Computing Fabrics (1998-2003)

On May 19, 2003: Eric Lundquist, Editor-in-Chief of eWeek, recognized that IBM's On-Demand Computing, HP's Adaptive Enterprise, and Sun's N1 are all movements towards Computing Fabrics as we first predicted them in 1998.

On January 7, 2002: eWeek called our 1998 Computing Fabrics Cover Story "Prescient"
and declared The Grid, a subset of Computing Fabrics, "The Next Big Thing".

Riding the
Third Wave

In the News 2002-2004

Computing's Next Wave 1998
(The First Report)

The Next Big Thing 2002
Computing Fabrics & Grids

The Three Waves of Computing


Defined & Compared


Conferences & Workshops

The Bigger Picture


The Next Big Thing 2002

Computing Fabrics and Grids

The Grid and Computing Fabrics
All in all, a Computing Fabric can be thought of as Grid architecture applied across a variety of levels of scale (from a network on a chip to global networks) and offering greater configurability.

Global Network xxx xxx    
Enterprise Network LAN xxx xxx xxx  
System xxx   xxx xxx
Subsystem xxx     xxx


Functional Unit xxx      

Grid Gotchas
Erick Von Schweber's inline comments on Peter Coffee's excellent analysis of The Paradox of Grid Computing.

Software Services Grid Workshop
A collection of groups including the Object Management Group, World Wide Web Consortium, and the Global Grid Forum recently held the "Software Services Grid Workshop" in Boston. This event was organized by Erick Von Schweber and Bob Marcus.

The Global Grid Forum 2001
"Towards a Model Driven Semantic Web"
Erick's presentation (in his position as CTO of Cacheon Inc.) to the Global Grid Forum summarizing and extending the findings of the Software Services Grid Workshop.

Related Stories
Get the Most Bang for Your Budget This Year
By Eric Lundquist (eWeek)
January 7, 2002
Open Source Is the Loom for Fabrics
By Rob Fixmer (eWeek)
January 7, 2002

The Grid and Computing Fabrics
By Infomaniacs
January15, 2002

The Paradox of Grid Computing
By Peter Coffee (eWeek)
January 7, 2002

Grid Gotchas
By Infomaniacs
January 15, 2002

Girding for Grids
By Rob Fixmer (eWeek)
January 7, 2002
Plugging into the Global Grid
By Tom Sullivan and Ed Scannell
August 3, 2001

in eWeek

On January 7, 2002 Rob Fixmer:

  • Described the Infomaniacs' 1998 Computing Fabrics article as "prescient"
  • Declared Grid Computing as The Next Big Thing
  • Decided"Fabric? Grid? The only difference is the density of the weave

Distinguishing between The Grid and Computing Fabrics

(i) a Computing Fabric denotes an architectural pattern applicable at multiple levels of scale, from global networks to networks on a chip (which are beginning to appear) whereas the Grid is an architecture specifically applied to campus through global networks (a consequence is that a Computing Fabric may be denser than a Grid as Rob noted),
(ii) a Computing Fabric admits to irregularity, where some regions of connected nodes may be very closely coupled and present a strong single system image while other regions may be very loosely coupled letting distribution concerns show through (the Grid by comparison is uniform), and finally
(iii) a Computing Fabric is dynamically reconfigurable and fluid whereas a Grid may be rigid or require periods of downtime for repair, upgrade, or reconfiguration.
All in all, a Computing Fabric can be thought of as Grid architecture applied across a variety of levels of scale and offering greater configurability.
By Erick Von Schweber
(C) Copyright Infomaniacs 2002. All Rights Reserved.
Updated January 15, 2002

Links between The Grid and Computing Fabrics

It's interesting that, not a whole long time after the 1998 cover story, SGI found their collaboration with Microsoft to be stalling, and began seeding Linux with several of their proprietary "fabric-like" technologies. Globus, Legion, and related open source efforts have certainly moved ahead with a speed that makes commercial efforts seem tortoise-like by comparison.

The links between the Grid and Computing Fabrics go back to the beginning - the Grid Forum (now Global Grid Forum) began with a Birds of a Feather session at SC98 immediately preceding our Computing Fabrics BOF in the very same room.

-Erick Von Schweber
In response to
Open Source Is the Loom for Fabrics

By Rob Fixmer.

January 7, 2002

in eWeek

The Paradox of Grid Computing
By Peter Coffee
January 7, 2002

Below we've quoted Peter's excellent analysis of the Grid Paradox in whole and embedded Erick Von Schweber's remarks based on several of Peter's points.

Infomaniacs Response: Grid Gotchas

"Experienced aviators warn novice pilots, "the problem with a multiengine airplane is that sometimes you need them all." During takeoff, for example, failure of even a single engine is a high-risk situation—but with more engines, there is a greater chance of at least one failure.

"Distributed computing systems, such as grid computing, involve a similar paradox: The more resources the system has, the greater the number of points where the system can fail or degrade—and the harder the task of ensuring adequate performance in all situations, without unacceptable overhead.

"A computing grid faces four new tasks, in addition to whatever problems it was built to solve. The grid must discover and allocate available resources as their availability comes and goes; it must protect long-distance interactions against intrusion, interception and disruption; it must monitor network status; and it must initiate and manage communications among the processing nodes to make each one's needs known to the others. There is no single optimal approach to any of these tasks but rather a family of possible solutions that match up with different types of problems.

"Delays in communication between widely separated nodes fall into two groups. Fundamental is the speed-of-light limit: A node at one location cannot possibly become aware of a change in state at another location in less than the straight-line, speed-of-light propagation time of almost exactly 1 nanosecond per foot of separation.

"That sounds good until it's compared, for example, with modern local memory access times of, at most, a few tens of nanoseconds. Tightly coupled applications, such as simulation or process control, are therefore disadvantaged in distributed environments "until science discovers a method of communication that is not limited by the speed of light," as Aerospace Corp. scientists Craig Lee and James Stepanek wrote in their paper published in April 2001 (which can be accessed via"

Erick> I had a dinner conversation with Greg Chesson a couple of years ago where we extrapolated one of the then recent results in experimental physics where a quantum superposition of states was maintained in the laboratory over a distance of 100 meters If one could harness and refine the techniques, we argued, and apply them to interprocessor communication then it would be possible - theoretically - to achieve instantaneous cache coherency across a very large multiprocessor system indeed (the heat dissipation problem would also rear its ugly head if one were to concentrate so many heat sources).

"There are problem decomposition techniques that aren't as badly handicapped by the speed of light: for example, Monte Carlo simulation, or the kind of data parceling strategies made famous by the SETI@Home project, which distributes sets of radio telescope data for intelligent-life detection by screen saver software. When problems lend themselves to this approach, they often don't need frequent synchronization and therefore aren't severely hampered by distance."

Erick> Another way of dealing with latency is to apply the techniques used in advanced ccNUMA architectures, specifically by migrating processes and replicating cache lines so as to achieve greater locality of reference. This notion, dynamically creating non-uniformity to better support the application work load, is one of the characteristics that distinguishes a Computing Fabric from a computational grid.

"What does affect the latter class of problem, though, is the limited bandwidth of networks and network interfaces. Plotting recent progress, Lee and Stepanek in the paper cited earlier find network access bandwidth, as determined by available interface cards, doubling every 2.5 years, ominously lagging the 1.5-year doubling time of processor performance, assuming continued Moore's Law improvement-which many project as likely through 2010."

Erick> At a distributed computing conference in San Fran last year Craig and I spoke (along with Dennis Gannon) and with photonics this disparity may be reversing, creating pipes that contemporary architectures may be unable to saturate except with "very" big block transfers.

"With processor speed outpacing the ability of interface cards to send and receive to the grid, it follows that some processing power will be best employed in boosting information content per bit: for example, by continuing the refinement of data compression algorithms using techniques such as the wavelet transforms in the JPEG 2000 standard."

Erick> At the same distributed computing conference there were several related presentations that applied Rich Wolsky's Network Weather Service to the task of guiding the selection compression and streaming protocols for distributing a Java class.

"Data compression developments such as these are offset, however-perhaps to devastating effect-by the growth of data overhead entailed in the use of XML syntax to make data more self-disclosing than it is in application-specific binary data structures. There's a difficult tradeoff to be made between ad hoc availability of data for unanticipated uses and efficient, cost-effective packaging of data."

Erick> Now that's a sore point you've touched on. Back at SC2000 in Dallas a paper was presented comparing SOAP with Java RMI (no speed demon itself). The result: SOAP required an order of magnitude more bandwidth than RMI and expanded the delivery latency by an order of magnitude. They suggested a form of protocol negotiation, beginning with SOAP (for greater likelihood of compatibility with a target remote system) and then using SOAP to negotiate advancing to a more efficient protocol to conduct the real work. Guided by Network Weather Service this could be a real boon.

"Sad to say, a great deal of processing power may also be consumed by the calculations needed to implement data integrity and security measures, such as encryption for authentication of messages sent and received. Grid computing, in an open environment such as an IP network, invites both attempts to read the mail between the nodes and to analyze the patterns of traffic for what they might reveal about concentrations of valuable information.

"If network and computer are the same, it follows that the network-an inherently exposed asset-is increasingly the locus of IT value. Enterprise IT architects and service providers will have to learn to protect it without crippling its hoped-for performance gains."

Erick> What we need, in my estimation, is security that can be turned on and off at a variety of levels of scale in the application, according to where it is (security context) at any particular time and what it's being asked to do. And this must be done as a fully separated concern, not mingled with other aspects of an application.

By Erick Von Schweber
(C) Copyright Infomaniacs 2002. All Rights Reserved.
Updated January 15, 2002



By Linda Von Schweber
& Erick Von Schweber

Copyright 1996-2004 by Infomaniacs. All Rights Reserved.
Updated May 28, 2003