Data, data and more data. From the analysis of banking operations to optimize the workload of airline flights. Study of the situation “in real time” is an opportunity to draw conclusions and make a balanced decision in a limited time frame. A great example is the Formula 1 race.
Before each stage, the riders conduct test races. Fireballs are
filled with hundreds of sensors that allow you to monitor the performance of all components and assemblies of the machine.
Sensors per car generate more than 1 GB of data for the race - this is information about the position of the car on the track, speed, gearbox operation and even tire pressure. These data allow you to get the most complete picture of the technical condition of the car and fix faults.
')
By the end of the season, the racing teams accumulate terabytes of data, which are subjected to rigorous analysis. The result is the success or failure of the team in further races. Such tasks are solved storage platforms and processing. For example, in the 2014 Formula 1 racing season, the McLaren team resorted to
SAP HANA technology .
In this article, we would like to dwell on the SAP HANA software and hardware complex and consider the existing
methods for calculating cloud capacity for this type of hosting.
/ Flickr / dima moroz / ccWhat is SAP hosting?
SAP hosting is a service of hosting SAP solutions in the provider’s cloud infrastructure where the supplier undertakes to fulfill the tasks of migration, administration, maintenance and maintenance of SAP applications.
As for SAP HANA, SAP provides the following
interpretation of the product abbreviation: High Performance Analytical Appliance. HANA is a high-performance analytics solution and includes tools such as transaction processing and large data processing tools. For example, SAP HANA provides the ability to automatically read the counters in various social networks.
IT-GRAD
offers SAP HANA on vSphere on special integrated data center equipment. In this case, you get up to 1 TB of memory and 32 physical cores (these are 64 virtual) for a single installation on VMware vSphere.
Cloud Capacity Calculation for SAP HANA
Now let's see how cloud capacity is calculated for SAP HANA hosting. In general, SAP has a number of configurations for SAP HANA and marks them with letters, as if you were choosing a t-shirt for yourself.
Having tried on a T-shirt of size S and finding that it is small for you, you can take a bigger size - M. Replace size M with L and so on. The same is done with resources: if they are not enough, you can always add them. The “sizes” indicated in the table satisfy most of the business tasks based on SAP HANA, however, in addition to them there are several more unique and specific configurations (presented below).
But let's proceed to the assessment of the required resources. Let's start with shared memory. Its size is determined by the amount of data that should be stored in it. Based on this value, the parameters of the disk subsystem and the processor are already being calculated.
In the requirements for RAM there are two components. The first is the static memory used by HANA, determined by the amount of data that needs to be stored in memory. The second - dynamic RAM - RAM, required when downloading new data blocks and applications.
On the recommendation of SAP, you should reserve the same amount of dynamic and static memory, so the total amount of memory will be equal to the amount of static RAM multiplied by two.
RAM
d = RAM
stRAM
total = 2 x RAM
stNext, there is a calculation of disk system resources (HDD). At the same time, we pay attention to the fact that the data set from the database is periodically copied to the so-called data volume in order to save information in the event of an unforeseen accident. The volume of the data volume is calculated according to the formula:
DISK
pers = 4 x RAM
totalYou should also consider the volume of the journal volume, where information on database changes is stored, so that it can be “rolled back” in case of force majeure. Volume is calculated using the following formula:
DISK
log = RAM
totalTotal total disk space is equal to:
DISK
total = DISK
log + DISK
persIn this case, we note that the primary results should be checked regularly to avoid trouble and performance loss with a sudden increase in data volume.
Now turn the resources of the CPU. The required amount of processor resources depends on the number of users. SAP HANA supports 300 SAPS (conditional units of load on iron) for each user. At the same time, the system supports 60–65 simultaneous active users per CPU. If you're interested, you can find more information about SAPS
here and
here . An example of a complete configuration calculation can be found at the link.
To simplify the process of calculating resources for SAP HANA, SAP offers to use the Quick Sizer utility. It allows you to calculate the amounts of RAM, HDD, CPU, and IOPS. The image below shows the results of the Quick Sizer tool. At the same time, the values of the CPU, HDD parameters are determined according to the “T-shirt sizes”, which were discussed earlier. The tool itself is available by
reference .
We also recommend you to pay attention to the SAP Standard Application Benchmarks test
results , which will help you choose the hardware configuration suitable for specific business tasks, since VMware virtualization successfully copes with the maintenance of highly loaded databases and applications that work stably in private and in public virtualized environments.
Other materials from our blog on Habré: