Hi Amber,
I recently gave an answer to similar question. What is important are requirements (functionality) for your BizTalk project. Scoping out the requirements for a BizTalk project is much the same as any other development project. It takes experience, judgement and know-how to be able to estimate how many man-hours (effort) will be required for a given set of BizTalk tasks and I don't know of any automated tools which will help (other than BizUnit, Team-Foundation). Approach for integration projects can be by using waterfall, (semi) agile and scrum methodologies. An important factor is discipline within development on way the work on project, their skill and how they work together. Those are important for having a controlled lead-time, and budget. In the end it really depends upon what the project is trying to accomplish and how requirements are going to be met (satisfying the business). Since there are many factors and dependencies involved estimation is hard. Having a project team that has done many BizTalk projects, has discipline and is skilled does make estimation easier.
HTH
Regards,
Steef-Jan Wiggers
MVP & MCTS BizTalk Server
http://soa-thoughts.blogspot.com/
Ambar, like you said : definitely NOT only configuration efforts and minimal coding!
It all depends on the team, whether they're confortable with Biztalk or not (and have they done integration projects - really different from doing a web site).
What we're doing is defining patterns (publisher, subscriber, service, etc.) and for each one the tasks necessary. Then, learn from our mistakes and experience. Our estimates much better now (but also a lot higher than what we initially thought ;p ).
Regards,
Hi Ambar,
definitely a mathematical equation is possible. Just base lining it.
effort =
{
(number of orchestrations * average size and complexity of orchestration * orchestration_complexity_weightage
+
number of schemas * average size of schema * schema_complexity_weightage
+
number of business rules * average size of business rule * business rule_complexity_weightage
+
number of maps * average size of map * map_complexity_weightage
+
number of custom pipeline components * average size of custom pipeline components * custom_pipeline_components _complexity_weightage
+
number of custom adapters * average size of custom adapter * custom_adapter_complexity_weightage
+
..
) *
(number of dependencies between artifacts (orchestrations, schemas, rules, ..) *
(number of interfacing applications_factor) *
(non functional requirements_weightage) *
...
}
/
{
// effort reducing factors are below
average_biztalk_experienceandskill * average_domain_knowledge * architect (s)_skill_rating *
client_requirementsClarity_rating * clientside_architect_cooperation :) (if he feels that biztalk is all about settings and compositions, it can demoralize)
...
}
definitely a mathematical equation is possible. somebody has to take care of developing it.
WBR,
prasadpnvrk
Hi,
My answer would be a combination of both Steef-Jan Wiggers and Pnvrkprasad. I have created Excel Template which codify Pnvrkprasad post (mostly), but that only provides a ballpark estimation. Once you are in the ballpark the experience and judgement of the person doing the estimate and their knowledge of the team who is actually doing the work comes into play.
William
Dear William,
Could you please send that excel template to ambar.ray@tcs.com It would be of great help. Thanks in advance
Where the 'architect' has failed is in his complete lack of understanding of BizTalk (not a good sign for somebody who is an architect).
BizTalk is a development toolset and application hosting environment. To say no development work is required is nothing more than ignorance. BizTalk offers a framework to develop applications for enterprise messaging and services. Lead time on a lot of development can be reduced by using BizTalk as it provides rich tools capable out-of-the-box.
As for estimation it's a dark art that differs from person to person. Simply put all developers are not equal. All a developer can rely on is his experience and gut feel for project work. I often find things that on the surface should take a day to produce land up taking a week due to some technical obscurity which is proving impossible to fix (rebooting your computer on a Friday often solves it).
Hi William,
Could you please upload your excel template to here? This will be very helpful.
Thanks and Regards,
Netra
I agree. Can you please share the xl format so it it useful for the community
Hi William,
Could you please forward the Biztalk Estimation Template to manohar.reddy@gmx.com
Hey Prasad,
Requst if you can please send that excel template to shailendra.mohan@wipro.com It would be of great help. Thanks in a
Hi,
Anyone got excel template?? Please upload here.
Thanks and Regards,
Netra
I don't think anyones "template" is going to help you. This is just something you learn from experience; and as has been mentioned the experience of your team matters (there are some of us on this forum that could work 10x - honestly 10x - as fast as everyone else because we know BizTalk so well). The best answer is to do a project and look back on it, then build a set of metrics that back into the project timeline. I don't really think you can count on a precise level because some things you just don't know (like if you really need an orchestration) until you execute them. I would focus on the use cases / functionality as a start. Also the existing knowledge that you (or your client) have around the end systems that are being integrated are important. If you don't have anyone that can help you untangle an end system and you have to dip into SQL or Oracle to do everything it's going to be a hard slog.
-Dan
Hi William,
Would you please email me the template as well? m.kharati@gmail.com
Cheers,