo product niche for premises deployment. Social media is all about converging interaction channels. It’s also about all things cloud, Web 2.0, and SaaS.
  42. "A Fact-Based Defense of Enterprise 2.0" ( jk--Huh? Did someone create a new religion and not tell the rest of us?6:50 AM Sep 8th from TweetDeck
    JK2—Who, if anyone, wrote the enterprise 1.0 gospel? Adam Smith? Andrew Carnegie? Peter Drucker? Tom Peters?
  43. Suffered 5 bee stings in one attack mowing the lawn Saturday morning. Little buggers ambushed me toward the very end of my mow.5:59 AM Sep 7th from TweetDeck
    JK2—I’m better now. Thanks for asking.