In this presentation, GigaSpaces' Nati Shalom and Shay Banon show how GigaSpaces' provided processing unit capabilities offer automated backup and scaling for critical applications. First, a discussion and justification of the space model is given, then Shay Banon walks through an actual deployment, with failures and guaranteed SLAs. They also define scalability, just in case you wondered what the specific meaning was. :)

Joe, don't you think it's enough now with Gagaspace? I consider it as quite penetrant and SPAM to read every 2nd or so post of that crap. Thanks, Andreas

If you have not noticed *Spaces-based systems are one of few areas where there is still real innovation going on and it is definately on the server side of things. Most of the time these sites is full of news boring product releases. That's spam to me. Products that drive innovation are exciting to read about. So if you do not like to read about new innovating products and technologies, shut up and read some other boring newssite. Thanks, Dag

Why not let the world be acquainted with such a great solution? +1 Dag -1 Andreas

Why not let the world be acquainted with such a great solution?

+1 Dag
-1 Andreas

-1 marketing blurb disguised as content -1 astroturfing

Joe,

don't you think it's enough now with Gagaspace? I consider it as quite penetrant and SPAM to read every 2nd or so post of that crap.

Thanks,
Andreas

+1 This constant bombarding of Gigaspaces brainwashing is becoming really tiresome. What innovation? JavaSpaces was originally spec'ed out in 1998. Space-based architecture is a load of marketecture blurb by a single vendor pushing a proprietary product solution. I don't hear of any other signifcant vendors or technology leaders pushing "Space-based" architecture. I wonder why ...

This constant bombarding of Gigaspaces brainwashing is becoming really tiresome.
What innovation? JavaSpaces was originally spec'ed out in 1998.
Space-based architecture is a load of marketecture blurb by a single vendor pushing a proprietary product solution. I don't hear of any other signifcant vendors or technology leaders pushing "Space-based" architecture. I wonder why ...

Right. I know Blitz and Gaga and that's it. I really don't know why they get the "Space" ;-) here... -- Andreas

This constant bombarding of Gigaspaces brainwashing is becoming really tiresome.
What innovation? JavaSpaces was originally spec'ed out in 1998.
Space-based architecture is a load of marketecture blurb by a single vendor pushing a proprietary product solution. I don't hear of any other signifcant vendors or technology leaders pushing "Space-based" architecture. I wonder why ...

Right. I know Blitz and Gaga and that's it. I really don't know why they get the "Space" ;-) here...

-- Andreas

Probably because people are FINALLY waking up and realizing that just having a webapp and a database just doesn't [always] cut it? Sure, space based architecture vendors are small, but the problem areas they solve also involve Coherence and ObjectGrid and any others like them. If you have analyitical, BI, large data and/or high transaction needs, you will be looking at these.

This constant

Constant? Search this site over the last year. Recent, more than normal posts? Sure.

bombarding

Guess you've never been bombed or bombed someone.

brainwashing

It can only be an attempt at it if you "know" you were being brainwashed. If if anyone has done brainwashing, it is DB vendors and those who can't see beyond it (google for the dba post on Map Reduce).

I don't hear of any other signifcant vendors or technology leaders pushing "Space-based" architecture. I wonder why ...

Hey, while you are at it, tell Linux vendors and Apple to quit trying to push their desktop products too.

If if anyone has done brainwashing, it is DB vendors

I would add J2EE vendors and RoR fanboys to make a list. I welcome more posts about alternative solutions for real problems. And yes it takes about 10 years for a really smart things to gain momentum - stupid things like SOAP spread like wild fire.

GigaSpaces should chill out for couple of days, let the things cool down a bit, cause community seems kind of pissed off after that BEA letter followed by trying to convince us it was not the marketing trick. Everyone knows it was marketing trick, and saying otherwise insults other people's intelligence.

I think when people come here they don't expect to only read pros about a certain solution especially if it is from a commercial vendor. I love software reviews, they are most needed, but a review usually contains a comparison to other sw products and the pros and cons the application. Gigaspace posts are more about self promotion, which is why i think many people here might find them offensive. which is think is a reasonable reaction, of which the ppl from gigaspace must be more sensible about it. Of course if they go FOSS all will be forgiven

The bottom line is, if you are going to promote your product you got to do it with style, like Cameron Purdy did.

I think when people come here they don't expect to only read pros about a certain solution especially if it is from a commercial vendor.

How long have you been coming here? It was pretty obvious what this post was linking to. If you don't like people talking about their product, ignore it.

Of course if they go FOSS all will be forgiven

Maybe by you.

.. but a review usually contains a comparison to other sw products and the pros and cons the application.

Really? From what I see, most are regurgitations of marketing materials. Thats why I seldom buy computer magazines anymore.

Hi all, I wanted to make few points clear: 1. This post was taken from a presentation that was given last year, at the TSS Symposium in LA. 2. We (GigaSpaces) had nothing to do with the decision of putting up this post or its timing. Regardless, I would encourage those who haven't done so, to read through this post or listen to the podcast before making any comment. One thing you'll find is that it has almost nothing to do with JavaSpaces and more to do with scalability issues in Tier-Based Architecture. It discuss why SOA/ESB are just a first step towards a broader solution for scalability. In this discussion we cover a set of patterns, which are based on real-life experience in achieving true linear scalability in mission-critical stateful environments, and overcome the limitations of the Tier-Based or SOA architectures. The pattern is based on the use of in-memory middleware, rather than a database or the file system, to achieve high availability and manage in-flight transactions. 2. Data bus -- suggests that it doesn't make sense to separate between the messaging tier and the data tier from a runtime perspective, and suggests a model in which the two can be integrated with each other (This is where JavaSpaces fits into the picture, as noted in other comments other technologies in the DataGrid space are covering that area as well). 3. Partitioning -- how to apply partitioning as a means to scaling data and the entire application. 4. Collocation -- we discuss how to use collocation as a means of improving latency, and simplifying deployment and scaling, by reducing the amount of moving parts. 5. Processing unit -- suggests a model in which we group together all services and middleware components that are tightly coupled from a runtime perspective (high-availability dependency, latency, performance, transaction flow). A processing unit determines the unit of scale and failover of our application. In addition, we discuss Spring-based methods for abstracting our code from the changes required to apply these patterns. Most of the discussion is provided in a generic manner that is not tied to a specific implementation. Obviously the code examples and demo are more product-specific but they are provided only as an illustration of the patterns. These same principles and patterns are used by other implementations: Amazon, Google and many others have been using these principles themselves in one way or another as i outlined in one of my previous post here. The main difference is that this is one of the first attempts to provide a holistic architectural view on the entire (end to end) application pattern, not just a specific use-case or component. It also addresses one of missing pieces in most discussions of this type -- how to apply these patterns to your existing Tier-Based application. We refer to the combination of all these patterns Space-Based Architecture (SBA). The word space in this case represents the 'cloud' more than it does JavaSpaces -- note that the name was given before Cloud Computing became popular. Probably a better name for it these days would be Cloud-Based Architecture :) It'll be refreshing to see the debate in this thread centered around these ideas and the principles behind them, rather than what one thinks should be posted or not. Nati S GigaSpaces