How to implement a Comprehensive Search Solution in SharePoint?

Question

I am looking to implement a search solution for one of my clients. Any comments will be highly appreciated. We currently have SharePoint 2007 Enterprise. We have the following requirements:

1. Users can search for SharePoint Items (Documents, Tasks, and Events etc.)
2. Users can search for SalesForce Items (Chatter Discussions, Documents etc.)
3. Users can search for Exchange Mailboxes Items (Emails, attachments etc.)
4. Users can search for Network Shared Folder items
5. All search results, which have the synonyms/related terms (defined by the business some where), will also be displayed. for example, if user searches for "TFS" then search results containing both "TFS" and "Team Foundation Server" should be displayed.
6. Search results from all different sources must be combined and presented on a single page but should have indicator of some sort (e.g. icon) to highlight their source of origin. For example, SharePoint result can have a different icon than a search result from SalesForce. This will help user to quickly glance over the results and get the idea of which result come from where.
7. Users must not be forced to learn different ways of searching for different sources; they just use normal search techniques as they are used to of doing in normal SharePoint search.
8. Search Results should respect the per item security from all sources. For example, if someone is not allowed to access "Z:\HR Documents", he should not see results from there.
9. Searching must be very fast or reasonably fast and should not heavily impact the content sources (Exchange Mailboxes, Shared Folder,Sales Force etc.)
10. Search infrastructure must be extendable so that we can add more content sources afterward, if required.

Basically, we want to provide a single point search experience to make sure one should not miss any information, which may be sitting somewhere in an organization, but due to inability to search for it, nobody knows about it.

Answer 1

1.Users can search for SharePoint Items (Documents, Tasks, and Events etc.)

Thats a standard functionality.

2.Users can search for SalesForce Items (Chatter Discussions, Documents etc.)

You could go with a federated search that would display results in the same way as e.g. Google presents its ads on the right side of the result page. A second way would be to integrate the data via DCS as external lists. I think AvePoint has even a SalesForce-SharePoint integration tool -but I have never used it.

3.Users can search for Exchange Mailboxes Items (Emails, attachments etc.)

You can setup a content source for your Exchange server. However, I think that only public folders are searchable. Personal mailboxes are not.

4.Users can search for Network Shared Folder items

Yes, thats no problem either. Just set up a content source for that share and give proper permissions to the crawl-account.

5.All search results, which have the synonyms/related terms (defined by the business some where), will also be displayed. for example, if user searches for "TFS" then search results containing both "TFS" and "Team Foundation Server" should be displayed.

Thats a functionality that the managed metadata service in SharePoint 2010 gives you. But it will only be available to SharePoint content.

6.Search results from all different sources must be combined and presented on a single page but should have indicator of some sort (e.g. icon) to highlight their source of origin. For example, SharePoint result can have a different icon than a search result from SalesForce. This will help user to quickly glance over the results and get the idea of which result come from where.

That works out of the box. Instead of an icon the path to the document is shown. If you want you can customize the XML/XLST of the search result webaprt.

7.Users must not be forced to learn different ways of searching for different sources; they just use normal search techniques as they are used to of doing in normal SharePoint search.

Yes, all your content is searchable from a single SharePoint search center.

8.Search Results should respect the per item security from all sources. For example, if someone is not allowed to access "Z:\HR Documents", he should not see results from there.

SharePoint search are security trimmed. However, that topic is pretty complex and I suggest you start reading following articles: http://msdn.microsoft.com/en-us/library/aa981236(v=office.12).aspx http://msdn.microsoft.com/en-us/library/aa981314(v=office.12).aspx

9.Searching must be very fast or reasonably fast and should not heavily impact the content sources (Exchange Mailboxes, Shared Folder,Sales Force etc.)

Searching, which means querying I guess, is very fast. A well performing SharePoint search should give you results in less than one second. However, crawling and indexing may take a lot of time. Depending on the amount of content it can take hours or even several days.
You can imfluence the impact on content sources with "crawler impact rules" (http://technet.microsoft.com/en-us/library/cc262861(office.12).aspx)

10.Search infrastructure must be extendable so that we can add more content sources afterward, if required

You can always add new content sources, that not the point when talking about extensibility. Whats important is that MOSS 2007 can only have one index server. That single point of failure is "fixed" in SharePoint 2010.