Our primary use case for this solution is for coding systems and applications. IBM PowerVM has a separate market segment for systems and servers. It has Intel systems and Power Systems. Intel-based or RISC and CISC-based systems are generic. The standard system is CISC-based, so it's a Compact Instruction Set of Computers. In contrast, Power Systems, which we usually procure, is RISC-based, a Reduced Instruction Set system.
Power Systems has a separate market segment regarding systems or servers. So you have Intel systems, and you have Power Systems. Intel-based or RISC and CISC-based are very generic. The standard system you get is CISC-based systems, so it's a Compact Instruction Set of Computers, whereas Power Systems that we usually procure are RISC-based, which is a Reduced Instruction Set system.
Primarily, they differ in the way they process data. Secondly, we have a workload requiring higher clock speed, so when we buy a processor, we look at multiple things. First, we look at how much clock speed the processor gives, which means the power for a single thread of the processor to process data. So let's say you have two processors, one with one gigahertz of thread speed and a second with 40 hertz. The 40 hertz is going to work four times faster when it is processing data. So clock speed is what matters. So when we need higher clock speeds, which means we have heavily dependent or heavy threads or multiple applications with heavy threads, we go for systems with higher clock speeds, and in that case, IBM PowerVM has the highest clock speed up till now.
Secondly, when you need higher memory bandwidth, data travel from disc to memory and from memory to processor for processing. So if the road connecting your disc with your memory and your professor is more comprehensive, you can track and send multiple higher traffic. The other important factor is that in use cases where you need higher bandwidth, throughput from memory is where you should use IBM PowerVM.
And third, IBM PowerVM has a higher cache processor. So when the data is processed in a processor, it must first travel to the cache. There are multiple levels, levels 1, 2, 3, and so on. So you will need higher cache memory, which means fewer round trips from memory to processor and cache. It means faster performance. So you procure IBM PowerVM or systems alike. We deploy the solution on-premises.
The only problem with the IBM system is improving its hardware licensing model. For example, when you procure a server with 40 physical codes, you must also procure licenses to activate the codes. In Dell or any system, you procure the hardware and have the full right to use the hardware. For IBM, it's different because when you procure a 40-code processor and have 40 physical codes, you only have a license to use 20. As a result, the 20 alert codes will always be in activated mode. Hence, the hardware licensing model could be improved because the licensing model is a bit different from the standard hardware procured.
