Middle School，MMP，OLTP OLAP

The full name of SMP is "Symmetrical Multi-Processing" (Symmetrical Multi-Processing) technology, which refers to a collection of processors (multiple CPUs) on a computer, and the memory subsystem and bus structure are shared among the CPUs. It is a parallel technology that is widely used relative to asymmetric multiprocessing technology. In this architecture, a computer no longer consists of a single CPU, but multiple processors simultaneously running a single copy of the operating system and sharing memory and other resources of a computer. Although multiple CPUs are used at the same time, they behave like a single machine from a management perspective. The system distributes the task queue symmetrically on multiple CPUs, thus greatly improving the data processing capability of the entire system. All processors have equal access to memory, I/O, and external interrupts. In a symmetric multiprocessing system, system resources are shared by all CPUs in the system, and workloads can be distributed evenly across all available processors.
    In the domestic market, the processors of this type of models are generally 4 or 8, and a few are 16 processors. But generally speaking, the scalability of the machine with SMP structure is poor, and it is difficult to achieve more than 100 multi-processors. The conventional one is generally 8 to 16, but this is enough for most users. The advantage of this kind of machine is that its usage is not much different from that of a microcomputer or a workstation, and the programming changes are relatively small. If a program originally written in a microcomputer workstation is to be transplanted to an SMP machine, it is relatively easy to change it. . The availability of the model of the SMP structure is relatively poor. Because 4 or 8 processors share an operating system and a single memory, once the operating system fails, the entire machine is completely paralyzed. And because the scalability of this machine is poor, it is not easy to protect the user's investment. However, the technology of this type of model is relatively mature, and the corresponding software is also relatively large, so a large number of parallel machines launched in the domestic market are of this type. The most common symmetric multiprocessing systems in PC servers typically employ 2-way, 4-way, 6-way or 8-way processors. Currently UNIX servers can support systems with up to 64 CPUs, such as Sun's Enterprise 10000. The most critical technology in SMP system is how to better solve the problem of mutual communication and coordination among multiple processors.

    To set up an SMP system, the first and most critical point is the need for a suitable CPU to cooperate. The CPUs we usually see are used alone, so we can't see any difference between them. However, in fact, it is not unconditional to support the SMP function. You can build a multiprocessing system with a few CPUs at will. Heaven and night talk. To realize the SMP function, the CPU we use must have the following requirements:

    1. The APIC (Advanced Programmable Interrupt Controllers) unit must be built in the CPU. At the heart of Intel's multiprocessing specification is the use of Advanced Programmable Interrupt Controllers (APICs). CPUs complete their communication by sending interrupts to each other. By attaching actions to interrupts, different CPUs can control each other to some extent. Each CPU has its own APIC (which becomes that CPU's local APIC), and also has an I/O APIC to handle interrupts caused by I/O devices, this I/O APIC is installed on the motherboard, but each The APIC on the CPU is indispensable, otherwise it will not be able to handle interrupt coordination between multiple CPUs.
     2. The same product model, the same type of CPU core. For example, although Athlon and Pentium III each have built-in APIC units, it is impossible for them to build an SMP system together, of course, even Celeron and Pentium III, the possibility of that is 0, even Coppermine core Pentium III And Tualatin's Pentium III also cannot build SMP system -- this is because their operating instructions are not identical, APIC interrupt coordination is also very different.
     3. Exactly the same operating frequency. If you want to build a dual Pentium III system, you must have two 866MHz or two 1000MHz processors, you can't use one 866MHz and another 1000MHz to build, otherwise the system will not be able to light up normally.
     4. Keep the same product serial number as much as possible. Even processors of the same frequency with the same core can cause incredible problems due to different production batches. When two production batches of CPUs are running as dual processors, it may happen that one CPU is overburdened, while the other is under less burden, so that it cannot exert its maximum performance, and even worse, it may cause a crash. , the same batch of processors should be selected as far as possible to form an SMP system.

MPP (Massively Parallel Processing), massively parallel processing system, such a system is composed of many loosely coupled processing units, it should be noted that the processing unit is referred to here instead of the processor. The CPU in each unit has its own private resources, such as bus, memory, hard disk, etc. Within each cell there is an instance replica of the operating system and management database. The biggest feature of this structure is that it does not share resources.



-------------------------------------------------- ------------------------------


         Since there are two structures, what are the characteristics of each? Which structure is more appropriate? Usually, the MPP system is a little less efficient than SMP because it needs to transmit information between different processing units, but this is not absolute, because the MPP system does not share resources, so for it, the resource ratio is higher than that of the SMP. There are more SMPs. When the transactions that need to be processed reach a certain scale, the efficiency of MPP is better than that of SMP. This is determined by the proportion of the communication time occupied by the computing time. If the communication time is relatively long, the MPP system will not have an advantage. On the contrary, if the communication time is relatively small, the MPP system can give full play to the advantages of resources and achieve high efficiency. In the currently used OTLP program, the user accesses a central database. If the SMP system structure is used, its efficiency is much faster than that of the MPP structure. The MPP system has shown advantages in decision support and data mining. It can be said that if the operations have nothing to do with each other and the communication between the processing units is relatively small, it is better to use the MPP system. On the contrary, it is not suitable. .

        For SMP, a key factor that restricts its speed is the shared bus. Therefore, for DSS programs, only MPP can be selected, not SMP. When the processing requirements of large programs are greater than the shared bus, the bus has no ability. After processing, the SMP system will fail. Of course, the two structures have their own advantages and disadvantages. It would be best if they could be combined to complement each other's strengths.





Today's data processing can be roughly divided into two categories: online transaction processing OLTP (on-line transaction processing), online analytical processing OLAP (On-Line Analytical Processing). OLTP is the main application of traditional relational databases, mainly for basic, daily transaction processing, such as bank transactions. OLAP is the main application of data warehouse systems, supporting complex analytical operations, focusing on decision support, and providing intuitive and easy-to-understand query results.
OLTP:
Also known as transaction-oriented processing systems, its basic feature is that customers' original data can be transmitted immediately Go to the computing center for processing, and give the processing results in a very short time.
The biggest advantage of doing this is that the incoming data can be processed in real time, and answers can be answered in a timely manner. Also known as real-time system (Real time System). An important performance indicator to measure the online transaction processing system is the system performance, which is embodied in the real-time response time (Response Time), that is, the time it takes for the computer to reply to the request after the user enters the data on the terminal. OLTP is done by the database engine.
OLTP databases are designed to allow transactional applications to write only the data they need to process a single transaction as quickly as possible.
OLAP:
Abbreviated as OLAP, with the development and application of database technology, the amount of data stored in databases has transitioned from mega (M) bytes and giga (G) bytes in the 1980s to terabytes (T) bytes and At the same time, the query requirements of users are becoming more and more complex, which involves not only querying or manipulating one or several records in a relational table, but also querying tens of millions of records in multiple tables. data analysis and information synthesis, the relational database system can not fully meet this requirement. In foreign countries, many software manufacturers have adopted the development of their front-end products to make up for the lack of support of relational database management systems, trying to unify the scattered public application logic, and respond to the complex query requirements of non-data processing professionals in a short time.
Online Analytical Processing (OLAP) system is the most important application of data warehouse system. It is specially designed to support complex analysis operations, focusing on decision support for decision makers and senior managers. Complex query processing of data volume, and provide query results to decision makers in an intuitive and easy-to-understand form, so that they can accurately grasp the business status of enterprises (company), understand the needs of objects, and formulate correct solutions.