I keep not finding the time to write as much about business intelligence as I’d like to. So I’m going to do one omnibus post here covering a lot of companies and trends, then circle back in more detail when I can. Top-level highlights include:
- Jaspersoft has a new v3.5 product release. Highlights include multi-tenancy-for-SaaS and another in-memory OLAP option. Otherwise, things sound qualitatively much as I wrote last September.
- Inforsense has a cool composite-analytical-applications story. More precisely, they said my phrase “analytics-oriented EAI” was an “exceptionally good” way to describe their focus. Inforsense’s biggest target market seems to be health care, research and clinical alike. Financial services is next in line.
- Tableau Software “gets it” a little bit more than other BI vendors about the need to decide for yourself how to define metrics. (Of course, it’s possible that other “exploration”-oriented new-style vendors are just as clued-in, but I haven’t asked in the right way.)
- Jerome Pineau’s favorable view of Gooddata and unfavorable view of Birst are in line with other input I trust. I’ve never actually spoken with the Gooddata folks, however.
- Seth Grimes suggests the qualitative differences between open-source and closed-source BI are no longer significant. He has a point, although I’d frame it more as being about the difference between the largest (but acquisition-built) BI product portfolios and the smaller (but more home-grown) ones, counting open source in the latter group.
- I’ve discovered about five different in-memory OLAP efforts recently, and no doubt that’s just the tip of the iceberg.
- I’m hearing ever more about public-facing/extranet BI. Information Builders is a leader here, but other vendors are talking about it too.
A little more detail, especially on Jaspersoft:
- Jaspersoft is not using multi-tenancy is to offer BI SaaS itself. But SaaS vendors were demanding the feature. What’s more, a couple of household-name corporations are using Jaspersoft’s multi-tenancy to give extranet BI access to their various customers or suppliers. Lawrence Livermore Labs seems to be a Jaspersoft extranet user too.
- The way Jaspersoft’s multi-tenancy works is that the concept of “organization” is added to the privileges hierarchy. Each organization sees its own virtual server. Only administrative superusers can span organizations.
- Jaspersoft also has a new memory-centric OLAP capability — with disk-based ROLAP for overflow — unrelated to the Mondrian MDX server. That’s a pretty common story in BI these days, I think, but I’ll confess to being unclear about exactly who is offering what when in that regard.
- Jaspersoft’s memory-centric OLAP is just a query accelerator, not a near-real-time data ingester like Aleri Live Update. Jaspersoft does handle real-time telemetry from at least one space mission (to Mars) — but how great can the bandwidth on that be?
- There’s also some kind of AJAX/Web 2.0/mash-up/whatever going on in Jaspersoft v3.5.
Some more detail yet, especially on Inforsense:
- Inforsense is focused on applications that answer a few questions rather than doing high-volume analytics, and are “designed to change.” This may be needed when there’s enough of an analytic business process that conventional BI tools aren’t a good fit (perhaps unless combined with some kind of composite application development tool or methodology).
- Inforsense’s application sweet spot to date is combining and moving around various kinds of health care data. (Especially laboratory data, both research and clinical.)
- Inforsense is a bit confusing because it was founded out of an academic research effort (Imperial College, London) to do data mining parallelized onto grids. That is no longer the company’s main focus, but the confusion continues with an occasional low-revenue, supposedly-high-prestige research award.
- Inforsense is further a bit confusing because, irrespective of focus, its analytic technology can supposedly be almost all things to almost all people. (Exactly the same thing complaint could be made about almost any other BI company.)
- What remains of the academic focus is what Inforsense characterizes as a “very flexible dataflow environment.”
- Inforsense can talk to lots of data sources and so on, including web services. It can also do updating, albeit not in demanding OLTP environments.
A few more notes, especially on Tableau Software:
- Tableau is built around a proprietary language VizQL. VizQL seems to be similar to SQL in that it focuses on filtering data. I haven’t yet read a paper Tableau sent, which should make it clearer what VizQL does that SQL doesn’t.
- Tableau is one of the new breed of “exploration” oriented BI vendors, encouraging users to just dive into data.
- I don’t know whether this is more a matter of technology or just astute marketing, but Tableau seems to be somewhat more focused than other vendors on the idea that you filter data, keep refining that filter as makes sense to you, share that filter with other people, and so on. It is hard to overstate how blind I think the BI industry is being in not aggressively developing and enhancing this kind of technology.
- That said, Tableau’s capabilities in this area still seem pretty primitive too.
- Like most software vendors, Tableau says its biggest competitor is incumbent/no decision. In Tableau’s case, the incumbent can be either BI tools or Microsoft Excel.
- Tableau says its second biggest group of competitors is other new/easy BI vendors such as QlikTech and LogiXML. Interestingly, both got mentioned with about equal emphasis.