SAP BI: Key aspects of designing fault-tolerance BO system – Part I
SAP BI is one of the popular BI system and has lot of production installs across the globe. It is a stable platform and gets simpler and better at every release with the new 4.x Architecture. There are lot of documents and guides on configuring BI system with methodology that proved to stand against time. So what is this blog for?
Well this series is for SAP BI / BO Administrator and Architects to understand the components needed in designing a truly fault-tolerance BO system outside the eco-system. This involves the Hardware, Software, the Network, Infrastructure and the other components that the typical BI Administrator does not get to work on. It also provides holistic overview of the things that are done by System Administrators, Infrastructure Administrator and DBA’s and some things to look for if you are designing one or planning to talk to other teams to get your BI implementation just right.
Parts of BO system
The BO system consists of the following essential components. The diagram has been simplified for the purpose of understanding.
In each layer of the component there are lot of configuration and components involved from Infrastructure to Application layer. Let’s go from basic to the application for each of the component.
File Store Repository
The File Store repository stores all the critical information of all the files needed for the BO system. While we know the file store repository setting is enabled in a service, let’s see what is actually behind the scenes of this simple folder.
The file storage is usually on high performance iSCSI drive connected with fibre optic SAN or SSD disks. It is usually in RAID so in case of disk failure the data is redundant and recoverable. The performance of the disk depends on the type of the disk and controller and the RAID configuration.
Usually the disk is mounted as separate drive on the File Server which is then shared on Network
- The Disk is configured as RAID for hardware failure
- Disk latency is low and throughput is high
The File repository folder is usually a shared path on the network and hosted on a file server. It can be accessed by either UNC path in Windows or a mount drive in Linux. The network path allows multiple BO systems to access the same file repository which is required for a cluster install.
Usually proper folder permissions and are given to the service account on which the BO systems are running. The File Server and the BO systems are on the same network.
- Whether the File repository is shared
- Necessary read/write permissions for the Service account to access the folder in Network
The necessary services for hosting the network share like Distributed file system or Network File system is setup and running. This works in providing the share in the Network
- The necessary services are configured and running
- Network path can be accessed directly
Finally coming to the BI Platform. The UNC or the mount path is specified in the properties of the Input and the Output repository services for all the BO nodes in a cluster.
- The property of the Repository services is updated in all the nodes
The Database is key for CMS to function and performs as the store for it. The Database can be any of the generic databases that are supported by BO. From Oracle to the newest SAP HANA, each has different settings and tuning. But the core concept behind it remains the same.
Infrastructure for the Database are usually memory efficient and have faster disk. The DB files will be hosted on a file server like powerful storage with the processing memory high needed for the DB. Memory are usually DRAM with high frequency.
- The Database is running as per the recommended hardware capacity
- The Database storage is on a redundant file storage
Databases are accessed by the network similar to how the file storage are accessed but are connected by the host and the port. So firewalls play a major role in allowing the database to be communicated from the BO system. Firewall port and services exceptions are used to control and enable the network access of the database from the BO system.
- Firewall exceptions are added for the BO server to access the DB
- Database is configured to support external access via port
Configuration at the Database depend on the specific database. Usually DBA does things like access control on the DB for the BO service account, Partition of data across clusters or tuning the database engine. The database is installed in a cluster so that there is high availability and redundancy.
- Database is installed in a cluster
- Service account and the CMS/Audit DB is created with necessary privilege
- Database is tuned for the BO load
Final step is the BI Platform where the Database connections and drivers are used to connect to the DB and setup during the installation or migration.
- Database drivers and connections are properly configured
- Database details like hostname, port, username , etc. are configured correctly
Continue in the part II of this series where we see the remaining components and their internals. Read SAP BI: Key aspects of designing fault-tolerance BO system – Part II