How In-Memory Database works? specifically SAP HANA

SAP HANA is a new technology build by SAP to fill lack of speed some other database were lacking, SAP HANA is a in memory column-oriented, relational database management system which stores all the data in memory and give your business an online analytic process to speed up the calculation, HANA is an acronym for ‘High Performance Analytic Appliance’.

As name suggested High Performance which means all the data stored in memory and give you advantage of higher speed this database is 90% is written in C++ and it’s used by all level of business which requires higher performance both for data analyses and transactional process.

SAP HANA is filling the gap of speed by both hardware and software innovation which means you cannot download and install HANA in any platform it need its own prosperity hardware to work, as you can see in screenshot above that in hardware it can do Multi-core architecture, Massive parallel scaling as well as you can use this in 64 bit addressing system and in the network side it have 100GB through put this hardware innovation will dramatically increase the speed and decrease the cost.

And in the software section SAP HANA will give you both row and column storage which is huge benefit on other database like in other database you can either have ROW or COLMUN store database but in SAP HANA you can have both in one place, you can also have higher compression if you use Column store you have more control over compression over here.

• Row and column store
• Higher Compression
• Partitioning
• No Aggregate Tables ( as everything is stored in memory)
• Insert only happens in Delta

Along with that you can do multiple partitioning in the SAP HANA which not only speed up your queries but also your data will be organized very well, as everything is stored in memory we will not be having any aggregate table and instead happens only in Delta.

Why In-memory database?

Well if you see current situation in business they all need data in real time they don’t want to wait until data is process and loaded in disk database so if you are looking for the speed and real time then you will go for Memory computing which SAP went in that business and created SAP HANA.

If we see the hardware speed in couple of years we see that everything is increased by three digit like memory speed increased by 250 times, CPU speed increased 150 times, network speed increased as well but the main part of hardware that normal database depends on to read and write data is increased only 25 times which is not even close to the speed of other hardware components.

You can read data from memory approximately 1 million time faster than reading from conventional disk and in conventional database the disk read is around 5 Millisecond but in Memory databases it’s 5 Nanosecond which you can see yourself how fast it’s.