计算机分类
v Micro/personal computer: individual
v Minicomputer: multi-user systems
v Mainframe Computer: has more I/O devices and larger storage capacities
大型机设备发展:
1960中期s360
1970早期s370
1970后期3030
1980早期4300、3080
1980后期3090
90年代s390,es9000,z系列
大型机操作系统发展:
v DOS/VSE: Disk Operating System/DOS/Virtual Storage Extended
v OS/360: 为s360系统定做,但是和DOS不兼容
v PCP: IBM-Primary Control Program,只允许一个程序在任何时间运行
v OS/MFT: Multiprogramming with Fixed Number of Tasks(MFT)允许多个程序同时运行,内存调用之前必须分配固定空间
v OS/MVT: Multiprogramming with Variable Number of Tasks(MVT)动态分配
v OS/VS1 and SVS:OS/VS1和MFT相似,只是分区是放在虚拟空间,同理OS/VS2和MVT类似,OS/VS2也被称作Single Virtual Storage(SVS)
v MVS: Multiple Virtual Storage
v Z/OS: It can run Java and web based applications.
大型机操作系统的特性:
v Virtual storage: the computer uses disk storage as an extension of real storage. Other data and instructions can be placed temporarily on disk storage, and recalled into main storage when needed.
v Multiprogramming: the computer lets more than one program execute at the same time
v Spooling: Simultaneous Peripherals Operations On Line.(队列同步的操作)Intercepts and redirects printer output to a disk file. Each programs' output stored separately.
v Batch processing(批处理): Work is processed in units called jobs. A job may cause one or more programs to execute in sequence. One of the problems that arise when batch processing is used is managing how work flows through the system. To manage this in the multi-user system, the Job Entry Subsystem (JES) processes each user's job in an orderly fashion.
Ø !!!differences between Batch processing and Interactive processing:
Time sharing: Time sharing makes on-line, interactive processing possible.通过终端直接输入命令与系统直接通信,Time-share processing通常称为Foreground Processing(前台处理),batch job processing is called Background Processing
v
大型机基础体系:
Channel是处理体系的一部分,它和控制单元通信,并加入了I/O设备。Channel从MVS处理I/O请求,当I/O操作完成后中断CPU。允许CPU处理I/O操作上的迭代。
v Computer System Controller: This component of the CPU controls all access outside the CPU. Its functions include coordinating all references to main storage for both instructions and data and overseeing the execution of instructions.
v Processor Complex: Processor complex refers to one of the servers which runs, controls, and delivers all the services of the machine. There may be more than one processor complexes.
v Uniprocessor: A Uniprocessor has one processor executing tasks, In a Uniprocessor, a single process has access to storage and to the channel subsystem.
v Multiprocessor: The term refers to the ability to have more than one task executing at the same time, not just two or more tasks running in the Processor Complex at the same time, but two instructions executing at the same point of time.
v Channels: The channels are processors, slower than the main processor, used to control movement of data from outside the Processor Complex to inside the Process Complex.
– provides a path between a processor and an I/O device. ( 4.5 MB/sec, 400-foot )
– Each channel can connect to upto 8 control units each of which is connected to an I/O device.
– A channel is intelligent ( a CPU in itself )
v ESCON( Enterprise System Connection )is a set of interrelated hardware and software products and services which provide fiber optic cabling transmission and reception.(ESCON是提供光纤传输和接收的相关硬件或者软件产品)
v Sysplex: SYSPLEX stands for “Systems Complex”. MVS/ESA version 4 introduces the concept of SYSPLEX. ---SYSPLEX是MVS系统的一个集合,它用某些硬件和软件进行工作,SYSPLEX和传统大型计算机系统的主要区别在于在SYSPLEX有不断增长的潜力,以及实用性的不同。SYSPLEX增加了那些可以协作的处理单元和MVS操作系统的数量。
I/O设备:
• Unit record devices(单元记录设备)---将每条记录作为独立物理单元进行处理
– card devices & printers
• Magnetic tape devices(磁带设备)
– sequential access only
• Direct Access Storage Device(直接存取设备)(DASD)
– disk drive, DASD, auxiliary, .secondary storage
– random access possible
DASD:
data format:
Fixed block data format: the blocks are of fixed size.
Count Key data format (CKD format):
-Count Area: Each block has an eight-byte area which contains flags, the address of the record, the length of the key, and the size of the key area (if any) and the size of the data block.
-Key Area:(可选):The key was used by two of the old data set architectures- Basic Direct Access Method (BDAM) and Indexed Sequential Access Method record.通常没有多大的商业价值,因为它包含在逻辑记录里,而不是物理模块外。There are some specialized data sets that are keyed in MVS. The VTOC (Volume Table of Contents) on a DASD volume is a keyed data set. Partitioned data sets have a directory at the beginning of the data set. The directory is a keyed area. The CVOL (Central Volume) catalog structure also had a keyed area
-Data Area:(可选但是通常出现)包含数据的物理模块,每个物理模块可以包括多个逻辑记录
CKD ( count-key-data) devices:
– Store data in variable-length blocks
– Each data block preceded by a count area and a key area
( which should be met before data in dir of rotation )
– Gaps to separate count, key and data areas.
DASD: control units
• Each type of DASD requires two types of control units to connect to a channel:
– String controller:
• attaches a group of DASDs of same type ( a string )
– Storage control:
• connects upto 8 DASD strings to a channel
The most popular storage control is the 3990. It can attach two string and can support multiple channel connections.
Data communications equipment:
Terminals
• 3270 terminals are the standard.
• It is a subsystem of many terminals, printers and controllers
• Terminals can be emulated on PCs which accesses the mainframe over wide area links.
概念和术语:
Real Storage(实际存储器) 包括main(主存储器), expanded(扩展存储器) and auxiliary(辅助存储器) storage. The CPU stores in any of these three as pages.主存储器在CPU的运行速度下工作,有固定大小;扩展存储器是主存储器的扩展,DASD对应的就是扩展存储器;数据必须在主存储器里执行。
Virtual Storage(虚拟存储器)
Address Space(地址空间)An address space is a complete range of addresses that can be accessed by a processor. 主存储器包括许多独立的存储区域,每个区域能存储1个字节的信息,这些区域的地址由内存中第一个区域地址决定。计算机的最大地址空间由能表示地址空间的最大数字决定。
– System 370 : 24 bit addressing, 2^24 =16 MB max The 16 MB line is called “Line”.
– 370 XA, ESA 370,390 : 31 bit addressing, 2^31=2GB max. The 2 GB line is called “Bar”.
– Z series – 64 bit addressing
Multiple Virtual Storage (MVS) MVS模拟了存储空间,同样利用实际存储器模拟地址空间,它们相互独立。DASD的实际存储器和存储器区域叫做page data sets (页面数据集),用来联合模拟虚拟存储地址空间。MVS能够模拟无限的地址空间。尽管MVS能同时支持多个地址空间,但是CPU一次只能读取一个。每个后台任务或者用户被授予自己的地址空间,所以每个任务或用户能独立的访问其它任务或用户16MB或2GB的地址空间。CPU能够访问那个地址空间地指令和数据,直到MVS给予任务或用户另外的地址空间。
Paging(分页)分页是单一的4kb的信息从主存储器到扩展存储器或辅助存储器的物理运转。MVS divides virtual storage into 4K (4096 bytes) sections called pages.
• Pages are 4K sections on virtual store (DASD).
• Page frames are 4K sections in real store (main memory).
• Page table maps pages to page frames.
• Page slots: the DASD area used for virtual storage, called a page data set, is divided into 4K (4096 bytes blocks) page slots, each of which holds one page of virtual storage.
• Page fault when program refers to data in a storage location not in real store.
• Page-in when MVS loads a new page into a page frame
• Page-out when data in a page-frame is written back to DASD.
• OS code responsible for paging cannot be paged out.
Expanded memory(扩展存储器): Acts as large buffer between real storage and page data sets (virtual store). System/390 or ESA processors now include a special type of memory called expanded storage. When a virtual storage page must be paged out, the processor moves the page's contents to expanded storage. This transfer occurs at CPU speeds rather than at DASD speeds, so the operation is instantaneous.
Swapping … (交换技术)
• Transfer of entire address spaces in and out of virtual store.
• Swapping is like paging, only at a higher level and across address spaces.
Swapping is the physical movement of all 4K pieces of information for a single job from main storage to expanded or auxiliary storage and back again. MVS periodically transfers entire address spaces in and out of virtual storage so that they are temporarily unavailable for processing, which is called swapping.
When an address space is "swapped out", its critical pages (the ones that contain the tables that keep track of the location of each virtual storage page for the address space) are written to a special data set called a swap data set. Later, when the system can accommodate the job again, the address space is "swapped in" so that it can be processed again.
Program modes(程序模式)
• Real mode
– Not pageable or swappable
– OS part responsible for implementing virtual memory, always resident
• Virtual mode
– Other processes that can be paged or swapped
v Some parts of the operating system are responsible for managing virtual storage,certain programs that communicate directly with channel devices or those have certain time dependencies, should run in real mode.
MVS/370-Memory map …
Recent System/370 family processors provide 31-bit addressing rather than 24-bit addressing. Programs that are written in 24-bit addresses should be able to run properly on processors that provide 31-bit addresses. In order to provide the compatibility and also to reap the benefits of 31-bit addressing, processors that use 31-bit addresses and the operating systems that support them provide a full 2GB address space, but they include special provisions for addresses below 16MB.
The conversion to MVS/XA provides 1.4 to 2.4 MB of VSCR, which comes from splitting the nucleus. MVS/XA does not completely remove the Virtual Storage Constraint. As the MVS/XA Operating System grows, more Virtual Storage will be needed for the Operating System. Most of the new code and data areas will be "above the line", the real limitation being application programs. As they are converted to fully utilize the 2 GB address space, then VSCR disappears.
The main difference from MVS/370 is that the private area starts after the Prefixed Save Area (PSA). In MVS/370, the Nucleus is in "low storage".
The address space is typically divided into three parts:
The system area: The System area is also called Nucleus(内核) holds the parts of storage that are required to be in central storage all the times.
– OS programs & data, common to all address spaces, nucleus, real mode
The private area: The private area is the region where the user’s programs run.
– User region + unallocated
The common area: The common area contains the OS routines that are common to all address space.
– OS programs & data, common to all address spaces
Memory map - XA and ESA
• First 16 MB can be addressed using either 24 or 31 bit addressing
– address space logically divided at 16 MB line
• No system area
– MVS nucleus combined with other OS data in common area
The total addressable memory in the 24 bit MVS/XA is 16 MB.
The MVS/ESA is 32 bit OS.
To provide compatibility the 32 bit is divided into 31 + 1.
The 31 bit is used for addressing and the 1 bit to classify between XA and ESA.
MVS/ESA - Dataspace
• ESA lets a job or user create one or more 2GB address spaces that can be used to hold large amounts of data
• Contents of dataspaces managed directly by user programs
• Dataspaces reside in normal virtual store subject to paging and swapping.
v MVS/ESA has extended virtual storage by incorporating data spaces into the processing environment. Data spaces have taken some of the burden of keeping data in virtual storage off of address space. Data spaces also allow greater accessing speeds by reducing I/O. This serves to free up large areas of the address space for other uses. Having additional address space for programs allows us to incorporate more functions in them without impairing(削弱) performance.
v There are two types of data spaces. Both may be 2GB in size and this entire amount may be used for data or programs stored as data. The first type is a regular data space. It is backed by main, expanded and auxiliary storage. Data that resides in a regular data space may be address space.
v Data spaces contain user data and user Application program.
Difference between Address space and Data space:
MVS/ESA - Hiperspace (high performance space).
• Similar to dataspaces
• Contents of hiperspaces managed by MVS/ESA and made available to user programs in 4KB units
• Hiperspaces reside only in expanded storage and are never in real store
• Facilitates hiperbatch to improve performance of certain types of batches
– hiperbatch transparent(透明的) to JCL that invokes them
v Regular data spaces are byte addressable whereas Hiperspace is 4K addressable. Data cannot be accessed from Hiperspace. The data should be moved to Regular data space or address spaces so that Application programs can access them.
The differences between data spaces and Hiperspaces:
v Hiperspaces are used by a new MVS/ESA feature called hiperbatch. This feature improves the performance of certain types of batch jobs by copying entire data sets into hiperspace where they can be accessed concurrently by several batch jobs.
By Michael_Mao