Link Details

Voted down for the only commercial product being Cacheonix :) I mean c'mon... Last time I downloaded it (~ about 1 month ago) it had no serious documentation, no examples, and had ~14 classes in public API. And in the same time you are omitting Coherence, GemStone and GigaSpaces, to name the few - all of which are years and years ahead in terms of the features and maturity.

We should not mix up a blunt marketing with lack of honest factual information...

Nikita Ivanov.
GridGain Systems.

Well, I hope you were kidding saying about 50 lines and complex SQLs. I would better fire one who executes SQL for every single raw in table ;)
Also I can say for sure that Oracle will collect data that you need in 5-10 ms +5 ms to process them on Java side. Don't get me wrong there are developers and Developers and the difference is exactly in those 5 seconds that you mentioned.

I personally agree with caching. Caching itself is a good solution in places where we need them. But I did not find anything new in this article.
I would probably vote up if you describe algorithm of distributed LRU + TTL based Hibryd cache made on the top of ANY database.

But anyway thanks fro the article. BTW. TimesTen + Oracle = 1-5 ms data access for example.

To Nikita:

Cacheonix is developing a usable caching and data grid API. Our goal is simplicity, not a hairy monsters like the products you mentioned. If you are looking for complexity, Cacheonix is not your product.

As for the marketing, you consistently fail to mention competing products in your speeches about compute grids, such as JPPF. Fix your double standards and we will talk.

To Dmitry:

No I am not kidding about 50 lines. It is a real-life example that benefited from caching hard to get. Unfortunatelly, long running database queries is a reality. As for timing, 100ms per query x 50 lines in the status report == 5 seconds. If you have never seen 100ms queries, you may consider doing practical software development fore claiming that something does not exist because you haven't seen it.

This is an introduction article, so it does not go into complexities. The goal of the article is to promote caching.

As for firing someone, I can refer you to HR of that company. You can discuss firing their engineers directly with them.

Slava,
Relax :) No hard feelings. Once you muster a bit more meat beyond few classes into you "project" there will be at least a reason to talk on our side. I mention JPPF less frequently and mention others more frequently - since that what our customers are mentioning. Calling Coherence/GigaSpaces/GemStone (data grids all) a hairy monsters just shows pretty stunning lack of basic understanding for this market and technology.

I can disagree w/GigaSpaces, for example, all day long on almost any topic. However, I always mention them to our *customers* because we are frequently compared to them. Keep in mind that if everything you provide is a hammer - every problem would like a nail to you...

Best of luck anyways - you'll need it.

Nikita Ivanov.
GridGain Systems.

Nikita, that's you who needs to relax. Not anything that has a grid in its name is your competitor. The goal of this article is to provide an introduction to caching, not a market overview. The next article will talk about data grids, and it will name all companies in this space we aware of.

Regards,

Slava Imeshev

Slava,
Well, I actually did spend time again looking at the project I downloaded about a month when I saw you released 2.0 version. Here is some statistics about it that will help readers not willing to go through convoluted process of downloading the software and the ones that will help them make an educated guess as for relevance to other "hair monsters" :)

1. All your examples consist of a one file with 10-lines empty JUnit stub. That's it.
2. Out of 4 scripts you are shipping - 2 are absolutely empty. Other two - are copied from someone else and I have no idea what they do.
3. You public API consists of 6 interfaces, 3 classes, and 3 exceptions (including configuration and logging) :)
4. You documentation consists of ~10 short wiki pages.
5. Current version is 2.0.
6. You charge $1950/cpu for this "product".

I'm seating on JSR-107 and have done some work on data and computational grid as you are well aware so I can speak on this topic with certain authority. To be brutally honest I find this whole situation is rather worthy of April 1st joke.

Very few people will invest time in cumbersome download and evaluation of your software. We did (with high hopes, btw). That's what we found. I think you have a very long, long road ahead and again I wish you best of luck.

Nikita Ivanov.
GridGain Systems.

Nikita,

Thank you for your comments. We are working on examples and the documentation as we speak. Cacheonix as API is simple and this is a good thing. If you think that Cacheonix API is missing something, you are welcome to comment in our support forum at http://www.cacheonix.com/forums/viewforum.php?f=2. We will be happy to add it. Cacheonix has started a local cache in early 2002 and has grown to a distributed caching system since. That is why current Cacheonix version 2.0.

Hope this helps.

Regards,

Slava Imeshev

I saw 20 min queries :) But reporting should not be cached IMHO. I just cannot agree with 100*50 lines. Get all 50 lines at once.

I'd like to point to the fact that 7 out 8 voted the article down are GridGain employees or associated with GridGain. This leaves me scratching my head, but the fact remains.

Slava,
Leave the petty feelings at home :) and start scratching you head while working on your product. You should thank GridGain with a big "T" as we not only downloaded your product and tried it, but we also not afraid to publicly comment on it without being afraid attached to it. You get honest, bullshit-free responses at least from me.

Folks from GigaSpaces, Oracle, etc. will never comment because they don't want to give you any free publicity. Cherish every response you got here (up or down) because it's pushing your post and makes people aware of your product. And if you'd look at my previous post you will see why I think many of us would vote it down - BECAUSE unlike everyone else we actually downloaded it and tried :)

Best,
Nikita Ivanov.