PDI was founded in 1980 by Carl Rosendahl with a small loan from his father. He was joined in 1981 by Richard Chuang and in 1982 by Glenn Entis. Richard and Glenn wrote the foundation of the in-house computer animation software that was to be used for the next two decades. They started work on 3D software at the end of 1981, and 3D production started in the fall of 1982. The initial goal of the company was “Entertainment using 3D computer animation”. By the time PDI reached its 25th anniversary in 2005, it had completed over 1000 projects and grown to over 400 employees.

Early years: 1980–1987

The first computer at PDI was a DEC PDP 11/44 with 128 kilobytes of memory. This was a lot of memory given that the computer had only 64 kilobytes (16-bits) of address space. It had a 20 megabyte disk. Attached to this was a $65,000 framebuffer which had a resolution of 512×512 and was 32 bits deep.

The first 3D image rendered at PDI was done on March 12, 1982. The image was simply a 4 by 4 by 4 grid of spheres of varying colors. The spheres were not polygonal, they were implicitly rendered and were fully anti-aliased. The resulting image was 512 by 480 by 24 (8 bits for red, green and blue channels) which took 2 minutes to render.

The PDP-11 was soon replaced by a DEC VAX-11/780 and later PDI shifted to another superminicomputer called the Ridge32. This machine was 2–4 times faster than the VAX-11/780 at a fraction of the cost.

The original in-house software evolved into a large suite of tools which included a polygon scan-line renderer

The initial investment to start the company was $250,000 (about $600,000 in 2005 dollars). Its original offices were in Sunnyvale, California working out of a garage owned by Carl’s father. PDI moved to its first real offices in 1985 (Sunnyvale), to its second offices in 1995 (Palo Alto) and to its current location in Redwood City in 2002. The growth of the company was financed solely through profit. The company was run as an open book, monthly financial reviews were shared with the entire company and a detailed monthly financial report was released. Money was never taken out of the company which maintained a 7% investment in R&D. PDI was debt-free when acquired by DreamWorks in 2000. This was quite an accomplishment for a low margin service business with a lot of risk.

PDI always kept an active 5-year long-term plan which was updated every year.

PDI’s first client was Rede Globo, Brazil’s largest TV network. This gave PDI the major client they needed to fund the building of most of the early software. This also sent PDI into the business of TV motion graphics and logo animation (flying logos). PDI designed some early show openings and other special projects for Rede Globo. The software written was also given to Rede Globo and is the only time the in-house software was given to another company. The contract ended in the mid-1980s, but Rede Globo continued to use the software for many years.

Most of the 1980s were spent creating broadcast graphics for most television networks around the world. PDI was working concurrently for ABC, CBS, NBC, HBO, Cinemax, MTV, VH1, TNT and Showtime. PDI focused on direct to video production as opposed to film output being done at other early studios. PDI modified the interface to a Sony BVH-2000 using parts put together from a trip to a toy store in order to do single-frame recording. All the rendering was done on fields at 60 or 50 frames per second (depending of the video broadcasting standard used locally).

PDI controlled a large percentage of this market during this time and they were really the first mass producer of computer animation. One year producing two major networks’ graphics packages meant specifically rendered images for over 400 local television stations. Some of the early production contracts included Globo, ”Entertainment Tonight” (produced for Harry Marks), ABC Sports 84 Olympic promos, and NBC News.

PDI planned and proposed a feature-length CG animation film in 1985. Unfortunately, they were unable to raise the funding needed to produce it.

While not the first computer graphics studio founded, PDI is the longest lasting. It has outlived all the other studios which existed in the early 1980s. Of the many reasons for this, one is that PDI never went into significant debt by purchasing expensive hardware. While other studios purchased Cray supercomputers, PDI only bought cheaper hardware, treating it as a commodity which would soon be replaced, enabling lower operating costs.

In transition: 1987–1992

PDI’s early focus was on network TV productions since they captured over 50% of that market in 1985. However, in 1990, PDI introduced the digital film scanning process. This process was used to popularize automated rig removal and image touch-up. PDI was also instrumental in introducing performance animation for theme parks, ads and movies. This started with a project for a real time performance character for Jim Henson Productions.

During these years of transition, PDI moved away from the motion graphics market and focused their attention on commercials and 3D visual effects for feature films. Notable among the commercials was the first Pillsbury Doughboy created in CG. Pillsbury was the first company to move an established icon to CG. Before this, all previous animated commercials were done with stop-motion. Other notable commercials include the “Bud Bowl” and “Scrubbing Bubbles” spots.

Early in the 1990s, Thaddeus Beier and Shawn Neely developed a method for morphing that resulted in a much more natural and expressive morph. This technique is called “feature-based morphing”. PDI used this technology to create various well-known sequences, including the Exxon car-into-tiger morph and the extended morph at the end of the “Black or White” music video from Michael Jackson. These morphing jobs were very easy to do with PDI’s software and the effect was in high demand. The algorithms invented by Beier and Neely were published at the annual SIGGRAPH conference and are now the basis of most image morphing tools. For many people, their first exposure to these algorithms was the SGI IRIX software called “Elastic Reality”.

PDI broke into the feature film visual effects business with contributions to ”Batman Forever”, ”The Arrival”, ”Terminator 2”, ”Toys”, and ”Angels in the Outfield”. At the time, the strengths of PDI included character animation, lip synch, rendering effects, the aforementioned rig removal and cleanup, and performance animation.

During this era PDI transitioned from the Ridge32 computer to SGI workstations running IRIX. They were not alone in this transition as most of the industry followed suit.

1987–1995: Character animation

Early in 1990, Tim Johnson and Rex Grignon officially formed PDI’s Character Animation Group with the mandate to develop a group of artists with the creative and technical skills needed to produce a feature-length CG-animated film. The group originally consisted of Johnson, Grignon, Raman Hui, Glenn McQueen, Beth Hofer, Dick Walsh, Karen Schneider, Simon J. Smith and Eric Darnell. Under the auspices of the group, PDI’s commercial character animation skills grew and numerous notable short films were produced. Among these are ”Gas Planet” (1992), ”Sleepy Guy”, ”Bric-a-Brac” (1994), ”Gabola the Great” (1997), ”Fishing” and ”Fat Cat on a Diet”.

This character group set the company off in a fun new direction that set the basis for development goals during this period. The shorts (short films) were a way to develop animation techniques as well as being a test bed for software and pipeline procedures and flow.

PDI has always allowed animators to pursue individual products and shorts. This has produced several award-winning short films in this category. Some of the more notable productions are ”Opera Industrial” (1986), ”Chromosaurus”, ”Cosmic Zoom”, ”Burning Love” (1988), and ”Locomotion” (1989).

By 1992, PDI was seriously looking for a partner to produce feature-length animated films. PDI’s first CG feature was planned in 1985, and Hollywood was still not ready to say “Yes”. PDI landed the “Last Halloween” TV special which won them an Emmy Award for the CG characters in the otherwise live-action special with Hanna-Barbera. This turned into PDI’s first 3D Character Animation pipeline in 1991. Using this pipeline they did a 3D stereo Daffy Duck for Warner Brothers and a CG Homer and Bart Simpson for the 1995 ”The Simpsons” Halloween episode “Homer3″.

The result of all these projects was, finally, a movie deal with DreamWorks SKG in 1995 to make the movie ”Antz”. At this time DreamWorks purchased a 40% share of PDI.

Glen Entis left PDI for the game industry in 1995, first joining DreamWorks Interactive as CEO. When Electronic Arts purchased DreamWorks Interactive, he moved to their Vancouver office to set up their next-generation games research group. He is a founding board member of Los Angeles’ Digital Coast Roundtable, and is chairman of the Academy of Interactive Arts & Sciences.

1995–present: feature films

In 1997 Simon J. Smith, a former member of The Mill in London and a CGI veteran, founded PDI’s layout department. The layout department produced PDI’s first feature film ”Antz” which was released by Dreamworks Pictures in 1998. This was followed by ”Shrek” in 2001. The layout department is now a major part of the PDI/DreamWorks film making process.

After the success of ”Antz”, Carl Rosendahl sold his remaining interest in PDI to DreamWorks. He left PDI in February 2000 to become managing director for Mobius Venture Capital, a board member of iVAST, an MPEG4 software company, and several other Bay Area technology firms. This sale essentially united the two studios, PDI and DreamWorks, into a single entity which went public a few years later as DreamWorks Animation (DWA). The PDI studio is now known as PDI/DreamWorks. Animators at PDI work on projects based at the PDI studio, but also assist in DWA projects based in the Glendale DWA studio. Today, the two studios essentially act as a single unit.

In 2008, Richard Chuang, the last of the initial three, left the company to pursue his own ventures.

Animated films

PDI/DreamWorks has produced six box-office hits with ”Antz” (1998), ”Shrek” (2001), ”Shrek 2” (2004), ”Madagascar” (2005), ”Shrek the Third” (2007) and ”Madagascar: Escape 2 Africa” (2008). With $442.2 million USD in box-office ticket sales, ”Shrek 2” is currently the fourth highest grossing film of all-time in the United States.

PDI won their first Oscar for Best Animated Feature for ”Shrek” in 2002.

Technical awards

PDI/DreamWorks has won four Scientific and Technical Academy Awards. The first was awarded to Les Dittert, along with others, in 1994 for work in the area of film scanning. The second was awarded to Carl Rosendahl, Richard Chuang and Glenn Entis in 1997 for the concept and architecture of the PDI animation system. This award in particular recognized their pioneering work in computer animation dating back to the founding of PDI 17 years earlier. Nick Foster was given an award in 1998 for PDI’s fluid animation system (flu) and in 2002 Dick Walsh was given one for the development of PDI’s Facial Animation System.

Adapted from the Wikipedia article Pacific Data Images, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Solaris has implemented a separate LWP layer since version 2.2. Prior to version 9, Solaris allowed a many-to-many mapping between LWPs and user threads. However, this was retired due to the complexities it introduced and performance improvements to the kernel scheduler.

UNIX System V and its modern derivatives IRIX, SCO OpenServer, HP-UX and IBM AIX allow a many-to-many mapping between user threads and LWPs.

Adapted from the Wikipedia article Light-weight process, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Bacula supports many features used by large scale, production networks, including:

Network options

*TCP/IP – client–server communication uses standard ports and services instead of RPC for NFS, CIFS, etc.; this eases firewall administration and network security

*CRAM-MD5 – configurable client–server authentication

*GZIP – client-side compression to reduce network bandwidth consumption; this runs separate from hardware compression done by the backup device

*TLS – network communication encryption

*MD5/SHA – verify file integrity

*CRC – verify data block integrity

*PKI – backup data encryption

Client-options

*POSIX ACL – needed to restore Windows NT ACE’s and Samba servers

*Unicode/UTF-8 – cross-platform filenames

*VSS – calls Microsoft’s snapshot service

*LVM – pre-script setup for Linux/UNIX snapshot

*LFS – backup files larger than 2GiB

*raw – backup devices without a filesystem

Backup devices

*pooling – allocates backup volumes according to job needs and retention configuration

*spooling – writes backup data to spool until target backup medium is allocated so jobs can continue uninterrupted

*media-spanning – such as spanning tapes

*multi-streaming – write multiple, simultaneous data streams to the same medium

*ANSI & EBCDIC – IBM compatibility

*Barcodes – reading tape barcodes in libraries

*most tape drives, including DDS, DLT, SDLT, LTO-1,2,3,4

Client OS

The client software, executed by a “file daemon” running on a Bacula client, on many operating systems

, including:

*Linux – most major distributions, including: CentOS, Debian, Fedora, Gentoo, Mandriva, OpenSUSE, Red Hat and Ubuntu.

*Solaris

*FreeBSD – all released versions

*NetBSD

*Windows (File daemon supported on all 32 and 64 bit Windows OSes)

*Mac OS X

*OpenBSD

*HP-UX

*Tru64

*IRIX

Adapted from the Wikipedia article Bacula, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

*1988–1991 – Early 3D modeling and animation functionality, early workflow definition, some 3D packages derived from CAD/engineering-focussed packages

*1991–1995 – New generation animation tools based on research on robotics and physics, intense innovation period, games focus, openness to SDKs, ended with live theater experiments

*1994–1999 – Acquisition by Microsoft. Mutation period, industry shakeup & buyouts, appearance of new platforms (PCs), drastic price drops, new generation architectures in development

*1998–2000 – Acquisition by Avid. New generation products based on new architectures

*2000–2003 – Production market mutation, more price drops, focus on productivity, non-linear animation systems

*2007–2008 – ICE (Interactive Creative Environment) in XSI 7, acquisition by Autodesk

1987

* Softimage President Daniel Langlois and engineers Richard Mercille and Laurent Lauzon begin development of the company’s 3-D application software. The user interface begins to take shape and would remain largely unchanged for the next 16 years of the product’s life.

1988

* Introduction of Creative Environment 1.0 at SIGGRAPH. A first for the industry, the software would offer modeling, animation and rendering in a single integrated environment.

* Release of Creative Environment v0.8, closely followed by the 1.0 release.

1989

* Creative Environment 1.65 adds texture mapping.

1990

* Creative Environment V

** new animation tools

** introduction of the new concept of object constraints

** new Dopesheet editor

** adds spline modeling

1991 – Actor Module

*Creative Environment 2.5 with Actor Module

** The Actor Module introduces Inverse Kinematics, a concept coming from robotics, to the animation market

**Bones

**Inverse Kinematics

**Flexible Skin (Envelopes)

**Rigid-Articulated Body Dynamics

* The Actor Module was awarded a Technological and Scientific Academy Award for its innovation

* Films: Terminator 2- Judgment Day (Academy Award for Best Visual FX)

* Director of sales Richard Szalwinski leaves to found Discreet and re-distribute Animal Logic’s image compositor Eddie

1992 – Eddie

*Creative Environment 2.52

**Motion Capture (Channels)

**SDK (DKit)

*Paint and Compositing package introduced with Softimage|Eddie

*Films

**Death Becomes Her (Academy Award for Best Visual FX)

* Softimage goes public on NASDAQ

* Acquisition of the compositing product Softimage|Eddie product from Discreet , co-owned and developed with Animal Logic

* Acquisition of Painterly Effects image processing library from ImageWare Research

1993

*Creative Environment 2.6

**Metaclay (“Metaballs”)

**Motion Control

**Clusters

**Shape Animation

**mental ray rendering

**Wave Deforms

**Flock Animation (macro particles)

**Standalone Particle System

**Rotoscopy

**3D Booleans

**Ghost display/Onion skinning

*mental ray and introduced.

* Softimage and mental images announce rendering technology agreement

*Creative Toonz debuts. The 2-D animation package automates the more tedious tasks involved in 2-D cel animation, such as inking-&-painting, while still maintaining the look of hand-drawn images and characters

*Beginning of Softimage Digital Studio on the Silicon Graphics IRIX operating system

*Films

**Jurassic Park (Academy Award for Best Visual FX)

1994 – Microsoft Softimage

* Microsoft acquires Softimage

* Creative Environment 2.65

**Complete new File and Database Management

**New Topological scene graph update

**Structure Keys

**Extended Constraints

**Expressions

**Ghost display/Onion skinning

**Animation par Shapes

**Toon rendering

*IDEAS: Interactive Developer’s Entertainment Authoring Software with ProPlay and ProPlay Plus solutions, includes: Softimage Creative Environment, Eddie compositing, video-effects software, distributed ray tracer and 3-D particles kit.

*Films

**Star trek: Generations / Industrial Light & Magic (1994)

**The Flintstones/ Industrial Light & Magic (1994)

**The Mask / Industrial Light & Magic (1994)

1995

*Softimage|3D 3.0

*NURBS Surfaces and Modeling

**Trims

**Instances

**Relational modeling

**Qstretch deform (squash and stretch)

**Custom hotkey remapping (swift keys)

**Spreadsheet

**Polygon Reduction (1st generation)

**Games Features

**Advanced polygonal modeling tools

**2D/3D Paint + UV Texturing + Painterly Effects, Color Reduction

**Game Export / filtering / on target viewing (SEGA Saturn)

**SI Live Virtual Theater (NAB 95)

*Introduction of Softimage|3D Extreme

** includes Osmose, Virtual Theater (real time capture and virtual set compositing) and mental ray.

*Softimage Toonz 3.5 and SOFTIMAGE|Eddie 3.2

*Films

**Casper

**Babe (Academy Award for Best Visual FX)

**Balto / Amblimation (1995)

**Casper / Industrial Light & Magic (1995)

**Juge Dredd / (1995)

**Jumanji / Industrial Light & Magic (1995)

**La cité des enfants perdus / BUF Compagnie (1995)

1996

* Softimage|3D 3.5

**Windows NT platform support

**User Data

**Image Library

**Saaphire SDK

**Ambulate, Stepmaler

* First prototype of Softimage|DS on Windows NT

* “Sumatra”(code name), the product that would become Softimage|XSI , and RenderFarm unveiled (a Sumatra module never released).

* Films and Games

**Virtua Fighter(game), Dragonheart, Independence Day (Academy Award for Best Visual FX),101 Dalmatiens / Industrial light & Magic (1996),12 Monkeys / Peerless Camera (1996), Basket spatial / Industrial Light & Magic (1996), Star Trek – First Contact / Industrial Light & Magic (1996), Pinocchio / MediaLab (1996), Malgré Picasso / Peerless Camera (1996), Draco : la légende du dernier dragon/ Industrial Light & Magic (1996), Eraser / Mass Illusion (1996), Joe’s Apartment / Blue Sky (1996), Island of Docteur Moreau / Digital Domain (1996), Chasseurs de fantômes / Weta Ltd. (1996), Mission : Impossible / Industrial Light & Magic (1996), Mars Attack! / Industrial light & Magic (1996), T2-3D / Digital Domain (1996)

1997

*Dominique Boisvert, Réjean Gagné, Daniel Langlois, and Richard Laperrière won a Scientific and Engineering Award for the development of the ‘Actor’ component of the Softimage computer animation system

* SOFTIMAGE|DS 1.0

* SOFTIMAGE|3D 3.7

**Sony Playstation Export / Import / Viewer + Attribute Editors

**Colors at Vertices (painting, OpenGL, Softimage + mr rendering, Saaphire)

**Direct3D Export / Import

**RenderMap (baking light maps)

* Films and Games

** Riven, Titanic (Academy Award for Best Visual FX), Alien : Resurrection / Blue Sky, Spawn / Industrial Light & Magic (1997), Anastasia / Fox Animation Studio (1997), Batman et Robin / BUF Compagnie (1997), Contact / Sony Pictures Imageworks / Weta Ltd. (1997), Men In Black/ Industrial Light & Magic (1997), Jurassic Park / Industrial Light & Magic (1997), Fifth Element/ Digital Domain (1997)

* First mention of Twister, a rendering module of Sumatra that would never be released

1998 – Avid Softimage

* Avid Technology acquires Softimage, begins to re-brand many of its visual effects products under the name “Softimage”

* SOFTIMAGE|DS 2.1

* SOFTIMAGE|3D 3.8

** GDK (Game Development Kit / high-level AP)

**dotXSI file format, import/export pipeline (Direct 3D, VRML, 3D Studio)

**GameFilter

**Merge

**Polygon Reduction

**Neural Quantizer (color reduction)

**Animation Sequencer (precursor to animation mixing)

**Audio Track for lip synch

* Announcement of an interactive rendering application called “Twister” , the first module of the future Sumatra product. The company later canceled this separate product to focus on Sumatra.

1999

* “Animation R3defined” (TM) campaign : First presentation of “Sumatra” (code name) as the first “Non-Linear 3D animation” system The Animation Mixer, which allows manipulation of animation as clips on tracks, similar to video non-linear editing system like Avid Media Composer.

* Softimage|3D v3.8 Service Pack 2

** Advanced Rendering – Caustics support, Global Illumination

**Nintendo NIFF toolkit

**Sony PlayStation HMD toolkit

**Bézier curves support

**Surface Continuity Manager (SCM)

**Drop & Slide Points

**GoWithThe Flow qui contraint les objets aux particules

* Films : Movie – The Mummy, The Matrix (Academy Award for Best Visual FX), Stuart Little / Centropolis FX, Fight Club / Pixel Liberation Front / BUF

2000 – Softimage|XSI 1.0

image:SoftimageLogo.png

* Softimage|XSI 1.0 is released

**New generation architecture, user interface, workflow, etc.…

**ActiveScripting

**Interactive Rendering (Render Region)

**Rendering Passes

**Render Tree

**GAP (Generic Attribute Painting)

**Surface meshes (Nurbs networks)

**Non-linear animation – Animation Mixer

**Integrated Particles

* Acquisition of The Motion Factory, Inc

* Softimage|3D v3.9

* Softimage|XSI 1.5

**Adds Polygonal modeling, Subdivision Surfaces, and texturing tools

**Animation Clip Effects/Offsets, Equalizer, Bridge Transitions

**Scripted Operators (scripted plugins)

**Soft-Bodies

**Cloth, Fluids (from Phoenix Tools)

**SDK Object Model (COM)

2001

*Softimage|XSI v1.5 begins shipping.

*SOFTIMAGE|XSI v2.0 unveiled at Siggraph 2001 and later ships

*XSI v2.0

** Introduces fully integrated compositor, based on Avid Media Illusion

** Introduces a Hair and Fur module, based on Joe Alter’s Shave

** Support for Linux (previously, supported unix was SGI IRIX)

** Electric Rain collaborate to bring Flash, EPS, AI and SVG exports to SOFTIMAGE|XSI customers.

*SOFTIMAGE|3D v4.0 unveiled at Siggraph 2001

2002

* Softimage releases the “SOFTIMAGE|XSI Experience” for version 2.0, a free educational software and training kit . Alias had previously released a Personal Learning Edition of Maya

* Softimage|3D v4.0 released

* Softimage|XSI v3.0 [http://www.edharriss.com/xsi/version3.htm User Feature Review]

** An evolutionary update

** Introduces Softimage|Behavior, a procedural animation system, marketed as a new Crowd Simulation engine . It is a re-branded version of The Motion Factory’s product called Motivate

2003

* Softimage|XSI v3.5 with Softimage|Behavior 1.1

** Brings back the Schematic View from Softimage|3D

2004

* Softimage|XSI v4.0 [http://www.edharriss.com/xsi/version4.htm User Feature Review]

** New Rigid Body Dynamics based on ODE

** Character SDK, Custom Display Host, XML-based UI definition, new XGS real time shader pipeline

** New Vector and Raster Paint tool in the compositing module

** Now ships with Syflex

* Softimage|XSI v4.2

2005

* Softimage|XSI v5.0 [http://www.edharriss.com/xsi/version5.htm User Feature Overview] [http://softimage.wiki.softimage.com/history/xsidocs_whatwasnew/new_Version50.htm Official Change List]

** first Windows 64-bit version

** New user interface elements to appeal to Autodesk Maya users

** Integrated Cloth with Syflex 3

** Updated Rigid Body Dynamics with physX

** New view: Shape Manager, for morph shape animation

** dotXSI 5.0

2006

* Softimage|XSI v5.1 [http://softimage.wiki.softimage.com/history/xsidocs_whatwasnew/new_Version51_511.htm Official Change List]

** Autodesk 3DS Max compatible keymap and GATOR plug-in for 3DS Max

** Collada import/export

* Introduction of Softimage|FaceRobot

* Softimage|XSI v6.0 [http://www.awn.com/articles/review/xsi-60-review-focusing-animation-and-other-resources User Feature Overview] [http://softimage.wiki.softimage.com/history/xsidocs_whatwasnew/new_Version60.htm Official Change List]

** Focus on animation tools, animation layers, motion transfer tool (“MOTOR”)

** New views: Material Manager, Material Panelw, Animation Layer Manager

** Elastic Reality morpher in the compositor

** Delta Referencing

2007

* Softimage|XSI 6.5 [http://softimage.wiki.softimage.com/history/xsidocs_whatwasnew/new_Version65.htm Official Change List]

** Mostly a release to adjust price and re-shuffle features between XSI Essential and XSI Advanced.

2008 – ICE, and Sale of Company Assets to Autodesk

* Softimage|XSI 7.0 [http://softimage.wiki.softimage.com/history/xsidocs_whatwasnew/new_Version70.htm Official Change List]

** introduces ICE (“Interactive Creative Environment”)

* October 23td, Autodesk announces intent to acquire all of Softimage 3D business assets from Avid Technologies

* November 18, 2008, Autodesk Completes Acquisition of Softimage

* Company ceases to exist as an entity

* Autodesk Media and Entertainment continues to develop Softimage|XSI product, re-branded as “Softimage”. FaceRobot rebranded Autodesk FaceRobot

Adapted from the Wikipedia article Softimage (company), under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Capacity

XFS is a 64-bit file system. It supports a maximum file system size of 8 binary exabytes minus one byte, though this is subject to block limits imposed by the host operating system. On 32-bit Linux systems, this limits the file and file system sizes to 16 binary terabytes.

Journaling

Journaling is an approach to guaranteeing file system consistency even in presence of power failures or system crashes. XFS provides journaling for file system metadata, where file system updates are first written to a serial journal before the actual disk blocks are updated. The journal is a circular buffer of disk blocks that is never read in normal filesystem operation. The XFS journal is limited to a maximum size of both 64k blocks and 128MB with the minimum size dependent upon a calculation of the filesystem block size and directory block size. Placing the journal on an external device larger than the maximum journal size will cause the extra space to be unused. It can be stored within the data section of the filesystem (an internal log), or on a separate device to minimize disk contention. On XFS the journal contains “logical” entries that describe at a high level what operations are being performed (as opposed to a “physical” journal that stores a copy of the blocks modified during each transaction). Journal updates are performed asynchronously to avoid incurring a performance penalty. In the event of a system crash, operations immediately prior to the crash can be redone using data in the journal, which allows XFS to retain file system consistency.

Allocation groups

XFS filesystems are internally partitioned into ”allocation groups”, which are equally sized linear regions within the file system. Files and directories can span allocation groups. Each allocation group manages its own inodes and free space separately, providing scalability and parallelism — multiple threads and processes can perform I/O operations on the same filesystem simultaneously. This architecture helps to optimise parallel I/O performance on multiprocessor or multicore systems, as metadata updates are also parallelizable. The internal partitioning provided by allocation groups can be especially beneficial when the file system spans multiple physical devices, allowing for optimal usage of throughput of the underlying storage components..

Striped allocation

If an XFS filesystem is to be created on a striped RAID array, a ”stripe unit” can be specified when the file system is created. This maximises throughput by ensuring that data allocations, inode allocations and the internal log (journal) are aligned with the stripe unit.

Extent based allocation

Blocks used in files stored on XFS filesystems are managed with variable length extents where one extent describes one or more contiguous blocks. This can shorten the list considerably compared to file systems that list all blocks used by a file individually. Also many file systems manage space allocation with one or more block oriented bitmaps — in XFS these structures are replaced with an extent oriented structure consisting of a pair of B+ trees for each filesystem allocation group (AG). One of the B+ trees is indexed by the length of the free extents, while the other is indexed by the starting block of the free extents. This dual indexing scheme allows for highly efficient location of free extents for file system operations.

Variable block sizes

The file system block size represents the minimum allocation unit. XFS allows file systems to be created with block sizes ranging between 512 bytes and 64 kilobytes, allowing the file system to be tuned for the expected use. Where a large amount of small files is to be expected, a small block size would typically be used to maximize capacity, but for a system dealing mainly with large files, a larger block size can provide a performance advantage.

Delayed allocation

XFS makes use of lazy evaluation techniques for file allocation. When a file is written to the buffer cache, rather than allocating extents for the data, XFS simply reserves the appropriate number of file system blocks for the data held in memory. The actual block allocation occurs only when the data is finally flushed to disk. This improves the chance that the file will be written in a contiguous group of blocks, reducing fragmentation problems and increasing performance.

Sparse files

XFS provides a 64-bit sparse address space for each file, which allows both for very large file sizes, and for ”holes” within files for which no disk space is allocated. As the file system uses an extent map for each file, the file allocation map size is kept small. Where the size of the allocation map is too large for it to be stored within the inode, the map is moved into a B+ tree which allows for rapid access to data anywhere in the 64-bit address space provided for the file.

Extended attributes

XFS provides multiple data streams for files through its implementation of extended attributes. These allow the storage of a number of name/value pairs attached to a file. Names are null-terminated printable character strings of up to 256& bytes in length, while their associated values can contain up to 64& KB of binary data. They are further subdivided into two namespaces, root and user. Extended attributes stored in the root namespace can be modified only by the superuser, while attributes in the user namespace can be modified by any user with permission to write to the file. Extended attributes can be attached to any kind of XFS inode, including symbolic links, device nodes, directories, etc. The attr program can be used to manipulate extended attributes from the command line, and the xfsdump and xfsrestore utilities are aware of them and will back up and restore their contents. Most other backup systems are not aware of extended attributes.

Direct I/O

For applications requiring high throughput to disk, XFS provides a direct I/O implementation that allows non-cached I/O directly to userspace. Data is transferred between the application’s buffer and the disk using DMA, which allows access to the full I/O bandwidth of the underlying disk devices.

Guaranteed-rate I/O

The XFS guaranteed-rate I/O system provides an API that allows applications to reserve bandwidth to the filesystem. XFS will dynamically calculate the performance available from the underlying storage devices, and will reserve bandwidth sufficient to meet the requested performance for a specified time. This feature is unique to the XFS file system. Guarantees can be ”hard” or ”soft”, representing a trade off between reliability and performance, though XFS will only allow ”hard” guarantees if the underlying storage subsystem supports it. This facility is most used by real-time applications, such as video-streaming.

DMAPI

XFS implements the DMAPI interface to support Hierarchical Storage Management. While this functionality has been ported to the Linux implementations of XFS, it is not yet a part of the mainline Linux kernel source.

Snapshots

XFS does not provide direct support for snapshots, as it expects the snapshot process to be implemented by the volume manager. Taking a snapshot of an XFS filesystem involves freezing I/O to the filesystem using the xfs_freeze utility, having the volume manager perform the actual snapshot, and then unfreezing I/O to resume normal operations. The snapshot can then be mounted read-only for backup purposes. XFS releases on IRIX incorporated an integrated volume manager called XLV. This volume manager has not been ported to Linux and XFS works with standard LVM instead. In recent Linux kernels, the xfs_freeze functionality is implemented in the VFS layer, and happens automatically when the Volume Manager’s snapshot functionality is invoked. This was once a valuable advantage as Ext3 system could not be suspended and volume manager was unable to create a consistent ‘hot’ snapshot to backup a heavily busy database. Fortunately this is no longer the case. Since Linux 2.6.29 ext3, ext4, gfs2 and jfs have the freeze feature as well.

Online defragmentation

Although the extent-based nature of XFS and the delayed allocation strategy it uses significantly improves the file system’s resistance to fragmentation problems, XFS provides a filesystem defragmentation utility (xfs_fsr, short for XFS filesystem reorganizer) that can defragment the files on a mounted and active XFS filesystem.

Online resizing

XFS provides the xfs_growfs utility to perform online resizing of XFS file systems. XFS filesystems can be grown provided there is remaining unallocated space on the device holding the filesystem. This feature is typically used in conjunction with volume management, as otherwise the partition holding the filesystem will need enlarging separately. XFS partitions cannot (as of August 2010) be shrunk in place, although several possible workarounds have been discussed.

Native backup/restore utilities

XFS provides the xfsdump and xfsrestore utilities to aid in backup of data held on XFS file systems. The xfsdump utility backs up an XFS filesystem in inode order, and in contrast to traditional UNIX file systems which must be unmounted before dumping to guarantee a consistent dump image, XFS file systems can be dumped while the file system is in use. This is not the same as a snapshot since files are not frozen during the dump. XFS dumps and restores are also resumable, and can be interrupted without difficulty. The multi-threaded operation of xfsdump provides high performance of backup operations by splitting the dump into multiple streams, which can be sent to different dump destinations. The multi stream capabilities have not been fully ported to Linux yet, however.

Atomic disk quotas

Quotas for XFS filesystems are turned on when initially mounted; this fixes a race window that is present with most other filesystems that first require to be mounted and where no quotas are enforced until quotaon(8) is called.

Adapted from the Wikipedia article XFS, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Physical methods

The earliest methods for generating random numbers — dice, coin flipping, roulette wheels — are still used today, mainly in games and gambling as they tend to be too slow for most applications in statistics and cryptography.

A physical random number generator can be based on an essentially random atomic or subatomic physical phenomenon whose unpredictability can be traced to the laws of quantum mechanics. Sources of entropy include radioactive decay, thermal noise, shot noise, avalanche noise in Zener diodes, clock drift, the timing of actual movements of a hard disk read/write head, and radio noise. However, physical phenomena and tools used to measure them generally feature asymmetries and systematic biases that make their outcomes not uniformly random. A randomness extractor, such as a cryptographic hash function, can be used to obtain uniformly distributed bits from a non-uniformly random source, though at a lower bit rate.

Recently, a team at Bar-Ilan University in Israel has been able to create a physical random bit generator at a 300 Gbit/sec rate, making it the fastest ever .

Various imaginative ways of collecting this entropic information have been devised. One technique is to run a hash function against a frame of a video stream from an unpredictable source. Lavarand used this technique with images of a number of lava lamps. [http://www.fourmilab.ch/hotbits/ HotBits] measures radioactive decay with Geiger-Müller tubes, while [http://www.Random.org Random.org] uses variations in the amplitude of atmospheric noise record

Another common entropy source is the behavior of human users of the system. While people are not considered good randomness generators upon request, they generate random behavior quite well in the context of playing mixed strategy games. Some security-related computer software requires the user to make a lengthy series of mouse movements or keyboard inputs to create sufficient entropy needed to generate random keys or to initialize pseudorandom number generators.

Computational methods

Pseudo-random number generators (PRNGs) are algorithms that can automatically create long runs of numbers with good random properties but eventually the sequence repeats (or the memory usage grows without bound). The string of values generated by such algorithms is generally determined by a fixed number called a seed. One of the most common PRNG is the linear congruential generator, which uses the recurrence

:X_{n+1} = (a X_n + b), textrm{mod}, m

to generate numbers. The maximum number of numbers the formula can produce is the modulus, ”m”. To avoid certain non-random properties of a single linear congruential generator, several such random number generators with slightly different values of the multiplier coefficient ”a” can be used in parallel, with a “master” random number generator that selects from among the several different generators.

A simple pen-and-paper method for generating random numbers is the so-called middle square method suggested by John Von Neumann. While simple to implement, its output is of poor quality.

Most computer programming languages include functions or library routines that purport to be random number generators. They are often designed to provide a random byte or word, or a floating point number uniformly distributed between 0 and 1.

Such library functions often have poor statistical properties and some will repeat patterns after only tens of thousands of trials. They are often initialized using a computer’s real time clock as the seed, since such a clock generally measures in milliseconds, far beyond the person’s precision. These functions may provide enough randomness for certain tasks (for example video games) but are unsuitable where high-quality randomness is required, such as in cryptographic applications, statistics or numerical analysis. Better pseudo-random number generators such as the Mersenne Twister are widely available. Much higher quality random number sources are available on most operating systems; for example /dev/random on various BSD flavors, Linux, Mac OS X, IRIX, and Solaris, or CryptGenRandom for Microsoft Windows.

An example of a simple pseudo-random number generator is the Multiply-with-carry method invented by George Marsaglia. It is computationally fast and has good (albeit not cryptographically strong) randomness properties. (note that this example is not thread safe):

m_w = ; /* must not be zero */

m_z = ; /* must not be zero */

uint get_random

{

m_z = 36969 * (m_z & 65535) + (m_z >> 16);

m_w = 18000 * (m_w & 65535) + (m_w >> 16);

return (m_z
Adapted from the Wikipedia article Random number generation, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

* 1961: IBM delivers the IBM 7030 Stretch supercomputer, which uses 64-bit data words and 32- or 64-bit instruction words.

* 1974: Control Data Corporation launches the CDC Star-100 vector supercomputer, which uses a 64-bit word architecture (previous CDC systems were based on a 60-bit architecture).

* 1974: International Computers Limited launches the ICL 2900 Series with 32-bit, 64-bit, and 128-bit two’s complement integers; 64-bit and 128-bit floating point; 32-bit, 64-bit and 128-bit packed decimal and a 128-bit accumulator register. The architecture has survived through a succession of ICL and Fujitsu machines. The latest is the Fujitsu Supernova, which emulates the original environment on 64-bit Intel processors.

* 1976: Cray Research delivers the first Cray-1 supercomputer, which is based on a 64-bit word architecture and will form the basis for later Cray vector supercomputers.

* 1983: Elxsi launches the Elxsi 6400 parallel minisupercomputer. The Elxsi architecture has 64-bit data registers but a 32-bit address space.

* 1991: MIPS Technologies produces the first 64-bit microprocessor, the R4000, which implements the MIPS III ISA, the third revision of their MIPS architecture. The CPU is used in SGI graphics workstations starting with the IRIS Crimson. Kendall Square Research deliver their first KSR1 supercomputer, based on a proprietary 64-bit RISC processor architecture running OSF/1.

* 1992: Digital Equipment Corporation (DEC) introduces the pure 64-bit Alpha architecture which was born from the PRISM project.

* 1993: Atari introduces the Atari Jaguar video game console, which includes some 64-bit wide data paths in its architecture.

* 1994: Intel announces plans for the 64-bit IA-64 architecture (jointly developed with Hewlett-Packard) as a successor to its 32-bit IA-32 processors. A 1998 to 1999 launch date is targeted. SGI releases IRIX 6.0, with 64-bit support for the R8000 chip set.

* 1995: Sun launches a 64-bit SPARC processor, the UltraSPARC. Fujitsu-owned HAL Computer Systems launches workstations based on a 64-bit CPU, HAL’s independently designed first-generation SPARC64. IBM releases the A10 and A30 microprocessors, 64-bit PowerPC AS processors. IBM also releases a 64-bit AS/400 system upgrade, which can convert the operating system, database and applications.

* 1996: Nintendo introduces the Nintendo 64 video game console, built around a low-cost variant of the MIPS R4000. HP releases an implementation of the 64-bit 2.0 version of their PA-RISC processor architecture, the PA-8000.

* 1997: IBM releases the RS64 line of 64-bit PowerPC/PowerPC AS processors.

* 1998: IBM releases the POWER3 line of full-64-bit PowerPC/POWER processors.

* 1999: Intel releases the instruction set for the IA-64 architecture. AMD publicly discloses its set of 64-bit extensions to IA-32, called x86-64 (later branded AMD64).

* 2000: IBM ships its first 64-bit z/Architecture mainframe, the zSeries z900. z/Architecture is a 64-bit version of the 32-bit ESA/390 architecture, a descendant of the 32-bit System/360 architecture.

* 2001: Intel finally ships its 64-bit processor line, now branded Itanium, targeting high-end servers. It fails to meet expectations due to the repeated delays in getting IA-64 to market.

* 2003: AMD introduces its Opteron and Athlon 64 processor lines, based on its AMD64 architecture which is the first x86 based 64 bit processor architecture. Apple also ships the 64-bit “G5″ PowerPC 970 CPU courtesy of IBM. Intel maintains that its Itanium chips would remain its only 64-bit processors.

* 2004: Intel, reacting to the market success of AMD, admits it has been developing a clone of the AMD64 extensions named IA-32e (later renamed EM64T, then yet again renamed to Intel 64). Intel also ships updated versions of its Xeon and Pentium 4 processor families supporting the new instructions.

* 2004: VIA Technologies announces the Isaiah 64-bit processor.

* 2006: Sony, IBM, and Toshiba begin manufacturing of the 64-bit Cell processor for use in the PlayStation 3, servers, workstations, and other appliances.

Adapted from the Wikipedia article 64-bit, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Universal Storage Platform V Specifications

*Frames (Cabinets) – Integrated Control/Drive Group Frame and 1 to 4 optional Drive Group Frames

*Universal Star Network™ Crossbar Switch – Number of switches 8

*Aggregate bandwidth (GB/sec) – 106

*Aggregate IOPS – Over 4 million

*Cache Memory – Number of cache modules 1-32, Module capacity 8 or 16GB, Maximum cache memory 512GB

*Control/Shared Memory – Number of control memory modules 1-8, Module capacity 4GB, Maximum control memory 28GB

*Front End Directors (Connectivity)

**Number of Directors 1-14

**Fibre Channel host ports per Director – 8 or 16

**Fibre Channel port performance – 4, 8 Gb/sec

**Maximum Fibre Channel host ports – 224

**Virtual host ports – 1,024 per physical port

**Maximum IBM® FICON® host ports – 112

**Maximum IBM ESCON® host ports – 112

*Logical Devices (LUNs) — Maximum Supported

**Open systems 65,536

**IBM z/OS® 65,536

*Disks

**Type: Flash 73, 146, 200 and 400GB

**Type: Fibre Channel 146, 300, 450 and 600GB

**Type: SATA II 1TB, 2TB

**Number of disks per system (min/max) 4-1,152

**Number spare disks per system (min/max) 1-40

*Maximum Internal Raw Capacity – (2TB disks) 2,268 TB

*Maximum Usable Capacity – RAID-5

**Open systems (2TB disks) 1,972 TB

**z/OS-compatible 1TB disks) 931 TB

*Maximum Usable Capacity — RAID-6

*

** z/OS-compatible (1TB disks) 796 TB

*Maximum Usable Capacity — RAID-1+

**Open systems (2TB disks) 1,130TB

**z/OS-compatible (1TB disks) 527.4TB

*Other Features

**RAID 1, 10, 5, 6 support

**Maximum internal and external capacity 247PB

**Virtual Storage Machines 32 max

**Back end directors 1-8

*Operating System Support

**Mainframe – Fujitsu: MSP; IBM z/OS, z/OS.e, z/VM®, zVSE™, TPF; Red Hat; Linux for IBM S/390® and zSeries®; SUSE: Linux Enterprise Server for System z.

**Open systems – HP: HP-UX, Tru64 UNIX, Open VMS; IBM AIX®; Microsoft® Windows Server 2000, 2003, 2008; Novell NetWare; SUSE Linux Enterprise Server; Red Hat Enterprise Linux; SGI IRIX; Sun Microsystems Solaris; VMware ESX and Vsphere, Citrix XENserver

Adapted from the Wikipedia article Universal Storage Platform, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

Unix and Unix-like

In Unix and Unix-like operating systems, kill is a command used to send a signal to a process. By default, the message sent is the termination signal, which requests that the process exit. But ”kill” is something of a misnomer; the signal sent may have nothing to do with process killing. The kill command is a wrapper around the kill system call, which sends signals to processes or process groups on the system, referenced by their numeric process IDs (PIDs) or process group IDs (PGIDs). kill is always provided as a standalone utility, but most shells have built-in kill commands that may slightly differ from it.

There are many different signals that can be sent (see ”signal” for a full list), although the signals in which users are generally most interested are SIGTERM and SIGKILL. The default signal sent is SIGTERM. Programs that handle this signal can do useful cleanup operations (such as saving configuration information to a file) before quitting. However, many programs do not implement a special handler for this signal, and so a default signal handler is called instead. Other times, even a process that has a special handler has gone awry in a way that prevents it from properly handling the signal.

All signals except for SIGKILL and SIGSTOP can be “intercepted” by the process, meaning that a special function can be called when the program receives those signals. The two exceptions SIGKILL and SIGSTOP are only seen by the host system’s kernel, providing reliable ways of controlling the execution of processes. S

Unix provides security mechanisms to prevent unauthorized users from killing other processes. Essentially, for a process to send a signal to another, the owner of the signaling process must be the same as the owner of the receiving process or be the superuser.

The available signals all have different names, and are mapped to certain numbers. It is important to note that the specific mapping between numbers and signals can vary between Unix implementations. SIGTERM is often numbered 15 while SIGKILL is often numbered 9.

Examples

A process can be sent a SIGTERM signal in four ways (the process ID is ’1234′ in this case):

*kill 1234

*kill -s TERM 1234

*kill -TERM 1234

*kill -15 1234

The process can be sent a SIGKILL signal in three ways:

*kill -s KILL 1234

*kill -KILL 1234

*kill -9 1234

Other useful signals include HUP, TRAP, INT and ALRM. HUP sends the SIGHUP signal. Some daemons, including Apache and Sendmail, re-read configuration files upon receiving SIGHUP, so the kill command may be used for this too. A SIGINT signal can be generated very simply by pressing in most Unix shells. It is also common for to be mapped to SIGTSTP, and for (backslash) to be mapped to SIGQUIT, which can force a program to do a core dump.

Related programs

*killall – on some variations of Unix, such as Solaris, this utility is automatically invoked when the system is going through a shutdown. It behaves much like the kill command above, but instead of sending a signal to an individual process, the signal is sent to all processes on the system. However, on others such as IRIX, Linux, and FreeBSD, an argument is supplied specifying the name of the process (or processes) to kill. For instance, to kill a process such as an instance of the XMMS music player invoked by xmms, the user would run the command killall xmms. This would kill all processes named xmms.

*pkill – signals processes based on name and other attributes. It was introduced in Solaris 7 and has since been reimplemented for Linux and OpenBSD. pkill makes killing processes based on their name much more convenient: e.g. to kill a process named ”firefox” without pkill (and without pgrep), one would have to type kill `ps& ax | grep& firefox | grep& -v& grep | awk& ‘{print& $1}’` whereas with pkill, one can simply type pkill firefox.

Microsoft Windows

Microsoft’s command-line interpreter Windows PowerShell, kill is a predefined command alias for the Stop-Process cmdlet.

Microsoft Windows XP, Vista and 7 include the command taskkill to terminate processes. An “unsupported” version of kill was included in several releases of the Microsoft Windows Resource Kits (aka “RezKits”) available for Windows NT 3.x, NT 4.0, Windows 2000 and Microsoft Windows Server 2003, the most efficacious of these being Version 3.5 of Kill, Copyright (C) 1994 Microsoft Corp.

GNU versions of kill have been ported via Cygwin and run inside of the Unix environment subsystem that Microsoft Windows Services for UNIX provides (Microsoft acquired Windows Services for Unix wholesale via their purchase of Softway Systems and their Interix product in September 17, 1999).

Examples

Find all processes beginning with the letter “p” that were developed by Microsoft and use more than 10 MB of memory and kill them:

PS C:>ps p* | where { $_.Company -like “Microsoft*” -and $_.WorkingSet -gt 10MB } | kill -confirm

Confirm

Are you sure you want to perform this action?

Performing operation “Stop-Process” on Target “powershell (6832)”.

[Y] Yes [A] Yes to All [N] No [L] No to All [S] Suspend [?] Help (default is “Y”): n

PS C:>

Microsoft Singularity

Singularity shell, the standard shell for Microsoft Research’s microkernel operating system Singularity includes a kill command to terminate background processes.

Examples

Stop the process with the name “SampleProcess”:

Singularity>kill SampleProcess

Cannot find active process Name=SampleProcess

Stop the process with the process identifier “42″:

Singularity>kill 42

Cannot find active process ID=42

Plan 9 from Bell Labs

Under Plan 9 from Bell Labs, the kill program does not actually perform this termination, nor does it take process IDs. Rather, it takes the actual names of processes and outputs the commands for rc, the shell used by Plan 9, to kill the process.

A similar command provided is called slay, which does the same but for processes that refuse to be killed this way.

Examples

For example, to kill all instances of troff, one types:

kill troff | rc

Adapted from the Wikipedia article Kill (command), under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.

When AppleTalk was first introduced the dominant office computing platform was the PC compatible running MS-DOS. The “TOPS Teleconnector” system enabled MS-DOS PCs to communicate over AppleTalk network hardware; it comprised an AppleTalk interface card for the PC and a suite of networking software allowing such functions as file, drive and printer sharing. As well as allowing the construction of a PC-only AppleTalk network, it allowed communication between PCs and Macs with TOPS software installed. (Macs without TOPS installed could use the same network but only to communicate with other Apple machines.) The Mac TOPS software did not match the quality of Apple’s own either in ease of use or in robustness and freedom from crashes, but the DOS software was relatively simple to use in DOS terms, and was robust.

The BSD and Linux operating systems support AppleTalk through an open source project called Netatalk, which implements the complete protocol suite and allows them to both act as native file or print servers for Macintosh computers, and print to LocalTalk printers over the network.

The Windows operating systems supported AppleTalk starting with Windows NT and ending after Windows 2003 on the server side and Windows 2000 Professional on the client side. Miramar included AppleTalk in its PC MacLAN product which was discontinued by CA in 2007. Group Logic continues to bundle its AppleTalk protocol with its ExtremeZ-IP server software for Macintosh-Windows integration which supports Windows 2008 Server and Windows Vista as well prior versions.

In addition,

Adapted from the Wikipedia article AppleTalk, under the G. N. U. Free Documentation License. Please also see http://en.wikipedia.org/wiki

No related posts.