BI Blogs

While researching and testing the default Web Report Studio metadata security I noticed this little trick for young players, which means changes to WRS/WRV Metadata Security for a user are not instantly applied/recognised in WRS/WRV.

Words from the SAS tech support note “Changes to SAS® Web Report Studio role memberships might not immediately be enforced“:

tyle=”color: #999999;”>“SAS Web Report Studio analyzes role memberships every 30 minutes. So, for example, if you move a user from the WRS Report Author role to the WRS Report Consumer role, the user will continue to have author privileges until the next time that he logs in after the next role membership analysis is performed. ”

So effectively the WRS/WRV applications are caching the Metadata security to remove the need for the web server to query the Metadata Server each time a user logins.

So try not to be fast when testing your WRS Metadata security changes, else you will think you have done something wrong.  And remember to logout and log back in after the cache has been updated.

You can update the LocalProperties.xml file to change the refresh time,  and therefore make the changes appear sooner.  Details on how to do this are in the SAS Tech Support note.

In our post Enterprise Software: Is there any one left to buy? we listed the best candidates to be acquired within the lucrative Business Intelligence (BI) and Data Warehousing (DW) industry:

• SAS (Private)
• Informatica (INFA)
• MicroStrategy (MSTR)
• Netezza (NZ)
• Teradata (TDC)

We have ruled out privately-held SAS because they are very reluctant to be acquired. (If they were willing then IBM would likely have bought them versus SPSS…no offense to SPSS but SAS is the 800-lb gorilla in the business analytics segment.)

The Usual Suspects

The high tech Titans - IBM (IBM), Hewlett-Packard (HPQ), Oracle (ORCL), Microsoft (MSFT), SAP (SAP) – are already very active in the DW/BI industry selling software, services and in some cases hardware. Are any of the acquisition targets on our list attractive to the Titans and would impact their market presence?

Oracle, SAP and IBM are top of the heap built with their previous acquisitions. It’s likely that they’re on the prowl for these larger acquisitions. (Oracle is always a wild card and sometimes acts as if it was the NY Yankees of software firms.)This leaves one Titan in our group, HP, which must be feeling it is being left behind. Sure it picked up EDS a year ago in services and has been squeezing out costs (it is awesome at that!) But the core of HP’s business is printers, printer supplies and low margin PCs. That generates a lot of revenue but it does not provide synergy for your services group to do the high-end systems integration work that IBM, Oracle and SAP services do. HP could try to boost its hardware revenue by acquiring either Teradata or Netezza, but that is not the break-out strategy they need and it just creates competing hardware lines that cannibalize each other.

HP already offers NeoView as their BI/DW appliance and sells a lot of hardware that is used in BI/DW implementations today. The break-out move for them is to acquire both Informatica and MicroStrategy. That vaults HP into the upper echelon in BI/DW markets coupling one of the best (or the best) data integration platform with a BI suite on par with anything the other Titans offer. HP can then move its EDS resources into higher end SI work and can finally leverage its purchase of Knightsbridge into projects bundling its hardware, software and services for years to come.

Throw in Tibco (TIBX) and HP could think about increasing revenue, profit and margins by growth rather than solely by cost cutting and waiting for a general IT recovery to lift printer and PC sales.

Breaking in

That leaves the two BI appliance vendors up for grabs. If not the usual suspects then who? There are two hardware titans, EMC (EMC) and Cisco (CSCO) to make some bold moves. EMC has been increasing its software exposure and made a bold move by acquiring VMware a while ago. Cisco has decided recently to go head-to-head with its former partners by selling servers and should be looking at how to strengthen itself against its former partners.

Cisco is the bigger firm and needs to make a much bolder statement than EMC. Cisco should pick up Teradata, thereby in one move acquire hardware, software and services assets to compete against the other Titans. This moves it out of its networking gear comfort zone, but every tech firm, if it dominates its market, needs to be looking for other markets to penetrate. Its ticket is Teradata.

EMC, with much less resources at its disposal, should acquire its fellow Massachusetts firm Netezza (with their respective headquarters just miles apart). EMC built its franchise with excellent engineering in hardware and software just as Netezza is doing. From an engineering and marketing perspective this combination should be very symbiotic.  It is a way for EMC play with the other Titans, leverage the massive amounts of their storage used for BI/DW and provide the boost Netezza needs to become a major player.

Dark Horses

These predictions are the bold bets that the Titan acquirers should make. But oftentimes companies are too locked into their current thinking that they do not even think outside the box. If the Titans do not buy these firms, then others, who do think out of the box, will consider them. You may be surprised as to who make the acquisitions.

The firms we are mentioning as acquisition targets are enjoying solid growth, but in the long-run it is likely the Titans will encroach on their market share. For both parties, the best time for these mergers is before the IT boom gets too far along.

Disclosure: I have current stock positions in the following companies listed in this post: INFA, NZ, ORCL, TBX, TDC.

Pues voy a gastar mi post nº 100 en enfadarme por la jeta que tienen algunos, en concreto estos cuatro individuos de Bangalore .

• Sunil Senan.
• Nitin G. Metri.
• Varun Babbar
• Gaonkar K.P

Voy a ver si patento la cadena de montaje que me forro.

Sistemas Decisionales, algo mas que BI

SSAS caching is an interesting subject: I’ve blogged numerous times about when caching can and can’t take place, and one of the most popular posts I’ve ever written was on how to build a cache-warmer in SSIS. However that last post was written a while ago, before the CREATE CACHE statement for warming the Storage Engine cache was reintroduced in SSAS 2005 SP2, and I’ve not been satisfied with it for a number of other reasons too. So as a result I’ve started to do more research into the issue and, with the help of Mosha, Akshai Mirchandani, and the chapter 29 of the indispensible “Microsoft SQL Server 2008 Analysis Services Unleashed” I’ve been collecting my thoughts on how to build a better cache warmer with a view to blogging about it again. This first post will deal with the relatively easy bit – warming the SE cache – and when I have time I’ll write about the more difficult but more useful task of warming the Formula Engine cache, and by hopefully then I’ll be in a position to create the definitive SSIS cache warmer package and be able to blog about that too.

Before I go on I need to mention that one other reason that has made me want to revisit this subject is that, at his MDX Deep Dive session at PASS last year, Mosha asserted that cache-warming was a Bad Thing. This intrigued me, since it seemed like such an obviously Good Thing to do, but of course it turns out he had good reason for saying what he did… SSAS caching is a lot more complex than it first appears and cache warming done badly can be counter-productive.

I’m also tempted to say that if you don’t need to warm the FE cache (ie you don’t have any complex calculations that take a long time to run) then there’s not much point in warming the SE cache. In my experience if you have the right aggregations built then you can get excellent response times even on a cold cache. Of course to build aggregations you need to know the granularities of data that your users’ queries are, but you need the same information to warm the cache. The only difference would be that you can make your cache warmer much more dynamic than your aggregation design: if your users’ query patterns change regularly it would be difficult to keep your aggregation design in synch with them, whereas with a cache-warmer you can record the most recent queries that have been run and warm the cache accordingly.

There are a number of things to bear in mind when warming the SE cache:

• We must not try to cache too much data. If we go over the LowMemoryLimit data will start to be evicted from the cache, which means we get no benefit from that point on; if we go over the TotalMemoryLimit it’s likely that just about the entire cache will be emptied. We also have to remember that the cache will grow as a result of normal usage during the day, so we should plan to use only a fraction of the available memory when warming the cache.
• We want to avoid cache fragmentation. The SE cache uses a data structure called the data cache registry to store its data, and when the FE requests a (potentially filtered) subcube of data from the SE that data subsequently gets stored in the SE cache as an entry in the data cache registry. As a result it is possible that instead of having one entry in the SE cache that contains a lot of data, the same data could be spread across multiple smaller entries. This is bad because:
  • It means that lookup in the cache is slower as there are more entries to search
  • The SE doesn’t bother to continue searching in the data cache registry after it has compared the current request with more than 1000 subcubes, and will then go to disk even if the required data is in the cache. If the cache is filled with lots of small entries then this will happen more often.
  • If the FE requests the slice {A,B} from a hierarchy in a subcube and the data for those two members A and B are held in different entries in the cache, then it will not be found. It will only be matched if a single entry in the cache contains data for both A and B.

f the above is subject to change in later versions, but I understand that this is the current situation with SSAS 2008 SP1.

How can we find out what data to use to warm the SE cache then? The MDX queries that have been run by users aren’t much help, but we can see what subcubes are being requested when a query is run by looking at the Query Subcube and Query Subcube Verbose events in Profiler; we could therefore run a Profiler trace to capture this data, but in fact the easiest option is to just use the Usage Based Optimization wizard’s Query Log since it contains the same data as Query Subcube. The following blog entry from James Snape shows how to interpret this information:
http://www.jamessnape.me.uk/blog/2006/11/09/SubcubeQueries.aspx 
A truly sophisticated SE cache warmer would be able to take this data, along with data about attribute relationships, and work out what the most recent, frequent (but slowest) subcube requests were, then try to find some lowest (but not too low) common granularities to use for cache warming, perhaps also slicing so that only the most recent time periods were used as well.

Last of all, how do we actually go about loading data into the SE cache? We could run MDX queries but this is a bit of a risky approach since we may not know what calculations are present and so we can’t be sure that our query requests the right data; the best option is to use the CREATE CACHE statement that is described here:
http://blogs.msdn.com/sqlcat/archive/2007/04/26/how-to-warm-up-the-analysis-services-data-cache-using-create-cache-statement.aspx

I’m off to Copenhagen this evening to deliver a keynote at the Danish IT Society’s Business Intelligence conference, and as the event is not Oracle-specific I thought I’d take a look at the BI and data warehousing market in general. The theme of the keynote is around the heightened demand for BI in this particular recession, and I wanted to take the opportunity to take a closer look at three vendors who have been particularly innovative in terms of “time to market” and “user-centricity”. The first product I’ll be demoing is the Oracle BI Applications, the second is Qlikview, and the last is the joint offering from Google and Panorama that I called out on this blog at the start of 2009.

Google have over the last few years been steadily developing applications that run in the cloud and that compete with the likes of Microsoft Office and Microsoft Exchange. Google Docs is a nice environment for creating and then collaborating on shared spreadsheets, documents and presentations and in fact we use this tool extensively within Rittman Mead for collaborating across all of the countries we operate in. Google Spreadsheets isn’t as fully-featured as Microsoft Excel, but it covers the vast majority of features most people want to use, and you can collaboratively edit documents and share them without any hosted server-side infrastructure.

Panorama Software have had an interesting journey over the past ten years or so, originally developing the OLAP technology that became Microsoft Analysis Services, and the front-end tool to accompany it that they then marketed as Panorama Novaview. All of this led to a fairly close association with Microsoft until Microsoft then went and bought ProClarity (prior to this they had a policy of not favouring any particular front-end tool or vendor), which then led to Panorama linking up with Google and offering a cloud-based version of their new OLAP engine, plus an OLAP query components that could be plugged into a Google spreadsheet as a “widget”. All of this is free, which means that in theory, you can start performing OLAP analysis for nothing using just Google Spreadsheets and the pivot table widget from Panorama. So how does it work in practice, and are we going to see OBIEE and Essbase customers moving their data into the cloud and using this ultra-low cost alternative?

The first step is to either create a new Google spreadsheet, or upload one from Excel. This is in fact where you hit the first issue, as Google spreadsheets can only contain a maximum of 200,000 cells, which sounds a lot but if you consider that your data might contain 10 columns (6 of dimension values, four of measures for example) then you are limited to a maximum of 50,000 rows, which isn’t far off the maximum for Excel but it’s still a bit limiting when it comes to real-world datasets. There is a way around this – more on this later on – but assuming you can work around this, you create your spreadsheet so that it looks like this:

You might recognise the data as the Oracle Sales History (SH) data set, which I exported to CSV and then uploaded to Google Spreadsheets via an Excel workbook. Once you’ve got your data uploaded you can then start the Panorama gadget. To do this, select Insert > Gadget from the Google Spreadsheets menu, like this:

Then you select the Analytics for Google Spreadsheets gadget from the list of table views.

This then adds the gadget to the spreadsheet you’ve just set up. At this point, you can set the data range for the pivot table input, where the data is coming from (initially, just this spreadsheet), and whether the first rows in the spreadsheet contain the data item column names.

You can move the pivot table to it’s own spreadsheet page at this stage, which I do, and then I get a whole worksheet to perform my analysis. At this point, you can select the data items for the X any Y axis, and I choose to put product category and subcategory on the Y, and state on the X as at this point, the gadget thinks that “Year” is in fact a measure.

Selecting Amount Sold as my measure, I create my first pivot table, which, for something that’s free, is pretty impressive.

Once data is in, you can apply filters to dimension members, like this:

You can also change the graph type, move it around, sort and rank the results and apply different aggregations to the data.

Interestingly, and going back to the original restriction on the size of the spreadsheet dataset, you can also use a tool called CubeSlice to take a subset of a Microsoft Analysis Services cube, copy it to your local filesystem and then upload it to the Google cloud, where you can analyze it using the Panorama gadget. This presumably is going to get you around the row limit restriction, and I guess confirms that converts from Microsoft Analysis Services and Excel are probably one of the key markets for this product.

So, is it game over for Oracle, is this to OBIEE what GMail was to Microsoft Exchange? Realistically, although it demos well, it’s a fairly limited product, suffers a bit from the latency of the Java/Flash UI and feels a little bit “flakey”. Although on the surface it seems comparable to the pivot table view in OBIEE, when you think about what users want to do with pivot tables once they’ve set them up – create page and section items, set up drill-through to detail, set up navigation, create custom calculations, measures and aggregates – it’s quite a limited solution, but if the alternative isn’t really OBIEE, it’s “no BI”, then this is certainly quite an attractive proposition. In addition, if you’ve got data in a local Microsoft Analysis Services cube and you’re considering whether to use Sharepoint plus Excel Services, but think that option is a bit heavyweight and expensive, this could be an interesting alternative particularly if you’re looking to share or collaborate on your analysis.

One interesting other feature is the ability to embed these OLAP reports into your iGoogle dashboard. If you look at the top right-hand part of the screen above, you’ll see an Add Gadget to iGoogle button, when you press this it lets you add the OLAP report into your iGoogle dashboard alongside your email, calendar and so on.

Now this is quite a neat way of sharing your analysis or making it available in a rudimentary dashboard or portal, although as usual “the devil is in the detail”. I couldn’t see a way of adding this to my Google Apps dashboard (which I’d need to do to be able to use my Google Apps email, calendar and so on) and you are quite restricted on the layout, I couldn’t get the report above to spread over the whole page, or run in two columns, for example. But if you’ve got some data in Google Spreadsheets, or want to start sharing some Microsoft Analysis Services data but don’t want to stump up for Sharepoint, it’s an interesting alternative.

The news hit the wires last night that Oracle were buying (certain assets of) Hyperroll, a technology company out of Israel with a head office in the USA. Hyperroll have been around for as long as I’ve been in Oracle BI consulting, and I knew them originally as offering an aggregation layer on top of Oracle Express (back in the pre-Oracle OLAP, pre-Compression days) that sped up cube rollups. According to Edward Roske they offered something similar to Essbase, until Hyperion brought out something suspiciously similar that became the Aggregate Storage Option. The introduction of ASO by Hyperion led to a patent infringement law suit that was in the end settled out of court, but that led to technology sharing and an investment in Hyperroll by Hyperion, which is probably where the germ of this deal came from.

These days, Hyperroll focused more on aggregating relational data sources, and branched out in to complementary areas such as real-time loading of financial management applications. From looking at the various press releases and FAQs, it looks like Oracle plan to position Hyperroll in the BI Foundation Layer within their EPM product stack, speeding up relational and OLAP queries and presumably supporting trickle-feed updates to their EPM applications.

This looks like quite a neat fit, and if they can sell it as some kind of “BI Server” aggregation layer you’ve got the potential for MOLAP-style query and rollup speeds across the whole range of OBIEE and EPM data sources. Given Hyperroll’s position in the market, their past support for Oracle OLAP/Express and Essbase and the relational nature of the BI Server this seems quite a clever acquisition on Oracle’s part, and it wouldn’t surprise me to see high-level mentions of it at Open World as part of the 11g BI and EPM stack. In terms of integration, I’d expect it to follow a similar pattern to TimesTen, where it’s initially offered as a stand-alone option to Oracle BI Enterprise Edition initially, with full integration coming a few years later on perhaps as part of the integrated BI/EPM stack that Oracle are working on.

More details can be found in the press release, FAQ and in this set of slides.

The following is a great article on Self Service BI. Click here to view the article.

Don’t say “cherry picking” to people at information-intensive businesses like banks, airlines, and telecommunications companies. You can spoil their lunch if a big customer has just run off to a competitor.

Mark Albala says he has a tool that will warn of such a move. He’s president of InfoSight Partners, and he’s about to offer a new Twitter and blog sniffer to companies in the upper-midmarket and bigger.

Customers often precede their moves with questions about the competitor. Asking “what do you know about …?” on Twitter could be the first and last hint. Albala’s tool monitors such chatter on social media, industry blogs, and other external sources to know when something’s up.

“Your surprises aren’t all coming from your own data,” he says, “but from outside your organization.”

This thing will push notifications to iPhones and Blackberries. It won’t be expensive, he says. He expects the beta to launch in November.

Researching definitions for each of the WRS default roles as part of the Advanced Metadata Security course.

The roles are documented in the SAS 9.1.3 Intelligence Platform: Web Application Administration Guide, Second Edition on page 130.  What it says is:

“By default, everyone who can log on to SAS Web Report Studio can view, edit, and create new reports.

To implement security, each user of SAS Web Report Studio can be assigned to one or more standard roles. A user’s role assignments determine which SAS Web Report Studio menu items are available to that user.

• By default, all SAS Web Report Studio users implicitly have the role. However, if you explicitly assign any members to the role, then only the explicitly-assigned members will have the role. This enables you to start using SAS Web Report Studio immediately after installation, yet still have the ability to restrict user access when locking down your deployment.

nt>

• “>Each role is a superset of the preceding role. For example, members of the “WRS Report Author” role have all the permissions that apply to the “WRS Report Consumer”.

font>

• “>Once you explicitly assign members to a role, you must explicitly assign membersto each superset role. For example, if you assign members to the “WRS ReportAuthor” role, then all of the subsequent superset roles (in this example, “WRSAdvanced User”) also become explicitly-assigned roles. The reason is that WRSAdvanced User is a superset of WRS Report Author.
• “>Once you explicitly assign members to a role, then any user who is not assigned to a role, or who has no metadata identity, can only view reports and manipulate reports (for example, select new data items to view in report objects).

e=”NewCenturySchlbk-Roman” size=”2″>

>WRS Advanced User

>WRS Administrator

>WRS Prohibited

colspan=”2″>*By default, WRS Report Authors can schedule reports, though you can change the default behavior and limit the scheduling feature to WRS Advanced Users. To do this, in your LocalProperties.xml file, specify true for the schedulingRequiresAdvancedUserRole property.

A

… well, at least they bought some of HyperRoll, and frankly, it’s about time. It all started back in the day when HyperRoll found a better way to roll-up Essbase data. Originally, there was scuttlebutt about Hyperion buying HyperRoll, but (so the story goes), the price was too high. So instead, Hyperion decided to create ASO which bore eerie similarities to HyperRoll. There was a nasty lawsuit involving Hyperion infringing on HyperRoll’s patent that got settled eventually for millions.

By that time, though, Hyperion had ASO firmly entrenched in Essbase customers world-wide and HyperRoll somewhat disappeared from the Hyperion landscape. They made a reappearance recently with their LiveLink technology that essentially let you build Essbase cubes from HFM data really, really fast. In some cases, it was almost real-time. I saw a demo of it where a user made an elimination entry in HFM and it was visible in an Essbase cube literally 3 seconds later.

So what exactly did Oracle buy? It’s unclear at the moment. HyperRoll’s website proudly proclaims “Oracle Buys HyperRoll.” Oracle’s release is a bit more coy on what exactly they purchased:

On September 29, 2009, Oracle announced it has agreed to acquire certain assets of HyperRoll. HyperRoll is a leading provider of financial reporting acceleration solutions with interactive reporting capability to support complex data calculation requirements. Oracle Enterprise Performance Management (EPM) already enables rapid, accurate and secure financial consolidation and reporting. With the addition of HyperRoll’s reporting acceleration technology, customers are expected to be able to execute a faster financial close.

HyperRoll’s employees are expected to join Oracle, and will bring domain expertise in multi-dimensional technologies and algorithms for large scale financial and business intelligence systems.

I posted earlier about the removal of the Public Kiosk in SAS 9.2 / Portal 4.2.

All the feedback I got stated that they turned off the Public Kiosk in SAS 9.1.3 / Portal 2.x as a matter of course.

Just noticed a SAS Tech support notice “Enabling unchallenged access to content in SAS® Information Delivery Portal 4.2” which outlines how to allow access to the portal without the need to login.

So obviously a few people still wanted it.

I’m very excited about the upcoming Big Data Summit in New York City on Thursday evening (October 1st). Sponsored by Aster Data, Microstrategy, and Informatica, we have an incredible speaker lineup including LinkedIn, comScore, and Colin White from BI Research. Check out the Facebook page for the event here.

To kick off the festivities, we’re holding a live webinar earlier in the day at 9 a.m. US Pacific time. Colin White and myself will be discussing Hadoop and data warehousing - how they’re similar, how they’re different, and what they can be used for (both separately and together). In fact, we’ll be making an important announcement of a new product offering that you won’t want to miss. If you haven’t already, I urge you to register by clicking here.

Mark it on your calendar - something “big” is coming on October 1st.

• Paso 1. Conocerse mutuamente.
• Paso 2. Desarrollar una relación.
• Paso 3. Definir roles y responsabilidades
• Paso 4. Establecer comunicación regularmente y de forma abierta, sincera.

• Soñando el mismo sueño, todos tenemos que tener los mismo objetivos, remar en la misma dirección
• Eliminando las capas de interferencia, muchas veces se ponen “personas intermedias” que trasladan las necesidades de negocio a los de IT. Esto hay que eliminarlo, se necesita un contacto directo para que ambos mundos se pongan a trabajar como partners, se necesita incentivar el contacto entre ellos, se necesita hacer sesiones de “un día en la vida del comercial” o “un día en la vida del analista de Business Intelligence” para que nosotros veamos lo difícil que es vender y que ellos vean lo difícil que es transformar datos en información y en indicadores.
• Confiando en que cada uno hará bien su tarea “confía en que el cocinero cocine”, que nadie este constantemente metiendo las narices en el trabajo del otro.

Sistemas Decisionales, algo mas que BI

 
Oracle Database 11g Release 2 introduces the successor to the good old Connect By based hierarchical querying, called Recursive Subquery Factoring. The basics are described in a previous article: http://technology.amis.nl/blog/6104/oracle-rdbms-11gr2-goodbye-connect-by-or-the-end-of-hierarchical-querying-as-we-know-it. This article will show some additional examples of using this recursive subquery factoring syntax.
The essence of this recursiveness: the subquery consists of two queries unioned […]

A

Just a short note to let you know that there are just two days left until the early-bird rate expires for our Training Days event later in October.

8217;ve got three days of intermediate to advanced-level instruction on OBIEE, Essbase, ODI and the BI Applications taught by myself, Venkat and Christian Berg. If you’re serious about OBIEE and want to go beyond what’s in the manuals and on the standard public courses, and want to spend some time with three of the best OBIEE developers in Europe (well, two plus me) then make sure you book now before the discount expires.

I’ve been talking with Dan Vlamis about a new book that he, Chris Claterbos and some of the Essbase team from Oracle have been putting together on Oracle OLAP and Oracle Essbase. Oracle Press were kind enough to send over a review copy the other week, and so I’ve been taking a look at it over the past few days in advance of it arriving in the shops.

“Oracle Essbase & Oracle OLAP: The Guide to Oracle’s Multidimensional Solution (Osborne Oracle Press Series)” is a joint effort written by Dan Vlamis and Chris Claterbos, who covered the Oracle OLAP side, and Mike Nader, Michael Schrader, Dave Collins, Floyd Conrad and Mitch Campbell covering the Essbase angle. What the book tries to do is to cover OLAP on the Oracle platform in general, initially going through some general OLAP concepts, then going through the two product offers and how they complement each other, and then diving deeper into the architecture, cube creation process and maintenance of each of the two OLAP platforms.

I’ve read a few Essbase books recently, most notably Edward Roske and Tracey McMullen’s “Look Smarter Than You Are With Essbase 11″, and there’s a fair bit of difference between how Essbase is handled in this book and the one by Edward and Tracey. The “Look Smarter” book is a bit more of a denser read, starts off with manual building of Essbase cubes and spends a lot of time on the internals, storage types, creation of rules files and so on. This book takes a bit more of a “product managers” approach to Essbase at the start, spending quite a bit of time positioning it vis a vis Oracle OLAP and going through some deployment scenarios and case studies. The “hands-on” part starts off by covering Essbase Studio first (the “Look Smarter” book only really covers this in passing at the end) which makes it probably a better book for beginners new to Essbase. Coverage of Essbase development is pretty comprehensive, and there’s chapters on usage with OBIEE (though more the front-end, not the importation and modeling of cubes in the semantic model) and some content on Planning, HFM and so on. This is a great intro to Essbase, and particularly useful if you’re considering the technology and not quite sure how it fits in with the rest of Oracle’s Fusion Middleware and BI/DW stack.

I was particularly keen to read the Oracle OLAP chapters though, as even thought I’ve worked with the product for many years, there are many Express-era concepts such as sparsity, dimension ordering, OLAP DML and so on that I could benefit from some of Dan and Chris’ thoughts on. These concepts are covered pretty well in chapter 3 of the book, with updates for 11g, though some of the newer concepts in 11g such as cube scripts, the new expression syntax, SQL access and partitioning could do with a bit more expansion – I guess this is the downside of sharing a book with Essbase, but it does mean there’s plenty of scope for an advanced, Oracle OLAP specific follow-on book if Dan and Chris have the time. One thing that is particularly useful though is coverage of the cube organized materialized view feature new to 11g – although there was a flurry of publicity around this around the launch of 11g, I haven’t seen a huge amount of uptake and the coverage and explanation of this should hopefully bring this to more people’s attention.

As usual for Oracle Press, the book is well edited and laid out and will be a great resource for customers new to OLAP who are wondering which of Oracle’s OLAP solutions to implement. From the Oracle OLAP side, you couldn’t get two more experienced developers and consultants to write a book on it, and I certainly picked up lots of useful tips and clarifications whilst working on the review. From the Essbase side, there’s some useful positioning stuff and a good introduction to cube building using Essbase Studio, and I’d say if you bought this book plus Edward and Tracey’s one as well, you’d have Essbase development pretty well covered off. Dan tells me that the book should be ready in time for Oracle Open World in two weeks time, so if you’re in SF and can get a chance to pop into the bookstore in Moscone West, you should be able to pick up a copy “hot off the presses”.

So i’m attending the advanced metadata security course for SAS 9.1.3 at the moment.

It is a complex area and really does my head in.  But here is a way that (I think) I understand how it works.

There are two different rules that are applied to any metadata object:

1) Direct - When security is applied directly to an object (i.e in Management Console you HAVE ticked a metadata security box for the object, so its white or green)

2) Inherited - When security is not applied to an object, it is inherited  (i.e in Management Console you have NOT ticked a metadata security box for the object, so its grey)

So here is how I think the rules for each of these works:

1) Direct - when Security is directly applied to an object.

The identity hierarchy model applies.  This means it looks at the levels permissions have been applied to the object and whether it is applied as an ACT or ACE.

• Permissions closest to user wins
• A deny always wins over a grant, if at the same level (and type)
• ACE’s always win over ACT’s

Or in something I can understand:

• If you ask your mother and she says yes, then you ask your father and he says no, the answer is no (deny at same level)
• If you ask your mother and she says yes, then you ask your grandfather and he says no, the answer is yes (permission closest to user)
• If you ask you mother (ACE) and she says yes, then you ask your teacher (ACT) and she says no, then the answer is yes (ACE always wins)

2) Inherited - When security is not applied to an object

The inheritance model applies.  This means every permission is placed in a bucket and if a grant is found anywhere, you have got it.

• Levels don’t count
• A single grant always wins over any and all denies
• ACE’s  vs ACT’s have no impact

Or using the same analogy:

• You are at a family reunion, you yell out a question, somebody yells yes, the answer is yes

And the last thing I learnt was if somebody said no (i.e you can’t see something as a user) then if you ask why, the only answer you will get is “because I said so”.

Which means it is very very hard to find out what stopped you seeing it.  Although I believe Paul at metacoda.com has some cool tools in the pipeline to help with this.

* Thanks to Adam Player for a great course and even better analogy's!