Cluster ....

It appears that there is a significant change that I had overlooked in the new OBI 10.1.3.2, compared to the Analytics 7.8 version.

I had noticed that there was a new webcat structure, that was easy, but one of our contacts spotted that there is a new way of clustering the new Webcat.

Previously, each Webcat server (BI Presentation server) had it's own copy of the webcat file, held within the Data folder.
If you had two or more webcat servers they would syncronise the contents of the webcat using a utility - something that did not work very well.

Now you have one webcat, irrespective of how many web servers you use. They all point to the same webcat.
This neatly solves the problems you get with Syncoronisation, but introduces one significant flaw - a Single Point of Failure.

What do you do? Put your webcat on a SAN, which has all the RAID you can get. Back up the whole webcat folder every night - obviously!
Now take snapshots of the webcat throughout the day, but be very careful how many you take. Too many will slow the system down.

One thing I have always felt though is that you should cluster for Customers' High Availability requirements NOT for performance.

The majority of performance issues should be dealt with by database design, prudent Caching (with seeding), adequate memory, and ETL design. Only when these have been worked through should you look for that extra 5% from clustered servers.

Long live the King

Siebel Analytics is Dead, long live Analytics.

Actually it's not, it's still a damn good product, but the son of Analytics, OBIEE, is the new King.

Why?

Because I'm now having fun without Siebel ;-)

The new world looks like this:

An Apex application is the portal.
Identity management is by OIM (OID), managed by the Apex aplplication.
SSO uses OID to authenticate users.
Links on the portal, to your dashboards and Answer subject areas.
ETL by OWB
Database design by YOU, customise as you like.

Is the grass greener, oh yes.



Of course, Elvis is still the real King

Mad Cowes

All quiet here due to me being on holiday.

We're having a fantastic week in Cowes.

Best of all we are leading the fleet going into the last day.

Overall Results

Help!

Help!, we need somebody, Help, not just anybody, Heelp!!

Can you tell I've been listening to the Beatles tonight?

Anyway, we really do need more Analytics people. We have OBI and Analytics work coming out of our ears, so if you are looking, or know someone with Analytics then please get in touch. Now.

We need all levels, experts who want to contract through to juniors with one project who want to work as a permanent in the UK. (Work permits can be arranged).

Hardly enough time to work on the book. But don't worry it is progressing.

By the way, do you know how how the get totals working in a report that has TopN (or BottomN) in? Just go to the function dialogue for the field and set Totals to 'None'

adrian.ward@majendi.com

Catching up

As I have been in hiding on a secret client site (with no access to the internet) for a while, I thought that I would go for a large blog today.


To start with, there is some new functionality in 10.1 that you should know about. One feature is called a Presentation Variable. Like repository based variables but it is created in the BI Presentation system.

Bascially you set it up from a prompt. Once set it can be used all over the place, but my favorite is the Title of a prompted report, or column of a prompted report.

Try this

On the prompt, set a variable, which can then be accessed in the Title of the report.
For example, if you select a County of 'Wiltshire in a prompt, you set a variable name called VAR_COUNTY,
then in the Title you can display 'Details for Wiltshire' by changing the title to: Details for @{VAR_COUNTY}{All Counties}.

Neat eh?

In the good old days (last year!) we had to rely heavily on False Titles to perform the above. These may still be useful today. These use Narratives in reports to create what loks like a title. The report uses the same dashboard prompt as the report it is presented with. Using the @1 syntax you can present the value in the column, the there have a title which says something like 'Sheep in Wiltshire'. Formatted properly it will look like a title. Place it above the query in the same section. Viola.
Now I just tend to use the Section names as the titles.






I started using some Oracle dev tools this week, and I'm actually beginning to like them. Call me fickle but I have to say that I may be throwing out TOAD and starting to use Oracle SQL Developer instead. The main reason is that it is free!!






Having a quick look through my emails the other day I came across this message.

Someone asked the following question


Hi,
We have a requirement to switch between the DBs based on the user login. We have two identical/similar dbs. But based on the users, the connection should switch.
Ex: user1 should see the data fram DB1 and user2 should see the data from DB2. Our plan is to keep a single stream of Physical, BMM ,Presentation catalogs and Webcat. Is this possible through Session variables. If yes, how?


Puna replied with this excellent answer


1) Create a table "db_details" in any db with the following details:
Login_Name dbname dblogin dbpassword

2) Populate this table as per your requirements. Create a connection pool in Physical layer which connects to this table (via database)

3) Create three Non-System Session Variables: dbname,dblogin,dbpassword. Create an initialization block with the following SQL Query:
select dbname,dbuser,dbpassword from db_details where Login_Name=':User' and polplute the the above created variables.
Set the connection pool of the initialization block which was created in step 2.

4) Now for the connection pool for which dynamicity is required as per the logged in user, in connection pool of that database in physical layer set the the following properties to:
Data Source Name: VALUEOF(NQ_SESSION.dbname)
User Name: VALUEOF(NQ_SESSION.dbuser)
Password:VALUEOF(NQ_SESSION.dbpassword)

Make sure if you are using ODBC or DSN's then DSN's with the name "VALUEOF(NQ_SESSION.dbname)" already exist in the system.
So now you will be able to switch DB's based on the Login_Name.
Hope this helps.


Nice one Puna.

If you read this (or know them) please send me your details and I'll invite you to our OBI Developers Forum.

Urgent Requirement

We need a Siebel Analytics consultant for a contract in the UK - not in the North :)

As usual, we have more work than people. If you are an Analytics consultant (i.e. know all about rpd development) then please get in touch.

There will be a technical interview, so please don't try bluffing me!

For the right candidate we'll pay £550 per day initially, rising to £600 after three months.


Start in mid July

If you know anyone who may be interested please get in touch.

Adrian
07801 480900
adrian.ward@majendi.com

SE One is here

Oracle have officially launched SE One - Like Siebel Analytics only cheaper.

This is the pricing package we have all been waiting for. There is no excuse for any company to not buy Analytics. This can be used for what Analytics is good at => RAD (DSDM) warehouse projects. Finance departments should take a really good look at replacing their Access Databases with this.

You'd think I had shares in Oracle (I Don't)!

http://searchoracle.techtarget.com/originalContent/0,289142,sid41_gci1262237,00.html

Freedom of the City

I had a great day today. I had to go to the Camberlains court and swear allegiance to the Queen and to the Lord Mayor. We (my wife and I) were then granted the Freedom of the city of London.
I can now drive sheep over tower bridge.

Interesting Note from Oracle

This note was published today. It may explain the questions I have been getting.

"The Standard Cost General table (W_STANDARD_COST_G) and the Exchange Rate General table (W_EXCH_RATE_G) are general reference tables.

Both tables have surrogate keys (ROW_WID). But unlike dimension surrogate keys, these surrogate keys are not referenced as foreign keys by any fact tables. Therefore the general tables do not participate in the star schema. They are not designed for use in reports and queries.

The general tables are intended for use as lookups during Extract-Transform-Load (ETL). For example, the Standard Cost General table provides cost information for a particular product in a particular inventory. The cost information will then be stored in the fact table directly. Similarly, the Exchange Rate General table provides the exchange rate between a pair of currencies for a specific day, given the exchange rate type.

The following are some WIDs in the general tables. They are not necessarily populated and they are not intended for use. These columns are:

W_STANDARD_COST_G.PRODUCT_WID
W_STANDARD_COST_G.CREATED_BY_WID
W_STANDARD_COST_G.CHANGED_BY_WID
W_EXCH_RATE_G.CREATED_BY_WID
W_EXCH_RATE_G.CHANGED_BY_WID

If you need to have a cost dimension, you need to design a different table. It is not recommended that you alter the Standard Cost General table for this purpose. "

Professional Indemnity Schedule

Just so I can point people to the evidence, here is our insurance certificate.

Professional Indemnity Schedule

Oracle Warehouse Builder Documentation

Today was spent installing Oracle databases and Warehouse tools onto one of the Majendi Laptops. It's good to keep doing these to keep your eye in.

A valuable resource if you cannot memorise everything is :

Oracle Warehouse Builder Documentation

A typical question

Please keep sending in your questions, it's interesting to see what challenges you guys (and Girls - Vicky!) face.


Here's a recent one


Hello Adrian,

I need little help.

Our company is implementing Oracle BI 10.x. We don't have much experience in Siebel Analytics/OBI.

We have few fact tables e.g. W_SALES_PICK_LINE_F, which we need/wish to extend. We are not sure about which methodology we should implement.
When we talked with few SA experts they mentioned we need to extend based on ROW_WID. Well, this column doesn't exists anymore in many fact tables, so currently we are thinking of creating ROW_WID on this F table manually and then create WC_SALES_PICK_LINE_FX and take these two tables in physical layer of RPD.

Additially, we need to create few Informatica mappings, again SA experts pointed out to use Image tables. Our data source is Oracle Applications, I don't have much idea about Image tables, but clearly there is not S_ETL_I_IMGAGe or so tables in Oracle Apps

Please guide us as how to proceed on these two issues.

Thanks in advance


and my answer

I am surprised to hear that ROW_WID has been removed from standard vanilla tables. I don't think your system will work without the ROW_WID field. Add the field back in quick! :)

My advice is to extend an existing table, not to create _FX tables. Providing you keep good control on your alter table scripts there are no issues with extending vanilla tables.

Take a copy of the vanilla mappings and customise these in your custom mappings folders. Do not update the vanilla mappings in place - they will be over written in an upgrade.

You can create your own image tables in the oracle db. The alternative is to use a last modified field in the table being imported.

You only need to create immage tables where there is lots of data, typically more than one million rows


Please can you send me more details on the project. It's good to understand where you are coming from, if you send me your rpd I can provide some feedback


Adrian
-------------------------------------------------------
Mobile: 07801 480900

Analytics Projects - Who wants one?

Whilst looking around for my next client I realise that we, as Analytics consultants, need to talk to each other more. Whats more, the consultancies and clients really need a place to go for reliable, experienced Analytics specialists.

I was discussing one new project with Haider and it turns out he worked on it before. It's a small world, and Siebel Analytics has a relatively small client base (my wild guess is 400-500 UK clients, half of which are active).
I know about 75 Analytics consultants. There are probably another 100 out there, perm employees, that may one day become freelance and pop up in the Majendi radar.

At the moment the Agencies have a good control on the Market. There are some that put out adverts for non existant jobs. They are useful to the liquidity of the market and some work hard for their cut of you pay. Just be careful they are not taking too much, the best ones have a fixed percentage with their clients. More than 15% is excessive.

So how do we help consultants? We probably should look at the networking capabilty that's out there, linkedIn, Xing, ecademy, Yahoo, etc, and perhaps set-up an OBIEE/Analytics forum. There we could share thoughts on prices to charge, new projects, help needed, technical stuff, etc. I have started one group on Yahoo, if you fancy joining in, let me know - it's invite only. I'm not sure if this helps the clients though.


Any thoughts?

On the subject of rate. You are what you charge. If you charge a cheap rate, your will be percieved as local quality, if you charge a premium rate you should deliver a premium level of service.
What are those rates? My view is that a certified consultant with 2 years (min 3 projects) Analytics consultant should be charging a minimum of 600 GBP per day. Lead consultants, with 7+ project should be GBP 750 per day. The consultancies will obviously need to charge clients a higher level. Some of the cheaper ones are charging 6-700 euro per day, the better ones, with good experience and efficient work, will be charging 900 plus.
One client I saw recently had a greatdeal with a low cost provider, the problem was they didn;t trust the work, and the estimate for time was nearly double what was really required. The client ends up paying nearly the same for inferior work. Somehow we need to educate that quality is very important to them, not cheapness.

Insurance time

I just had to renew the Professional Indemnity insurance, which this years comes to over 1,300 pounds!

We use Trafalgar, seem to provide good cover, but not sure if they are the cheapest. However, you can do it all online which saves a huge amount of time.

Use the link to get a quote.





Dear *NAME*,

This mail has been sent to you by *YOUR FRIEND* from *YOUR COMPANY* through our website.

Since 1995, Trafalgar Risk Management has focused on providing computer contractors with specialist insurance cover – Professional Indemnity; Tax & VAT Investigation; Employers’ & Public Liability.

Using state of the art computer systems we can provide quotes; accept applications; accept premium payment and allow you to print out all your relevant documentation out online – within minutes!

Furthermore, by streamlining our purchase & renewal procedures, the reduction in administration costs can be passed back to you, as can be seen in the extremely competitive premiums we offer.

We hope the above will be of interest to you and for further information please visit our website through the following link:

http://www.trafalgar.uk.com/index.asp?r=12692

Regards,

Trafalgar Internet Sales
Trafalgar Risk Management
The Lloyds Building
12 Leadenhall Street
London EC3V 1LP
Tel: 0845 685 2610
Fax: 0845 685 2611

Trafalgar Risk Management Limited is authorised and regulated by The Financial Services Authority under registration number 311662.

Trafalgar Risk Management Limited is a member of the British Insurance Brokers Association

FAQ 2289: How Do You Customize Default Header Links in the Oracle BI Presentation Services Interface?

This supportweb note came out today. It's well worth looking at if you need to customize the links in your BI Publisher.

FAQ 2289: How Do You Customize Default Header Links in the Oracle BI Presentation Services Interface?

Majendi Day out pics

The team went for a day out on the Company Yacht!

Skype

Some say "SCIPE" others "SkyPee"

Whichever you say, it's worth getting it. Free calls. We use it for conferencing.

you can find me on the system as username 'adyward'

www.skype.com

Link to Me!

Come on you lot. Let get networking!

The 'WebCatalogServiceSOAPKPI' Outbound Service Web Port

Here's a techie from Justin Townsend. We're working on some interesting ideas and clients (and client bids), where this is relevent.


Using the integrated application (Core Oracle/Siebel modules combined with Analytics & Marketing functionality) relies quiet heavily on the use of web services. These services are invoked in a 'traditional' SOAP interface, providing the communication links between the applications.

In a standard installation of the integrated application, web services are used for: pass-through authentication, execution of Marketing jobs and serving up of Analytics and Marketing resources within the integrated environment.

WebCatalogServiceSOAPKPI, as the name may suggest, serves up resources from the Analytics application which are stored in the web catalog file. Using this service you can access: dashboards, requests, iBots, segments, segment trees and list format definitions.

This WebCatalogServiceSOAPKPI is particularly important where the business has many pre-defined reports (visible on dashboards or via requests). In combination with the definition of a symbolic URL, the web service serves up these pages to the user when they access dashboards related screens in the integrated application.

Symbolic URLs are defined in the core Siebel application; when the application's services are started up, the definitions for these symbolic URLs are stored in the web server cache.

When making changes to symbolic URLs, you need to ensure the cache is cleared to make sure the amended definitions are used by the web service.

Symbolic URLs are defined in:

Administration - Integration > Symbolic URL Administration;

the cache is cleared in:

Administration - Marketing > Servers, Outbound Web Service Port (hyperlink) > Outbound Web Services

What Does Metadata mean

I'm i the middle of writing the book (along with Mark) and I was using the word Metadta alot. Then something happened on a client project which made me realise that few people probably understand what this word actually means.

The OraFaq makes a good start
"Data that is used to describe other data"

But does that clear it up?

How about
"Metadata is information about a particular data set which may describe, for example, how, when, and by whom it was received, created, accessed, and/or modified and how it is formatted. Some metadata, such as file dates and sizes, can easily be seen by users; other metadata can be hidden or embedded and unavailable to computer users who are not technically adept"

Every other definition I see just says "Data about data"

So that clears it up then!