Archive for category Big Data
HP and Hortonworks deepened their relationship last week, and the deal says a lot more about the former than it does the latter.
The news is that HP is investing $50 million in Hortonworks for about a 5% ownership stake in the company (Hortonworks’ Series D valuation is estimated at $1.1 billion) and a seat on Hortonworks’ board. HP will resell the Hortonworks Data Platform (HDP) and provide Tier 1 support to customers. The two companies will also work together to certify the HP Vertica analytic database on YARN.
Organizations are increasingly coming to the realization that data is a core strategic asset, the new source of competitive advantage. Success in today’s economy is predicated on how organizations leverage data as much as any other corporate function. The challenge organizations face, then, is how to leverage this asset to its maximum potential in ways that are as efficient as possible and also minimize risk.
This is no small feat. It involves data itself (identifying and managing sources of data), technology (tools and systems to ingest, process, store, analyze and share data), governance (ensuring data is used ethically and in compliance with relevant policies/regulations) and people (aligning various stakeholders and business objectives.)
Despite the apparent contradiction, Hadoop and other emerging Big Data approaches are at the same time complementary to and disruptive to established data warehousing and business intelligence practices in the enterprise. I recently spoke with my colleague Stu Miniman about this and other findings from Wikibon’s Q2 2014 Big Data Analytics Survey in the below Cube Conversation. The survey, one of two major Big Data surveys Wikibon will undertake this year, is part of Wikibon’s new Big Data research service. The new service is focused on primary data-driven research designed to uncover how Big Data is practically applied in today’s enterprise, explore the impact on existing modes of data management and analytics, and to understand its implications for existing and start-up Big Data vendors. To find out more about Wikibon’s new Big Data research service, please email
This week there are two important enterprise technology conferences taking place. One – SAPPHIRE 2014 – will see an old guard enterprise tech giant attempt to show it is capable of adapting to a technology landscape increasingly dominated by the cloud and Big Data. The other – Hadoop Summit 2014 – will see dozens of start-ups born in this new world out to prove to cautious CIOs that their technologies and platforms are ready for enterprise-level workloads.
It’s an interesting juxtaposition. SAP is determined to join the ranks of the “cool” cloud and Big Data companies (Salesforce.com, Hortonworks, Amazon Web Services), while those cool companies are equally determined to join the “enterprise-grade” club dominated by IBM, Oracle and, yes, SAP.
The Hadoop ecosystem is an eclectic mishmash of start-ups, mid-sized vendors and IT heavyweights with products and services up and down the Big Data stack. Inevitably the ecosystem will consolidate and thin itself out through mergers, acquisitions and – unfortunately for some of these start-ups – bankruptcies.
Consolidation is part of the natural evolution of any given technology market after an initial period of frenzied innovation, and the Big Data market is no exception. I believe we are witnessing the start of this consolidation today. It will take several years to play out, but the first phase of consolidation is manifesting itself in the form of strategic technical partnerships between vendors that play in different segments of the Hadoop market.
IBM’s annual revenue last year dropped below $100 billion for the first time since 2010. The company’s fourth quarter results were particularly weak, coming in 5.5% below expectations. This was due in large part to IBM’s struggling hardware business, with revenue dropping a staggering 27%.
I’ve already laid out my predictions for Big Data in 2014, but I also wanted to let the Wikibon community know how my colleagues and I plan to cover Big Data in the year ahead. We’ve organized our research agenda into three major buckets.
Technology. Clearly the technologies and products that collectively make up Big Data – including Hadoop, NoSQL data stores, analytic databases, data visualization tools and more – are maturing at a rapid pace (much faster, for example, than relational databases did in the 1980s.) Big Data is also applicable across industries, meaning these technologies are inevitably and increasingly intersecting with adjacent technology movements, namely the cloud, mobile computing and social media. As we have for the last several years, Wikibon will devote significant coverage to these developments with an eye on putting technology innovations in context for enterprise Big Data practitioners (both technology practitioners and line-of-business practitioners.)
After a (very) brief respite, #theCUBE is back on the conference circuit. Starting tomorrow at 10:20am ET, theCUBE is live for two days at the HP Vertica Big Data Conference (#HPBigData2013) in Boston. Unlike some Big Data shows that focus more on vendors, the Vertica show is heavy on customers and real Big Data end-users.
Everyone knows that the shift from traditional print and broadcast advertising to digital advertising is taking a huge toll on the media industry. Newspapers can’t charge as much for online ads as they do for print ads, and revenues are shrinking precipitously.
But the impact of this shift is being felt not just in the media industry but in the advertising industry itself. Advertising agencies, whose bread and butter is building creative ad campaigns and negotiating print and broadcast ad placement, are likewise under pressure to make the transition to a data-based marketplace.
There’s an old saying in the data management world: garbage in, garbage out, or GIGO. It means that the results of any data analysis project are only as good as the quality of the data being analyzed. Data quality is of critical importance when data sets are relatively small and structured. If you only have a small sample of data on which to perform your analysis, it better be good data. Otherwise, the resulting insights aren’t insights at all.