Perspective of Basis in BI...

Background & Dialog Processes
A process that runs with dialogs, means with screen processing is call dialog processing and it is a foreground processing. The process which runs with out a dialog involvement and which runs without manual interference is background processing. The process here is managed by background processors. The background job, as the name suggests completes the job in the background, therefore allowing the user to continue accessing the application.A batch process comes into picture when two or more background jobs run as a batch. Like when you trigger a load the back ground process runs with as a batch process.
Background users are - access to Privileges to execute Background Jobs (reference users)
Dialog users are -Normal users to execute the interactive tasks.

RZ10-If a job is run as a dialogue it takes up a SAP session so as a user you will be unable to do any other work. There is also a time limit when you run a job as dialogue which if reached will not complete the job. This can be change in RZ10.

SM50 - Allows you to identify how many processes you have set up for dialogue (DIA) or background (BGD). If either are reached it can cause performance issues.

RSBATCH - Setting Parallel Processing of BI Processes
BI background management enables you to process the BI load and administration processes in parallel. You can configure the maximum number of work processes for BI processes. In parallel processing, additional work processes for processing the BI processes are split off from the main work process. The parallel processes are usually executed in the background, even if the main process is executed in dialog. BI background management is instrumental in attaining a performance-optimized processing of the BI processes. In addition to the degree of parallelism, you can specify the servers on which the processes are to run and with which priority (job class). BI background management thus supports an efficient distribution of the work processes in order to reduce the system load when processing the BI processes. Another advantage is that logs are created for monitoring the BI processes.

Suggestions:
1) First check the resource availability.
2) Check the extraction process how fast the records are generated in the source, how much time the background job is taking to finish. If that is fine then check how the idocs are moving from source to BI.
3) And also if the extraction job finishes in time then check the dialogue processes available at that time.
4) Check and find the difference at peak and free times by running differently(just a comparison).
4) Check the data packet size at BI level and Source level.
5) Check is there critical transformations/routines at BI level.

Reference

BI Jargon

There are countless of 'BI jargon' used in communicating the technical aspect of datawarehousing and these terms are not to be taken for granted if we want to truly appreciate the art and value of a good business intelligence architecture.

Some of the terms I want to list down are:
LSA - Also known as the next generation of datawarehousing stands for Layer Scalable Architecture. Most organizations with global footprint are considering this to be the way how the datawarehouse architecture should be.LSA consist of 5 main layers namely Data Acquisition Layer,Corporate Memory Layer, Propagation Layer (contain original 'unflavored' data ), Transformation Layer and Reporting Layer. This article explains everything about LSA.

Timestamped record - Discrete vs Continuous

Nonvolatile records - Data in datawarehouse is not subject to change in contrast with OLTP

Corporate Direction in BI

Data is the asset of an organization and making sense out of the data is invaluable for an organization. Each organization will decide its own BI strategy that is aligned with the business strategy and goals.The fundamental of a successful BI environment is based on the ability to maintain tight management over the data warehouse architecture while making the best use of information.

I stumbled upon a good article written by Prashant Pant , a senior Deloitte BI Consultant which point out the 10 essential components of a successful BI Strategy. I found a particular interesting point he highlighted which is called the metadata roadmap. This is so due to the nature of my work to address any potential impacts in a new project or a new enhancement to the existing system. It stated that metadata explains how, why and where the data can be found, retrieved, stored and used in an information management system. Metadata management like the one my company is having for infoobjects is crucial especially when the objects are shared across different BI systems. The attributes has to be correctly set to the master data and is able to differentiate the disparate systems. If not there won't be a single version of truth and user's would be confuse with the reported numbers when they examine granular data in different ways.

Check out Prashant's article

My Sap BI Blog

I started this blog because I needed a place to act as a 'repository' to the BI information I can retrieve from my day to day experience as a SAP BI Consultant. Plus I thought this would be ideal for me to churn out what I understood and to shed some lights to some of the work related issues.