Technology Musings - Latest Comments

Re: Android – Rise of the Amazon Marketplace, Part 2

Bruce Bent II — Fri, 04 May 2012 03:15:36 -0000

Technology plays a major role in a business. Fast exposure to prospect clients are gained and it benefits almost 90% of businessmen. World without great technologies, well, I can't imagine what would it looked like. All I can say is that, it really had helped a lot of entrepreneurs.

Re: Consideration For The Technical Implementation of an SOA

Franck MIKULECZ — Sat, 04 Dec 2010 19:57:36 -0000

Surprised not more comments came up, I find your research fascinating. It is based on simple facts and produces powerful decision fodder for people in your use case situation. I am somehow one of them and will be discussing this with our Dev. teams for immediate review...

Thanks for the good work
(oh shit, we are still heavy on Java)

Re: How To Build an SOA Based, High Performance, Scalable and Reliable Twitter on Steroids

rohanpillai — Sun, 02 May 2010 14:03:11 -0000

Excellent work & design. I also work on Datapower and appreciate the wonders it does!

Re: Here is my hammer. Show me your screw!

eight — Sun, 03 Jan 2010 16:31:54 -0000

screws, lol

Re: The New Economics of Technology Startups?

eight — Sun, 03 Jan 2010 16:30:55 -0000

My company's reliance on advertising is overwhelming, they pay $1000/week

Re: The New Economics of Technology Startups?

Sam495 — Tue, 22 Sep 2009 18:34:08 -0000

we need more free products supported by text ads in the bottom corners, beats premium any day

Re: Here is my hammer. Show me your screw!

Robert Morgan — Sun, 20 Sep 2009 11:17:20 -0000

nail gun!

Re: Here is my hammer. Show me your screw!

nnj — Fri, 18 Sep 2009 05:42:38 -0000

should it not be i have a hammer, give me a wall and a nail :)

Re: Consideration For The Technical Implementation of an SOA

Robert Morgan — Sun, 13 Sep 2009 16:48:47 -0000

holy crap my head hurts

Re: Enterprise 2.0 Needs To Stop Being So Naive

Robert Morgan — Sun, 13 Sep 2009 16:48:19 -0000

bring in the probe! lol

Re: The Need For Speed

techmusings — Thu, 10 Sep 2009 08:20:24 -0000

I don't think so. XBRL is currently only being used for SEC filings, etc. Its not something that is catching on more broadly in financial markets at this time. Generally we don't use a lot of XML because of the high cost of parsing it. Even with the filings, I think you will see some of the currently data vendors take the XBRL from the SEC, parse it once and re transmit it using legacy formats.

Re: The Need For Speed

nnj — Wed, 09 Sep 2009 09:53:07 -0000

With trades happening so fast, exchanges moving toward dark pools will XBRL be the technology platform of choice for market sensitive news dissemination?

Re: Enterprise 2.0 Needs To Stop Being So Naive

Bob Gourley — Mon, 07 Sep 2009 07:49:01 -0000

Thanks very much for this big dose of sense. I appreciate it, it fits totally with my observations of the challenges of enterprise IT.

Re: Why a Business Process Modeling (BPM) Approach to SOA Usually Fails

techmusings — Sat, 05 Sep 2009 18:19:32 -0000

I agree with you re the BPM. The intent here is not to say that you should not do a BPM exercise as part of the SOA design process. You definitely need to do that. The problem is that most people stop there and end up making their design too process specific. You have to take the design process a bit beuond the BPM and look at foundational services and more importantly in my experience you need to identify the fundamental data subjects the architecture needs to deal with as the associated fundamental data objects (or types).

As for the performance issue, the answer is it depends. Let me cross post the comment I made to a similar question on the original article on technologymusings.com.

The answer is it depends. Keep in mind that this is a logical architecture at this point. While we may choose to deploy it in a physical realization that matches this with each component on a separate server and every call being through a network hop, etc, it it not necessary. Also, we have not indicated that each service interface needs to be web services (which is slower, but not slow by most people standards (sub 1ms per call typically unless the interface takes big XML data objects in the call)). That said, you can do this message based and have very low latency per hop/call (as low as 3 microseconds with the latest high speed messaging). Either way, a single call to a traditional on disk database is going to swamp the Web Services and the layered network calls. If you want to be low latency the whole design end to end needs to account for that goal. You may put critical service components on a single box to minimize network hops. You may use a binary service interface for those high performance calls, etc. There are many ways to skin this. The key is to remember that SOA does not require statelessnes, SOAP or REST. SOA applies independent of the technologies we choose to use to implement it.

But your right if you have too many layers and you don't choose your technology implementation well, you could slow it down. But it would take a lot to be noticeable vs the database calls or the internet hop costs themselves.

Hope that helps

Re: Why a Business Process Modeling (BPM) Approach to SOA Usually Fails

venkumar99 — Fri, 04 Sep 2009 12:42:14 -0000

Paul,
I don't think this is an issue of BPM approach having failed the SOA implementation. All BPM is supposed to identify a process. The actual modelling of as a contact or a customer or a user is part of your SOA and how you define them. That is more an abstraction ant really part of a business process.

I agree that the design you came up with makes more sense. Lot is also going to depend on how the company goes about its BPM/SOA journey. If one dept of group implements SOA without understanding the business context across other functional areas, you will end up a architecture similar to what you mentioned.

Re: Enterprise 2.0 Needs To Stop Being So Naive

Mark Fidelman — Tue, 01 Sep 2009 17:06:17 -0000

Agree Paul, we need more definition around E2.0. I am developing a blog post that I hope sheds some more light on the subject (although it's only my own view). I'd be interested in your feedback.

Re: Enterprise 2.0 Needs To Stop Being So Naive

techmusings — Tue, 01 Sep 2009 09:22:02 -0000

I agree. Large enterprises are definitely not early adopters as a rule. I do think the thoughts and technologies around Enterprise 2.0 will be embraced earlier by the SMB sector and startups in general. Even then, the definition and focus of Enterprise 2.0 is still too narrow in my opinion. It can't just be about Social Networking tech and cloud, etc. its needs to be more encompassing if we are going to apply such a grandiose label to it.

Re: Enterprise 2.0 Needs To Stop Being So Naive

Mark Fidelman — Tue, 01 Sep 2009 08:31:50 -0000

Good post Paul, yet I think your focus is on the late majority and not the early adopters. I too have come across similar resistance and issues that you have faced at the CXO level and specifically the CIO (think old manufacturing).

The fact remains, some will embrace the E2.0 movement sooner than others. But, you need to start somewhere.

Re: Enterprise 2.0 Needs To Stop Being So Naive

Brandon Bohling — Tue, 01 Sep 2009 00:36:38 -0000

Ahhhh, but it's all about hype these days...not just for the the traditional media, but for the blogosphere as well. And then after all the buzz, people can start making money via conferences, consulting, workshops, etc. Where's the fun in working with dead, er, legacy systems!? Who wants to live in the REAL world? Again, spot on Paul.