This is David Bressler's Typepad Profile.
Join Typepad and start following David Bressler's activity
Join Now!
Already a member? Sign In
David Bressler
New York
Recent Activity
Sanjay, I apologize for the delay in posting your comment, and replying. I was on vacation. I'm familiar with that whitepaper (I wrote it). Your question, is a good one, and it gets to the crux (key point) of the matter. "How can a tracer added to a message be truly non-impacting?" It is all about semantics, but let me give you the key bullets as to how we define non-impacting: 1. No application code changes are required. 2. No application configuration changes are required. 3. Messages can be fully encrypted, and Actional still works. We are not "in the middle" and therefore don't require decryption of the message to insert the tracer. 4. Does not need to be everywhere. Meaning, you can have Actional at one point in the message flow, but not others and still have full functionality (on the node where Actional is plus one-hop away) without having to do anything special on the nodes where Actional is not because of the now-present "tracer" that the un-instrumented nodes don't know anything about. For full disclosure, for the RMI protocol, the Actional tracer needs to be on both ends... so that's one case where you do need Actional in two places for it to work properly, but the same non-intrusiveness applies - it can be in those two places, but nowhere else, etc. 5. Does not affect performance. This speaks for itself! 6. The management server does not become a single point of failure; nor does it block messages if it fails. 7. The application being monitored doesn't not need to be managed any differently over time, so there isn't any increased cost-of-ownership over time. One other key Actional difference by the way, is that it can run all the time, with all functionality, so that you don't need to turn it on when you need it (which defeats the purpose of discovering what you don't know so that you can anticipate and prevent problems). Finally, these non-instrusive key differentiators mean that you can deploy Actional at any phase of a project. You don't need to architect or deploy an application any differently than you would without Actional (other than installing Actional on the base platform). I hope this makes sense and is helpful. By the way, I think these are the key competitive differentiators too - while many vendors might "talk this story" I've not seen one that can deliver it. CA Wily, IBM Tivoli, Oracle Amberpoint, OpTier... they all are hard to configure and affect performance too much to run all the time. "Must a policy be defined first before transaction tracing is enabled, or is there tracing based on automatic discovery?" Tracing the aggregate is done before policies, but if you want to trace a specific transaction you need either a policy, dimension, or business process (not an orchestrated process, Actional uses the phrase business process to mean any arbitrary repeatable flow that you want to track that you assign a name to - like provisioning process, or shipping process). You can drill into the "flow map" and see lots of details about what's happening in the collection interval, again at an aggregate, once data is collected. This information is broken down by service, operation, and if a cluster of machines is instrumented, by individual member in the cluster. A great way to see this functionality is to download the free Actional Diagnostics tool (http://web.progress.com/en/actional/actional-diagnostics.html). That tool is limited to a single node - but you can still see the traffic between components presented in a robust way showing what is happening within the node.
Julianna! Welcome to the blog. Great post!
Mark, I agree more or less in totality with what you say. But, I might have said it differently. I think press releases are good "milestone documents" in that they provide an "official party-line" on release milestones, customers acquired, key partnerships, or industry recognition. Any time, however, that you get a one-way release, especially from a private company, to me it loses all credibility without the social media validation. For example, you do a good job of representing Streambase, of keeping it real. But, without that direct contact, it seems there isn't much accountability. I've got a competitor who constantly uses customer names without permission (you can tell because it's not a joint press release, though I happen to know the inside stories), or who releases information about new features coming that never make it (press releases are much stronger with customer's participating actively in the PR). Another important point is to write press releases so that they are meaningful. I think David Meerman Scott said it best when he talks about gobbledygook press releases (http://www.webinknow.com/2009/04/top-gobbledygook-phrases-used-in-2008-and-how-to-avoid-them.html), and I find what he says to be true. Most PR teams use such watered down terms that you can't differentiate one company to the next. And that, to me, makes them of very little value. Personally, I like Cisco's old strategy... they'd put out a press release outlining a direction/strategy, and the key milestones they were going to accomplish. Then as they accomplished them, they could put out releases about how well they were executing. Anyways, I hope to see fewer press releases and more blogs from "authentic people" who have something to say. An example (get ready for some shameless self-promotion) is a recent post I put out about some features in a new Orbix release (http://davidbressler.com/2009/11/05/corba-is-now-actional-ized/), explaining authentically (I hope) why it's important. Doing this in a blog, not a press release, gave me a lot more ability to communicate clearly, and also gives readers the opportunity to ask for clarifications if they choose - something they don't have the ability to do with a press release (easily). Anyways, I'm glad to see as leader of a company you are embracing some of these ideas to do things differently. Good luck. David Bressler http://blogs.progress.com http://davidbressler.com
1 reply
This is actually quite a comical post. It's true... and it highlights, at least to me, that most of the problems that need to be solved are not hard ones. We (problem solvers) just need a systematic approach to identifying problems, so that they can be solved. Most of the time, buyers just walk away. I will walk away from a store where there is a line of smokers (often employees) standing in front making me breathe their exhaust in order to enter the store. One of my personal favorites, along these lines, is "extra clicks" on many sites. They say "click here to login" -- well, why couldn't they just put that little uname/password entry widget/form into that place, instead of making me click to get it?
1 reply