Wednesday, February 17, 2010

Attunity Oracle-CDC for SSIS - A Great Product

Despite a year of Azure, MVC/MVP/MVVM extravaganza, the promise of .Net 4.0 with declarative WCF/WF, CQRS confusion, and of course Oslo's purported subsumption into EF, it's clearly been awhile since I've felt absolutely compelled to write about something until now.

Anyone who has toiled in SSIS knows the many faces of frustration. Whether with the product itself, or with 3rd party drivers and addons which disappoint. But every now and then a real winner comes along giving hope and respite - Attunity is just that sort of vendor. Their 'Oracle-CDC for SSIS' product is solid and their customer support highly responsive. It would be hard to ask more of a vendor.

The product simplifies not only the process of setting up CDC for Oracle (requiring minimal Oracle DBA involvement), it also generates SSIS packages that make consuming the captured data changes to a destination database table a snap. Throw in an admin console and a straightforward mechanism for deploying to a production SSIS machine and it simply can't be beat.

At this point I couldn't imagine doing 'raw' CDC - the product is that good.

Thursday, February 26, 2009

OakLeaf Systems Objects to a Red Dog "Skunkworks"

From Roger Jennings' blog:

I don’t believe the skunk works approach is appropriate for Azure. Developers need a full description and timeline for required features. Failure to deliver promised features in a timely manner is what led to A Mid-Course Correction for SQL Data Services.

My reply: As an ex-DEC SE and someone who has also been prodding Global Foundation Services management for Cutler press access and more RD details, I would disagree with you in regard to your skunkworks comment.

First, Cutler only works one way - with a small, expert and tight-knit teams; you never get anything out of him or his team until they're ready to release. Second, if the problem they're addressing - virtualizing and automating a compututational fabric over and across data centers - required a bunch of developer level interactions then they will have essentially failed.

And last, the RD project / team is in no way associated with the typical MS app-stack confusion (especially the lingering cloud of indecision around persistence) - Cutler would have walked before subjecting his team to such vagaries. I also very much doubt they get particularly involved in the Azure / .NET Services app layer sagas and tribulations in more than a passing way.

Thursday, January 22, 2009

More on Sagas with Arnon Rotem-Gal-Oz

A bit more conversation around extended transactions and Sagas on Arnon Rotem-Gal-Oz blog.

Saturday, January 17, 2009

Sagas of the Service Bus...

A recent comment I made in response to Oren's 'Rhino Service Bus: Sagas and State' post:

"My understanding of sagas has always been from the perspective of long-lived transactions (LLT) and typically those running / orchestrated across multiple enterprises within some supply / value chain. That the [various] transactions and compensating transactions within each enterprise involved in an LLT are 'bundled' into a saga which is bound to an LLT by some mechanism.

The idea essentially being that each saga is responsible for maintaining the consistency of its [component] state thereby contributing to the consistency of the parent LLT state. But after all this discussion I can now see this saga pattern / notion exhibits somewhat of a 'fractal' characteristic relative to its application at various scales both within and across enterprises."

Wednesday, January 7, 2009

Rules Server Post Tweetstream

healyje @anupriyo @ChrisRomp The last 1/3 is around authoring of the rules / policy which should, like Oslo, expose views for both analysts and devs about 8 hours ago from web in reply to anupriyo

healyje @anupriyo @ChrisRomp Another 1/3 is similar to the infrastructure around managing /versioning / deploying services about 8 hours ago from twhirl in reply to anupriyo

anupriyo @healyje @ChrisRomp What is the rest 2/3? I'm thinking you mean rule management, health & activity monitoring, authoring tools, etc. about 8 hours ago from web in reply to healyje

healyje @anupriyo @ChrisRomp There are .NET Rules APIs / ports of engines, but the engine is only about a 1/3 of the entire enterprise picture. about 8 hours ago from twhirl in reply to anupriyo

anupriyo @healyje @ChrisRomp Java has a Rule Engine API. http://is.gd/eOyR I would like to see something similar in .NET -- for interoperability. about 9 hours ago from web in reply to healyje

healyje @martinfowler Managing those rule bases w/o a formal enterprise Rules / Policy management infrastructure is a relentless & endless nightmare about 8 hours ago from twhirl in reply to martinfowler

healyje @martinfowler The are many business domains with thousands of rules and thousands more exceptions to those rules all mutating slowly or fast about 9 hours ago from twhirl in reply to martinfowler

healyje @martinfowler Issue is of thresholds - at what point do rules engines make more sense than home spun libraries, DSLs, or smaller frameworks. about 9 hours ago from twhirl in reply to martinfowler

martinfowler RulesEngine: Should I use a Rules Engine? A rules engine is all about providing an alternative computational mo.. http://tinyurl.com/8azkl3 about 9 hours ago from twitterfeed

healyje @anupriyo You can consume rules from your code, but that is not on the scale of what comprises real enterprise rule / policy management. about 9 hours ago from twhirl in reply to anupriyo

anupriyo @healyje Doesn't BizTalk already have a separately deployable rule engine? I thought I read that somewhere but I may be mistaken. about 10 hours ago from web in reply to healyje

ChrisRomp @anupriyo The BRE is standalone, technically, but is licensed/sold w/ BizTalk Server. There's a robust API for calling it via code. about 10 hours ago from OutTwit in reply to anupriyo

healyje
Blogged: Microsoft BizTalk Rule Server 2010 - http://is.gd/eNWq about 11 hours ago from twhirl

Microsoft BizTalk Rules Server 2010

Ok, I'm back on the same 'Rules Server' rant I've been on for the past ten years or so.

I would very much like to see Microsoft to develop an MS Rules Server 2010 and an accompanying Enterprise Rules / Policy Manager - it could certainly be branded under 'BizTalk' as far as I'm concerned. However, this server should be [structurally] independent of, but used by BizTalk / SharePoint / WF / WCF and other Microsoft products as well as by .NET developers directly in code (yes, I know we can do that today after a fashion). Ideally it would utilize Oslo's DSL and Repository capabilities as well. This could either be a head-on competitor with ILOG and OSS rules software or simply be an open framework everyone could plug their implementation into (open like Oslo). Either way would be o.k. with me.

In some respects it's like rules / policy management are software's final frontier and gets little traction despite the obvious enterprise imperative in industry after industry. Insurance and Healthcare alone would make it worth doing. But every other year when I've revisited the issue with friends I know at Microsoft, Oracle, and IBM I've always heard the same refrain: "TURF, TURF, TURF" is the reason none of them can work through the space to a product. With the advent of Oslo and a major refactoring of BizTalk, it seems like there is an opportunity now for Microsoft to move beyond whatever obstacles have been in the way of the development of such a product.

How the PDC ended up costing $18k extra...

I was blissfully attending Oslo sessions at the PDC in LA when I began to feel like someone had placed a bicycle innertube around my chest at armpit level and was progressively tightening it down with a stick behind my back. This started about 11am and by 2pm I was fairly disconcerted by this strange sensation and found myself pressing on my sternum with both hands as if to try and dislodge something. By 3:30pm it was getting alarming and I now had the added lovely symptom of feeling like something the size of a ping-pong ball was stuck in my throat.

It was at this at this point I tried calling my Sister who, by happy coincidence, is an ER doc in LA; but, she didn't pick up and I decided to just persevere through the end of the day's session. I did finally get ahold of her around 5pm as we were boarding the bus back to the hotel and was told in no uncertain terms to "get off the bus and take a cab to Good Sam's ER.

I did just that and by 6pm was fully embroiled in all manner of prognostications by a flurry of doctors, nurses, and techs in Good Sam's ER. I was nitro'ed (tongue spray and chest patch), aspirined (baby), lidocained (milkshake), Nexium'ed (intravenously), Xrayed, blood tested, and subjected to other kinds of other interesting interrogations - all the while never more than three feet from a set of defribulator paddles.

At the end of my six hour ER ordeal I was transferred up to the cardio unit for observation and to do a echocardiogram stress test the next morning given it was going on midnight by then. The stress test was interesting in and of itself in that you got to watch your heart beating from all angles even if you were desperately gasping for breath for half of it. I did somehow survive it, however.

A couple of hours later my new cardiologist concluded the combination of the flu shot I'd gotten four days prior, along with 'unspecific stress' (clearly Oslo excitement), had triggered a gastric reflux event which I'd never experienced before. But, given my blood workup was clean and I had maxed out their stress test system (which he said was unusual for anyone let alone an old guy) they relented and let me get back to the PDC. So I made good my escape, but not in time to catch the Quadrant session, which was a bit of a bummer, though being alive was a bonus.

Anyway, the bill just arrived for my 20 hour stay in the hospital - where nothing more invasive was done to me than a blood test - and by the time you rolled the doctor's bills on top of the hospital it came in at $18k, or $900 per hour. I'm pretty sure the docs and nurses aren't getting much of that so I'm still wondering where it's all going. That, and I think I really need to up my rates before the next PDC.

The final word from my Sis - and the point of this post - is never wait that long if you start getting weird sensations in your chest / throat. I was lucky in that I'm in good shape and nothing bad was going on cardio-wise, but she said many folks aren't so lucky and a lot of them don't make it when they [unwisely] try to ride out their symptoms the way I did...

[ P.S. Good Sam's cardio unit was outfitted with cell phone jammers, had no WiFi in range, and anyway I'd left the USB cable for cell CDMA connections at the hotel. Damn if I didn't actually have to read - good thing I'd just nabbed one of the Oslo books... ]

[ P.P.S. Yes, I do have insurance as a self-employed person, so it's not all on me - it's on all of us. Will be having a 3rd party claims auditor go over it all now... ]