IBM Parallel Sysplex: Difference between revisions

Revision as of 16:14, 29 March 2009

Revision as of 02:43, 28 April 2009

Line 1:

In computing, a '''Parallel [[Sysplex]]''' is a [[computer cluster|cluster]] of [[IBM mainframe]]s acting together inas a single system image, usually with [[z/OS]]. AUsed for disaster recovery, Parallel Sysplex combines data sharing (same disk under control of sharing structures in a Coupling Facility) and [[parallel computing]] to allow a cluster of up to 32 computerssystems to share a workload for [[high performance computing|high performance]] and [[high availability]]. Parallel Sysplex provides [[horizontal scaling]].

{{Mergefrom | IBM Geographically Dispersed Parallel Sysplex | Talk:IBM Parallel Sysplex#Merge proposal |date=March 2009 }}

[[File:GDPS.svg|thumb|360px|Schematic representation of a [[Geographically Dispersed Parallel Sysplex]] (GDPS)]]

==Sysplex==

In computing, a '''Parallel [[Sysplex]]''' is a [[computer cluster|cluster]] of [[IBM mainframe]]s acting together in a single system image, usually with [[z/OS]]. A Parallel Sysplex combines data sharing (same disk under control of sharing structures in a Coupling Facility) and [[parallel computing]] to allow a cluster of up to 32 computers to share a workload for [[high performance computing|high performance]] and [[high availability]]. Parallel Sysplex provides [[horizontal scaling]].

In [[IBM]] [[mainframe computer]]s, a '''Systems Complex''', commonly called a '''Sysplex''', allows multiple processors to be joined into a single unit, sharing the same ''Sysplex name'' and Couple Data Sets. Put another way, a Sysplex is a single logical system running on one or more physical systems. Sysplexes are often isolated within a single system, but Parallel Sysplex technology allows multiple mainframes to act as one.

Components of a Sysplex include:

The forerunner to Parallel Sysplex was '''Virtual Coupling''', a technique which allowed up to 12 [[IBM 390]] systems to execute jobs in parallel. The true Parallel Sysplex was introduced with then-new mainframe models in April 1994<ref>http://www.redbooks.ibm.com/redbooks/pdfs/sg244356.pdf System/390 Parallel Sysplex Performance - IBM Redbook. Retrieved 17-09-2007.</ref>. IBM continues to improve and enhance Parallel Sysplex.

* A Sysplex Timer which synchronizes all member systems' clocks;

* [[Global Resource Serialization]] (GRS), which allows multiple systems to access the same resources concurrently, serializing where necessary to ensure exclusive access;

* Cross System Coupling Facility ([[IBM XCF|XCF]]), which allows systems to communicate [[peer to peer]];

* Couple Data Sets (CDS);

==Parallel Sysplex==

[[File:GDPS.svg|thumb|360px300px|Schematic representation of a [[Geographically Dispersed Parallel Sysplex]] (GDPS)]]

The forerunner to Parallel Sysplex was '''Virtual Coupling''', a technique which allowed up to 12 [[IBM ESA/390]] systems to execute jobs in parallel. The true Parallel Sysplex was introduced with then-new mainframe models in April 1994.<ref>http://www.redbooks.ibm.com/redbooks/pdfs/sg244356.pdf System/390 Parallel Sysplex Performance - IBM Redbook. Retrieved 17-09-2007.</ref>. IBM continues to improve and enhance Parallel Sysplex.

Major components of a Parallel Sysplex include:

* [[Coupling Facility]] (CF or ICF) hardware, allowing multiple processors to share, cache, update, and balance data access;

* [[Sysplex Timer]]sTimers or [[Server Time Protocol]] to synchronize the clocks of all member systems;

* High speed, high quality, redundant cabling;

* Software ([[operating system]] services and, usually, [[middleware]] such as [[IBM DB2|DB2]]).

The Coupling Facility may be either a dedicated external system (a small mainframe, such as a [[System z9]] BC, specially configured with only coupling facility processors) or integral processors on the mainframes themselves configured as ICFs (Internal Coupling Facilities).<ref>{{cite Bothweb are popular|url=http://www.pcmag.com/encyclopedia_term/0,2542,t=Coupling+Facility&nbsp;&mdash;i=40413,00.asp there|title=Coupling areFacility minorDefinition technical|publisher=PC advantagesMagazine.com and|accessdate=April 13, 2009 }}</ref> It is recommended that at disadvantagesleast betweenone external CF andbe ICFused installations.in Fora achievingparallel highestsysplex.<ref>{{cite availabilityweb without|url=http://www-ti.informatik.uni-tuebingen.de/os390/sysplex/sysplex/couplfac.pdf unnecessary|title=Coupling spending,Facility IBM|accessdate=April generally13, recommends2009 combining}}</ref> a singleA externalParallel CFSysplex withhas anat ICF.least However,two withCFs [[zand/OS]]or 1ICFs for redundancy.8 andEvery highermainframe participating in a Parallel Sysplex does not need an ICF-only implementationor isits own external CF&nbsp;&mdash; mainframes merely attach, via cables, to the external CFs quiteor viableICFs.

Server Time Protocol (STP) replaced the Sysplex Timers beginning in 2005 for System z mainframe models z980 and newer.<ref>{{cite web |title=Migrate from a Sysplex Timer to STP |url=http://publib.boulder.ibm.com/infocenter/zos/v1r9/index.jsp?topic=/com.ibm.zos.r9.e0zm100/sttostp.htm |publisher=IBM |accessdate=April 15, 2009 }}</ref> A Sysplex Timer is a physically separate piece of hardware from the mainframe<ref>{{cite web |title=Sysplex Timer |url=http://www.symmetricom.com/resources/compliance-certifications/sysplex-timer/ |publisher=Symmetricom |accessdate=April 15, 2009 }}</ref>, whereas STP is an integral facility within the mainframe's micocode.<ref>{{cite web |title=IBM Server Time Protocol (STP) |url=http://www-03.ibm.com/systems/z/advantages/pso/stp.html |publisher=IBM |accessdate=April 15, 2009 }}</ref>

A Parallel Sysplex has at least two CFs or ICFs for redundancy (or at least one of each). Every mainframe participating in a Parallel Sysplex does not need an ICF or its own external CF&nbsp;&mdash; mainframes merely attach, via cables, to the external CFs or ICFs.

Server Time Protocol (STP) replaced the Sysplex Timers beginning in 2005 for newer mainframe models. While the Sysplex Timers are physically separate machines, STP is an integral feature within the mainframe itself. With STP and ICFs it is now possible (and common) to construct a complete Parallel Sysplex installation solely with two connected mainframes and cabling. Moreover, a single mainframe can contain the internal equivalent of a complete physical Parallel Sysplex, souseful therefor isapplication notesting technicaland requirementdevelopment topurposes.<ref>{{cite haveweb a|url=http://www.zjournal.com/index.cfm?section=article&aid=308 multi-frame|title=MVS ParallelBoot SysplexCamp: atIBM leastHealth forChecker application|first=John testingE. and|last=Johnson development|publisher=z/Journal |accessdate=April purposes.15, 2009 }}</ref>

* [[==Geographically Dispersed Parallel Sysplex]]==

Parallel Sysplex enables another major capability: online maintenance. For example, with Parallel Sysplex it is possible (and common) to configure [[DB2]] for z/OS to deliver continuous business service, permitting whole version upgrades and patches while DB2 continues to run and serve users.

'''Geographically Dispersed Parallel Sysplex''' ('''GDPS''') is an extension of Parallel Sysplex of mainframes located, potentially, in different cities. GDPS includes configurations for single site or multiple site configurations:<ref>{{cite conference |first=Riaz |last=Ahmad |date=March 5, 2009 |title=GDPS 3.6 Update & Implementation |publisher=SHARE |___location=Austin, TX |url=http://ew.share.org/proceedingmod/abstract.cfm?abstract_id=19145 |accessdate=April 17, 2009 }}</ref>

* GDPS/HyperSwap Manager: It is a synchronous [[Peer to Peer Remote Copy]] (PPRC) technology for use within a single data center. Data is copied from the primary storage device to a secondary storage device. In the event of a failure on the primary storage device, they system automatically makes the secondary storage device the primary, usually without disrupting running applications.

==See also==

* GDPS/PPRC: It is a synchronous data mirroring technology (PPRC) that can be used on mainframes {{convert|200|km|mi}} apart. In a two-system model, both sites can be administered as it it were one system. In the event of a failure of a system or storage device, recovery can occur with limited or no data loss automatically.

* GDPS/XRC: It is an asynchronous [[Extended Remote Copy]] (XRC) technology with no restrictions on distance. XRC copies data on storage devices between two sites such that only a few seconds of data may be lost in the event of a failure. If a failure does occur, a user must initiate the recovery process. Once initiated, the process is automatic in recovering from secondary storage devices and reconfiguring systems.

* GDPS/GM: It is an asynchronous [[IBM Global Mirror]] technology with no restrictions on distance. It is designed to recovery from a total failure at one site. It will activate secondary storage devices and backup systems.

* GDPS/MGM & GDPS/MzGM: These are configurations for systems with more than two systems/sites for purposes of disaster recovery. GDPS/MGM and GDPS/MzGM are based on GDPS/PPRC and GDPS/XRC, respectively.

==See also==

* [[Geographically Dispersed Parallel Sysplex]]

* [[System z]]

* [[LPAR]]

* [[IBM Tivoli System Automation]]

==References==

Line 31 ⟶ 45:

==External links==

* [http://www-1.ibm.com/servers/eserver/zseries/pso/ IBM Parallel Sysplex site]

* [http://www-1.ibm.com/servers/eserver/zseries/gdps/ IBM GDPS page]

[[Category:IBM mainframe technology]]

Line 36 ⟶ 51:

[[Category:Cluster computing]]

[[Category:Parallel computing]]

[[Category:1994 introductions]]

[[de:Parallel Sysplex]]