SharePoint Blogs / SharePoint University
SharePoint Blogs and SharePoint University - all in one place!
Need SharePoint Training? Attend a SharePoint Bootcamp!

Please delete cookies related to sharepointblogs.com and sharepointu.com to resolve login issues!

Site Collection Logical Architecture

This post is probably not relevant to most people, but I keep running into the same discussions with companies that are setting up SharePoint for the first time or have already set it up and are asking a lot of questions about why their environment is not as flexible as they thought it was going to be. I also seem to get into discussions about this with other SharePoint integrators who don't feel that this topic is important. Maybe they assume people already know the concept or maybe they don't understand it themselves.

I feel that understanding the architecture around the definition and maybe the purpose of sites and site collections is the building blocks of SharePoint. Most people coming into SharePoint for the first time see it as if it were a single web application with a ton of features and settings. They also look at it as if the taxonomy of the sites were a physical heirarchy on a file system. Its hard for some people to grasp the understanding that everything is stored in a database and there is no real physical site hierarchy.

SharePoint Architecture 

With the use of managed paths, site collections can impersonate a physical hierarchy for a means of organization, but the site collections are still on the same level as far as SharePoint is concerned. A Farm, Web Application, and a Site Collection are all "Containers" (or in active directory thinking "organizational units").

As far as when to use a site collection as opposed to using a sub-site, that is based on the organization itself. Usually it is the IT department’s goal to implement the SharePoint system. In most circumstances these people are infrastructure experienced and not experienced in deploying applications. This results in a hodge-podge installation of SharePoint with a single site collection to house everything known within the organization. This will result in a failed installation, since there is no room for scaling. Plus IT will either be burdened with the task of maintaining all the structure and security, or they will demand it (sometimes IT doesn't get the idea of distributed security).

Site collections will allow the IT department freedom to maintain just application itself without the worry of security or content hierarchy maintenance. The following is a list of what an individual site collection offers.

For the Users:
  • Dedicated Recycle bins
  • Dedicated usage Reports
  • Distributed administration (site collection administrators)
  • Dedicated search scopes, keywords, and best-bets
  • Custom feature deployments
  • Dedicated language translation maintenance
  • Dedicated galleries for web parts, master pages, content types, site columns, site templates, and list templates
  • Dedicated shared libraries, such as site collection images and site collection styles
  • Dedicated real estate (Self Containment)
For the IT Administrators:
  • Site quota templates
  • Distributed administration
  • Site locking
  • Database maintenance options
  • Backup / Restore abilities
  • Content Deployments
  • InfoPath forms services global template targeting

I have 2 big, at least what I think is big, points on why to use site collections. The first one is site quotas and recycle bins. The issue is the recycle bin is based on site collections and the quota for a site collection. If everyone shares a site collection, then they share the recycle bins storage size. The example I usually give is HR deletes 1MB per day and IT deletes 1GB per day. With a 5GB site quota, HR content will be flushed through the system a lot quicker if they share a recycle bin. This results in having to restore a database to get back a single document. (You know it will happen, Murphy’s Law). If they were in separate site collections, then the HR recycle bin would be valid for months maybe years with a 5GB quota because it is only affected by their deletions.

The second point is distributed administration. For most small companies this might be a moot point, but for high content driven organizations with a small IT force. It is a godsend for IT. IT doesn't know who should be able to see what content, besides how it should be organized. This is the job of the content owners and users. SharePoint site collections offers IT the ability to create a site collection for a project, team, department, document, or whatever the needs are, then assign an owner and hand it off to them. Now IT has time to work on IT related issues and the content owner now has their real estate to start developing. By this I mean, they now have the ability to quickly create libraries, calendars, meetings, wikis, whatever they feel they need to efficiently organize their content in a meaningful manner to the people who will be using it.... THEM. Plus the ability to allow users to either read or contribute to the site. This is huge since the content owners best know who needs to author, read, consume the content, plus how it needs to be organized. IT usually has no clue which then leads IT down a long time consuming road to interpret how the content owner needs it and will probably have to keep changing it constantly since they are the owner. IT has their own work to do as well.


Posted 06-25-2007 8:50 AM by dwollerman

Comments

Andrew Woodward wrote re: Site Collection Logical Architecture
on 06-25-2007 9:39 AM

Dave,  excellent posting on sites and site collections.

Even Microsoft fail at times to understand the use of site collections (as is the case with the Microsoft Learning Gateway).  

Toby Getsch wrote re: Site Collection Logical Architecture
on 06-25-2007 5:40 PM

Excellent descriptions w/ great bulleted lists for users and IT Administrators.  This further clarifies strategy choices I've already made, and gives more credibility to my vision of where and why we're moving towards empowering the people who are closest to the data, but still maintaining control.  Thanks.

Greg wrote re: Site Collection Logical Architecture
on 06-25-2007 10:06 PM

Very nice. One question though. I understand that each Personal site is it's own top level site collection. Your diagram above makes it look like all personal sites are stored in a single site collection. Can you clarify? Thanks.

Sharepoint link love 6-26-2007 at Virtual Generations wrote Sharepoint link love 6-26-2007 at Virtual Generations
on 06-26-2007 5:24 AM

Pingback from  Sharepoint link love 6-26-2007 at  Virtual Generations

Dave W. wrote re: Site Collection Logical Architecture
on 06-26-2007 7:08 AM

I agree, I didn't clarify that good enough in the diagram, but yes each personal site is in its own site collection. What I was representing in the diagram is that all the personal site collections are stored in their own web application outside of the SSP but still within the Farm. I have updated the diagram to try and represent it a little better. Thank you for the feedback.

Links (6/26/2007) « Steve’s SharePoint Stuff wrote Links (6/26/2007) « Steve’s SharePoint Stuff
on 06-26-2007 9:02 PM

Pingback from  Links (6/26/2007) « Steve’s SharePoint Stuff

Mike Walsh's WSS and more wrote WSS FAQ additions and corrections LXI - 25th June - 1st July 2007
on 07-01-2007 1:37 AM
Matt B wrote re: Site Collection Logical Architecture
on 07-03-2007 5:29 AM

Great Discription, Thanks.  It's actually quite logical once it's explained!

Mirrored Blogs wrote Site Collections Logical Architecture
on 07-27-2007 5:02 PM

Dave Wollerman has posted a great article on the use of site collections and the benefits they have for

Angie R. wrote re: Site Collection Logical Architecture
on 09-21-2007 3:00 PM

Can content be shared across site collections? That is, what if we need unique site collections for each of our business units (for all the benefits you described above) - but 80% of the content ends up being the same across all sites? We won't have to build each page from scratch for each unique site collection, will we?

dwollerman wrote re: Site Collection Logical Architecture
on 09-21-2007 6:44 PM

think of site collections as individual containers. Each one is separate on to itself. The content in each site collection is separate from other site collections. Templates and site definitions can be used to create a default workspace with defined lists and libraries along with workflows and features that are activated.

Angie R. wrote re: Site Collection Logical Architecture
on 09-21-2007 9:24 PM

So even if the content is exactly the same, I would have to upload it multiple times, to each unique site collection? Can Sharepoint be extended/customized to overcome this? Or is it a fundamental limitation in the product, no matter how it's configured? Really appreciate your insight.

dwollerman wrote re: Site Collection Logical Architecture
on 09-22-2007 8:08 AM

site collections are designed to store like and similar content. If each site collection is storing the same content then maybe site collections are not the way to go.

I am not sure what the 80% of content is, but if it is content that should be unique or requires audits and such, then having multiple locations ofr the same content would (or should) fail the audit.

Not knowing your content, you can possibly use content deployment functions as part of MOSS using central administration (or WSS you have to use the STSADM command line utility) to push the content to all the site collections.

Angie R. wrote re: Site Collection Logical Architecture
on 09-24-2007 8:21 AM

Thanks for your perspective. To give you an example of  "common" content, let's say we're running a number of food-related sites. We would have articles on wine-food pairings, that could apply across all of the sites. So, rather than upload the article on wine-food pairing separately to each site collection, I'd like for only one instance of the content to live in some common repository, and simply "assign" it to the various sites.

dwollerman wrote re: Site Collection Logical Architecture
on 09-24-2007 8:38 AM

Angie,

There are a few ways you can do this, but not sure if any of the ways are what you are thinking...

#1: Content Deployment

Deploy content to multiple sites. This requires paths and jobs setup though central administration or manually using the command line STSADM utility. Still duplicates will then exist on every site.

#2: Page Viewer Web Part

Create a site that contains the common content then have pages created with page viewer web parts that render the content from the common site. This way there is only a single source for editing / maintaining.

#3: Site Definition or Features

Create a site definition template or a feature that contains the content as ghosted. This way the content exisits on the file system with the site def or feature and sites can then use that content. This still gives the single source benefit, plus allows for each site to customize if needed.

hopefully this helps.

Angie R. wrote re: Site Collection Logical Architecture
on 09-26-2007 5:24 PM

Thanks for all the info/feedback! You've been a great resource as I try to reconcile the options.

Robert Aston wrote re: Site Collection Logical Architecture
on 10-19-2007 2:05 AM

Hi David

Great article, sheds so light on some of the questions I have been having about the architecture of site collections.

I have a question though similar to Angies. Say you were to create a site collection for HR and a site collection IT. If the HR site collection had a list of custom training codes and you wanted to use that list as a look up field in a list in the IT site collection how could you achieve this? The two lists are not visible to each other right?

dwollerman wrote re: Site Collection Logical Architecture
on 10-19-2007 6:17 PM

Nope, site collections are not able to see other lists in other site collections. You would have to develop a custom field type to accomplish this.

Alessandro Galanti wrote re: Site Collection Logical Architecture
on 10-22-2007 10:23 AM

One question about that.

Can a Site Collection use multiple content DB?

thank you

dwollerman wrote re: Site Collection Logical Architecture
on 10-22-2007 10:32 AM

No. Part of the hierarchy of site collections are that an entire site collection is contained in a single content database. You can have multiple site collections in a single content database but you cannot have multiple content databases for a single site collection.

SteveSelf wrote re: Site Collection Logical Architecture
on 10-24-2007 3:22 PM

I relaly enjoy your perpective on the structure...

here is my question

lets say that i whant to use site collections and i have a  site www.abc.com/a/b/c  

i whant to create a site collection for Site C and have no content in site a or site b.... however i whant the ability to create a site collection at a laterday for Site b.. and hopefully not effect any of the data in A B or C  

Is there a way to do this... I have kinda experimited with this and it has broken... so any fixes would be welcome also

thanks

dwollerman wrote re: Site Collection Logical Architecture
on 10-24-2007 7:42 PM

I do know that SharePoint will choke if you try to create a site or already have a site created where a managed path will be used. I have ran into that in the past and it was a pain to get it resolved (at least in 2003 it was).

I haven't tested this out, but what I was thinking you would have to use is explicit managed paths. You would need one for a/b/c and a/b. I am leaning towards this not working. I honestly don't think you can have site collections that conflict with managed paths.

Jacob Egholm wrote re: Site Collection Logical Architecture
on 10-31-2007 8:43 AM

I planning to build multiple web sites that needs to use the same masterskins, page layouts, content types and web parts. The content and navigation structure on these sites will not be the same. Should I use a separate Site Collection for each web site or put all web sites in one Site Collection?

dwollerman wrote re: Site Collection Logical Architecture
on 10-31-2007 12:21 PM

well, this depends on the deployment of the master pages, layouts, content types and web parts.

by default sharepoint shares these items within a site collection, so the quick answer would be to keep it all in a single site collection.

but if you are deploying these items as a feature / solution package then you can deploy them to multiple site collections keeping the solution project as the source file for these items. There is more work involved in the deployment method, but you might benefit in the future using a deployment scheme allowing you to use the mutlitple site collections.

you will have to wiegh the benefits of site collections vs sub site plus the pros and cons of creating a deployment package.

rebecca wrote re: Site Collection Logical Architecture
on 11-08-2007 10:22 PM

thanks for the great article...

I have been struggling with this age old sitecoll v site question myself...

I know the pros, recycle bin, quotas, turning on and off services in cent admin, but...

isn't it more of a pain to plan navigation between site collections than using sites and sub sites? and what about master pages, themes, page layouts & content types, CQWPs etc? how difficult is it to use these elements across site collections as compared to within the same site collection?

Backups can occur at the site level, no?

Here's my big question: can a top level site and its subsites be migrated later if it is determined a site collection is needed? How difficult is the process?

What about records centers, can there only be one per site collection?

Sorry for the novel size comment. Some implementations can be so hard to decide which to use...others are easy, slam dunks. In each case I want to be sure my client has the correct taxonomy to begin with-a solid foundation for the organization.

dwollerman wrote re: Site Collection Logical Architecture
on 11-09-2007 7:12 AM

I get this question alot, and usually it comes down to what i most important to what the customer is doing with their environment. You need to balance the flexability and features of a site collection with its limitations.

The limitations of a site collection are there for a reason and that reason is because site collections are not suppose to be related to anything other than itself. Site collections are a means to provide a stand alone secure container to allow like content to be shared within. This means that site collections and their duties can range from a corporate portal down to a simple project adhoc workspace.

Also where customers get confused in the concept is that they expect all naviagion and flow of the portal down to those projects sites to be seemless. Although Microsoft does give you the ability to attempt this with managed paths, and "connect to portal", the reality is that it doesn't encapsulate everything.

Using CQWP across site collection is not possible, you need to write a custom query part to do that.

Master pages, page layouts, content types, site columsn, etc.. would have to be packaged and shared through a solution and features.

Themes are at a farm level and are automatically avaliable for all sites in a farm.

Backups in 2007 can be done from a farm level down to a database level, plus with STSADM you can backup site collections. There is no full backup for a site level.

In 2003, you had "smigrate" which was the frontpage backup site method. That is gone in 2007 and replaced with content deployment mechanisums "export and import". These operations can be found under STSADM or under central administration content deployment section (only for MOSS).

Record center is just a site template, you can apply that just like any other template at any level in any site collection as many as you need.

dwollerman wrote re: Site Collection Logical Architecture
on 11-09-2007 7:27 AM

Also, to elaborate on the CQWP Across site collections, you can do alot with enterprise search mapped properties to site columns. Dan Attis wrote a greate article as an example:

devcow.com/.../SharePoint_2007_How_to_Rollup_Content_from_multiple_Site_Collections.aspx

Sandar Van Laan wrote re: Site Collection Logical Architecture
on 11-15-2007 6:16 PM

My thoughts, exactly, Dave. This conversation continues to come up, and separate site collections always seems like the way to go.

Sandar

Trackback: vanlaan.typepad.com/.../site-collection.html

Ivan WFP wrote re: Site Collection Logical Architecture
on 12-06-2007 8:28 AM

Hi Thanks for your great insights. I am having one problem when running the stsadm - o backup and restore commands. It restores everything quite nicely except the images and the templates that I downloaded off the web and included onto my sharepoint sites. I have looked everywhere for a solution but cant find any do you perhaps know if you can get stsadm back up option to include these sites when it does a single file back up?

dwollerman wrote re: Site Collection Logical Architecture
on 12-06-2007 8:57 AM

Everything in the site collection will be contained with a backup of a site collection. Make sure that the images and templates you downloaded exist on the farm and web application that you restored the site collection to. The content in the site template gallery should come over with the backup, any images that are located outside the site collection refereneced relativly might not have the same relative path in the new location. There is not enough information in your comment that I can offer a direct answer to. Hopefully this helps.

Aditya wrote re: Site Collection Logical Architecture
on 12-06-2007 2:36 PM

Hi Dave,

A great article on how SharePoint is architected. I had a question, when we type in a URL (managed path) and hit enter, what is the sequence of events that retrieve the data? I believe the URL is broken into chunks and the query to DB performed for each of the chunk? Can you please explain what events take place, by giving an example?

Kat wrote re: Site Collection Logical Architecture
on 12-14-2007 9:23 AM

Hi Dave,

Thanks for this terrific information! This is a subject that has been a constant source of concern and frustration for me. I'm an Intranet Manager trying to plan for an upgrade from SPS 2003 to MOSS. In our current intranet environment, we have little content at the portal level (just general news, events, and other non-department-specific info), and we have a site collection for nearly every department and several ad-hoc groups and general purpose sites. Most department and general purpose sites have a top-level site that is readable by all staff, and private collaboration sites are built underneath. We try to use the portal to link to key content in the "public" top-level sites. The problem is when folks expect to find a "forms" library or something like that which contains all forms for the organization. That type of thing has to be manually created w/ a links list that links to the forms stored in the individual "public" top-level sites.

Anyway, when looking at moving to MOSS, the content query web part seemed really appealing for being able to display content from multiple sites in one spot. We were thinking of consolidating into fewer site collections in order to take advantage of the CQWP. But I constantly am concerned about hitting storage allocation limits, having long lists of random list and site templates, breaking permission inheritance everywhere, etc. Your post states most of the reasons I like using the multiple site collections.

We really like having web sites that can be managed by individual departments. They can set it up the way they like, display the information and documents people need to get from them, and manage the security, which makes them happy. It allows them to take more ownership of their content, which helps keep info current. The problem really is always that nagging..."this belongs here...and here...and here" and no one wants to create a bunch of manual links that go dead the second something is moved or deleted.

Any additional advice for this nervous intranet manager? :)

Thanks!

dwollerman wrote re: Site Collection Logical Architecture
on 12-16-2007 12:06 PM

Use Search... :)

Think of the intranet as how a user uses the internet. On the internet there is no way to keep up with the information that is available by using manual links. Yes there are more important pieces of information that will require a link to make it more easily available. That content is usually not moving around alot and has a set place where it needs to be located, but for more targetted information for an individual, search is mostly used.

On your intranet, it would be virtual impossible to keep up with manual links that are important to all individuals. If an individual uses the search functionality of sharepoint they would be able to find more targetted information that pertains to them and the information they are looking for.

This of course all depends on the search being configured properly for the information people are searching on regularly.

Hope this helps.

Kat wrote re: Site Collection Logical Architecture
on 12-17-2007 9:23 AM

Thanks, Dave! That does help.

Dave Wollerman's SharePoint Blog wrote Site Collection Taxonomy Thoughts
on 01-02-2008 1:11 PM

This post is inspired by a comment I recently recieved from someone asking about when to use site collections

Robin wrote re: Site Collection Logical Architecture
on 01-09-2008 5:08 PM

Hi,

Thanks for this excellent article.

I am new to Sharepoint and have a question. For internet sites, we would like to have content organized according to countries and use the end-user's browser settings to detect the language/country and provide appropriate country specific content.

In this case, would you recommend creating seperate site collections per country or 1 site collection with multiple sites for each of the countries? Some additional details:

1. Each of the country pages on the site share a common navigation structure - i.e. it is just content that varies but the categories (or sub-sites) are similar.

2. On a few occasions, we also have a need to have the French site show selective German content (in addition to the French content)

3.  In a few cases, within a country site, we have to set up variations for 2-3 languages (say Canada - French and English), Belgium (Danish, Finnish). Can we have multiple such variations within a single site collection?

Thanks, again.

dwollerman wrote re: Site Collection Logical Architecture
on 01-09-2008 7:11 PM

For the most part depending on what needs to be shown and if there are no security issues, you could create audiences for each country.

As for the variations, yes they can be contained in the same site colleciton. and in conjunction with the audiences, you can use variations for the languages.

for the most part it might be easier to have each country be their own site collection to keep the management of the content using audiences down. It will depend on if you want to have duplicate content in multiple site collections or manage a single copy of the content allowing only specific content by using audiences.

Nicole wrote re: Site Collection Logical Architecture
on 01-16-2008 4:03 PM

Thanks so much.  I have been looking ALL DAY for this.

I wanted to know when to start a new site collection and you clearly answered my question.

Sanjar wrote re: Site Collection Logical Architecture
on 01-22-2008 1:53 AM

Excellent article Dave, struggled to understand from other sources but, you have made it simple.

John wrote re: Site Collection Logical Architecture
on 02-22-2008 6:21 PM

On your second big point (distributed administration): Isn't it possible to assign administration at the site level?

So for instance, when creating  a site you can choose to not inherit parent site permissions.  At that point it will ask you for the names of the groups that you want for Reader, Contributor, and Owner.  Then you can add whomever you want.  Can't anyone in the Owner group do anything with that site and its subsites?

dwollerman wrote re: Site Collection Logical Architecture
on 02-22-2008 8:24 PM

Yes, you are correct that you can apply unique security at the stie level, but keep in mind that sharepoint groups span the site collection level. this means that IT groups then are managed in the same area as HR groups or any other sharepoint groups. Also site collection administrators are not useful in this scenario since you cannot give IT user or HR user site collection admin status as they will automatically have admin permissions everywhere. There is also the "all user information" area. all users who visit the site are automatically entered into the all users information area, this can get quite full and confusing for managing users at any level throughout the collection.

The point i was making is that since site collections are unique onto themselves and do not interact with other site collections, you can assign owners to the sites. These owners can then manage their own security groups and members without affecting or being affected by other departments, teams, or projects.

Joe wrote re: Site Collection Logical Architecture
on 03-13-2008 9:31 AM

Good article.  I was confused by this section though:

The example I usually give is HR deletes 1MB per day and IT deletes 1GB per day. With a 5GB site quota, HR content will be flushed through the system a lot quicker if they share a recycle bin. This results in having to restore a database to get back a single document. (You know it will happen, Murphy’s Law).

What do you mean by "HR's content will be flushed through the system quicker"?  Do you mean that, because IT is filling up the quota at a fast rate, the Recycle Bin will be emptied more often than if HR was in its own site collection?

dwollerman wrote re: Site Collection Logical Architecture
on 03-13-2008 12:15 PM

The recycle bin has 2 stages. The first stage is based on a time limit (default 30 days). The second stage is based on the a percentage of your site quota (default is 50%). Plus the first stage is counted against your site quota limit, where the second stage is not. The second stage bin runs a first in / first out process.

So HR deletes a very important 100k fle on day 1.  IT deletes 1GB of content over the next week. Your site quota is 2GB, which means that your second stage recycle bin is 1GB by default. In 30 days this content gets moved out of the time based bin into the second stage bin which has a quota of 1GB. Since IT has deleted 1GB of content that week, the HR 100K file will be flushed out of the system on day 31 or so.

Now if HR had their own site collection and recycle bin storage. That 100K file could be around for a year compared to just over a month.

Karuana wrote re: Site Collection Logical Architecture
on 03-28-2008 6:08 PM

Great article... thanx for this input and everyone elses... One note about sharing data across site collections.. it seems to me a problem that MS should have addressed but since they didn't I use CorasWorks to roll up executive data across site collections and maintain master lists of information that are used across the Enterprise... For me we couldn't really implement the best of SharePoint without Corasworks...

Thanx!

Gregory S. MacBeth - View Gregory MacBeth's profile on LinkedIn wrote Decisions About When To Use a Farm, Web Application, Content Database or Site Collection
on 04-23-2008 10:15 AM

I wanted to say a few things that may help when deciding upon when a Farm, Web Application, Content Database

JayGolden wrote re: Site Collection Logical Architecture
on 04-23-2008 2:05 PM

Thanks, I was just firming up a SharePoint structure methodology document I have in place. This helped to add a few points that I missed.

Cheers,

Jay

Harhar wrote re: Site Collection Logical Architecture
on 04-30-2008 6:47 AM

Thanks for the article, very informative for a nube to SP.

Going back to "Also where customers get confused in the concept is that they expect all naviagion and flow of the portal down to those projects sites to be seemless. Although Microsoft does give you the ability to attempt this with managed paths, and "connect to portal", the reality is that it doesn't encapsulate everything."

I work in an organisation moving to MOSS 2007 that wants a lot of site collections (IT, HR etc.) but also wants the Intranet to have a common Global Navigation.

Is the easiest option to do the following:

Connect the Windows SharePoint Services Site to the Portal Site

Although this step is optional, Microsoft recommends that you configure the Windows SharePoint Services site to connect to the portal site. To connect your Windows SharePoint Services site to your portal site:

1.    Connect to your Windows SharePoint Services site, and then click Site Settings.

2.    On the Site Settings page, click Go to Site Administration under Administration.

3.    On the Top-level Site Administration page, click Configure connection to your portal site in the Site Collection Administration area.

4.    Click Connect to portal site, type the URL of the portal site in the Portal Web Address box, specify a friendly name for the portal in the Portal Name box, and then click OK.

dwollerman wrote re: Site Collection Logical Architecture
on 04-30-2008 5:54 PM

What you are describing is the ability to link the site up to the "Portal", which is whatever you decide to enter in as a URL. This connection will show up in the global breadcrumbs in the upper left of the screen. It will provide a link back up to the corporate intranet.

If the top or left navigation expectations are that they are the same no matter what site collection you are on there is two ways to do this.

1.) manually maintain the navigation (which removes the posibiliy of security trimming)

2.) Custom develop a site map provider that builds the appropiate navigation heirarchy based on site collections.

HarHar wrote re: Site Collection Logical Architecture
on 05-02-2008 3:44 AM

Thanks, thought that manaually updating would be the easiest way. Foruately it is going to be a pretty locked down structure so controlling it shouldn't be a problem

Dave wrote re: Site Collection Logical Architecture
on 05-02-2008 12:41 PM

Dave, we have lots of site collections. How can we create a list/report to identify who in the Site Collection Administrator groups for each collection?

Thanks

Dean

Brian Merrifield wrote re: Site Collection Logical Architecture
on 05-02-2008 5:04 PM

So, I'm a little confused in regards to multiple site collections within a web application. If I have my default web application running on port 80, doesn't that mean that only one URL will be "active" for one site collection?

I have our Intranet site setup really basic right now, but its time to add a seperate department from another site onto our SharePoint server. They want a custom URL, so I don't think a managed path will work (under our current URL).

I've read the specs for multiple site collections in a web application, but just don't see how it relates to a web application.

Thanks,

Brian

dwollerman wrote re: Site Collection Logical Architecture
on 05-02-2008 6:55 PM

Dave, there is nothing out of the box that will give you a list/report of the site collection administrators for every site collection. You will have to write a custom piece to extract that information.

You could possibly use the content databases and write a SQL query to pull the information out, but I do not recommend it since it is not supported by Microsoft.

dwollerman wrote re: Site Collection Logical Architecture
on 05-02-2008 7:12 PM

Brian, Web Applications are designed to handle multiple site collections. You can think of a web application as the equivelant to a IIS Virtual Server. The web application is technically just a container when created. It does absolutly nothing until you create a site collection.

When you create a site collection, you need to specify which web application the site collection will be contained. By default there are 2 managed paths. One is a root level (/) and the other is called "sites".

The root level managed path represents an explicit path which means that only one site collection can be created at that location. The "sites" managed path is a wildcard path, which means that multiple site collections can be created using that path.

Multiple web applications can use port 80, but they have to either be on a different IP address, or use a host header. Site collections run under the context of a web application, so one web application running under port 80 means that many site collections can run under the same port..

If they want a custom URL, you might need to consider creating a new web application for them, or look into using site collections with host header mode. Host header mode allows you to specify a DNS name for a site collection. I am not sure of the exact details, but you can see the option under the stsadm -help createsite command.

James wrote re: Site Collection Logical Architecture
on 05-06-2008 9:11 PM

This is an excellent clarification of Site collections.     I have a client that wants to take the existing single Site Collection with Subsites and split the subsites into separate Site Collections.  

Current hierarchy within One Site Collection:

hostname/Site Collection Top level Site

hostname/Subsite1

hostname/Subsite2

,,,

hostname/SubsiteN

The target architecture will have all subsites in its own Site Collection

hostname/Site Collection Top level Site

hostname/Site Collection for Subsite1

hostname/Site Collection for Subsite2

,,,

hostname/Site Collection for SubsiteN

There is a dependency in the current architecture such that Columns and Content Types are inherited by the subsites.

Can this type of site collection migration be accomplished using STSADM (export/import/backup/restore)?    Would the export/import operations export the any inherited items like columns and Content Type?  

How would you approach this migration from subsites to top level site collections?

Thanks in advanced!

dwollerman wrote re: Site Collection Logical Architecture
on 05-07-2008 5:45 AM

I don't think you are going to get the content types and site columns to come over with the export/import commands. The recommended method is to create a feature that contains all the shared content types and their related site columns. That way you can deploy to any site or site collection in your organization. There is a tool on codeplex that allows you to export content types to build the feature for cross farm deploymnet (www.codeplex.com/MOSS2K7CTypesViewer).

James wrote re: Site Collection Logical Architecture
on 05-07-2008 7:51 AM

Dave,  Thank you for the speedy response and advice.  I will give the Content Type deployment by Feature a try!  

Thanks again!

sharepoint site verses subsite wrote sharepoint site verses subsite
on 05-09-2008 5:46 AM

Pingback from  sharepoint site verses subsite

tmmurphy wrote re: Site Collection Logical Architecture
on 05-16-2008 1:54 PM

Just like James, Above - I'm taking over admin-ing a SP farm that has many subsites and now wants many site collections.  

The trouble is, they were all set up under the "/" explicit Managed Path, and now they all want to be listed under that "mainsitename.domain.edu/sitecollection" site names.  

James didn't mention this as a problem, but from what I see above (about the explicit path "/" taking only one site collection) and what I've experienced trying to do this (and I've tried using SP Designer as another "backup/Restore" agent ).  So - will it be possible to do this - move sub-sites up under the main site name, all as Site Collections?  I could even move them all to our development server, change something on the main site, then bring them back.

Thanks for any/all help!

-tmmurphy

dwollerman wrote re: Site Collection Logical Architecture
on 05-16-2008 2:21 PM

tmmurphy,

There are two types of managed paths, wildcard and explicit. By default a web application has a "/" explicit managed path and a "sites" wildcard managed path.

The explicit managed path allows only a single site collection to exist at that path. So anything underneath that path will be sub sites of that site collection. (Example: http://www.domain.com/)

The wildcard managed path allows multiple site collections to utilized that managed path. So for example the default "sites" wildcard managed path will allow multiple site collections to utilize it as shown below...

http://www.domain.com/sites/sitecollection1
http://www.domain.com/sites/sitecollection2
http://www.domain.com/sites/sitecollection3

If we break it down...

http://www.domain.com = Web application
/sites = wildcard managed path
/sitecollection# = top level site (site collection)

if we break down the explicit managed path...

http://www.domain.com = Web application
/ = explicit managed path & top level site (site collection)

So basically you can use either sharepoint designer or the stsadm export/import commands to take a sub web (ex: http://www.domain.com/subsite/demo) and turn it into a site collection using a wildcard managed path (ex: http://www.domain.com/sites/demo). You can create custom managed paths for say departments or projects. (ex: http://www.domain.com/departments/hr) that way they look like they all fall under the "/" root level of the web application in a more defined heirarchy other than just using "sites".

Chris H wrote re: Site Collection Logical Architecture
on 05-20-2008 8:33 AM

Dave, is there a way to change the managed path for an existing site collection?  For example, if I have a site collection going to the managed path of hostname/helloworld and I wanted to change it to hostname/hw.....is there a way to re-associate the existing site without doing a backup/restore to a blank site collection under the new managed path?

thanks for your help!

dwollerman wrote re: Site Collection Logical Architecture
on 05-20-2008 8:43 AM

You can try to use the stsadm o renamesite operation, but based on Microsofts KB article (support.microsoft.com/.../939535) it doesn't work with path based site collections.

If the stsadm command does not work for you, which based on the information it probably won't, you will have to do a backup / restore of the site collection into the new managed path. Obviously make sure the new manage path exists before attempting to restore.

stsadm -o backup -url http://hostname/helloworld -filename hw.bak

stsadm -o restore -url http://hostname/hw -filename hw.bak

Alan wrote re: Site Collection Logical Architecture
on 05-23-2008 4:38 PM

Hi Dave - great post.  We are currently running WSS 3.0, with no use of site collections.  We are expanding our use of Sharepoint, and will be implementing MOSS 2007 next quarter.  I would like to start using site collections as much as possible (e.g., one for each department), but one of my biggest concerns is in regard to searching across site collections.  We basically have three categories of sites:

1. Cross-department - access for all authenticated users (basically this is the corporate "intranet" function)

2. Department-specific - access for all authenticated users (department-specific pages on the "intranet")

3. Department-specific - limited access (collaboration areas for various departments)

Again, my problem is with search.  Having the ability to search across the corporate "intranet" has been a stated requirement (showing results from both #1 and #2), but having the ability to search across both a department's public and private sites has also been requested (showing results from both #2 and #3).  Assuming a user cannot be given an option to search across site collections, it seems that we won't be able to use site collections much without relaxing one of these two requirements.  Any suggestions?

dwollerman wrote re: Site Collection Logical Architecture
on 05-23-2008 9:32 PM

Alan,

With MOSS you gain the ability to search across site collections. WSS you can only search within the site collection.

MOSS will provide you the ability to search across the entire farm, plus additional areas such as exchange public folders and file shares. The search results are security trimmed so if the current user has permissions to see a certain area, they will see results from those areas.

Srini wrote re: Site Collection Logical Architecture
on 06-19-2008 7:03 PM

I am still not able to see the diagram. Is there a way you could provide a link to the image please?

Srini

dwollerman wrote re: Site Collection Logical Architecture
on 06-19-2008 7:34 PM

www.sharepointblogs.com/.../original.aspx

It was on another service, i moved it over so you should be able to see it now.

shadycat wrote re: Site Collection Logical Architecture
on 07-08-2008 6:32 AM

Excellent, thanks

Mike wrote re: Site Collection Logical Architecture
on 07-16-2008 7:20 AM

I'm very new to MOSS and our company is just at the design stage. I think I grasp Site Collections but when would I use multiple Web Apps?

We want to have an intranet and a customer facing extranet, I was planning on using seperate Web Apps for this (can take one offline with out affecting the other / restore one without affecting the other / more involved security on the extranet etc) but should I use seperate Site Collections instead?

Thanks

dwollerman wrote re: Site Collection Logical Architecture
on 07-16-2008 7:40 AM

Determining on when to use web applications is usually related to authentication types. You can have multiple authentication types in MOSS, but only 1 per web application. Another decision point is the content and how you are required to organize it and secure it.

From what you said you have an intranet and a extranet. Does the extranet require different authentication? if yes, then it has to be a new web application. Does the intranet contain content that is sensitive that should not be exposed outside the company? if so, then the extranet has to be a new web application.

Another thing you can do is if your intranet can be exposed to the outside, but still requires different authentication, then you can create a new web application that extends your intranet. What this does is allows two web applications to access the same content each with their own authentication.

I wouldn't use a site collection for this because you still have to expose the entire web application to the outside to get to the site collection. You are better off thinking in terms of web applications for the intranet and extranet. Plus, the extranet won't be a single site it is a network of sites, so your extranet web application most likely could contain multiple site collections itself.

Mike wrote re: Site Collection Logical Architecture
on 07-16-2008 10:20 AM

Thanks that really helps, it's likely that the intranet & extranet will have different authentication and some content that is specific to each so I'll use two web apps.

There'll be some content that needs to be shared accross both web apps but I should be able to get round that with search.

Thanks very much and sorry to hijack your Site Collection blog with Web App queries!

Site Collections Logical Architecture | 21apps wrote Site Collections Logical Architecture | 21apps
on 07-23-2008 12:27 PM

Pingback from  Site Collections Logical Architecture | 21apps

Propecia pill. wrote Propecia.
on 08-05-2008 1:51 AM

Side effects from taking propecia. Tadalafil and proscar propecia success. Propecia generic.

A.G. wrote re: Site Collection Logical Architecture
on 08-06-2008 2:56 AM

Dave, Thanks for this great article.  I have the following situation:

Client is composed of several business units (BUs) with unique branding and Internet presence but shared corporate governace in terms of HR, IT and Legal.  Client wants to create an Intanet where the BUs share the same HR/legal/IT announcements, forms (vacation reqs, contract approval req, hardware reqs) and related workflow.  In the meantime, client wants the BUs to maintain certain level of autonomy when it comes to their BU site's Look and Feel, BU specific content and possible day to day management.  Employees do move between BUs, but they have to be in the same BU for a minimum of a year.  

After reading your article and other documentation, I'm leaning towards creating separate collections for Corporate and each BU.  My big issue is how does corp publish targeted content to all BUs only once? and how would the different BUs access the "same" corporate content, and what to do about the corporate forms and workflow?

I greatly appreciate your insight.

dwollerman wrote re: Site Collection Logical Architecture
on 08-06-2008 7:12 AM

The problem is in the "pushing" of documents (and maybe content). You shouldn't have to push documents to the business units (such as forms and such) The employees should come to it. Each business unit could provide an informative page that allows their users links to get the information from corporate.

Think of corporate as just another business unit. If someone in BU-1 wants more information about what BU-2 is doing, they would go to the BU-2 site. They shouldn't expect to find that information on their site since it doesn't relate to BU-1. Point is.... Keep the content at it's home and have people come to it to consume it rather than having multiple pieces of the same information floating around out there.

For sharing announcements or aggregating content across site collections, I recommend using search. If you are running MOSS, then you can create your announcements in a structured manner (such as content types and site columns), then configure search managed properties to map to those columns, then you will be able to use the search core results web part (Again, MOSS only) to retrive that spceific content and render it with a custom XSL layout. With this option the content is still "at it's home" just being rendered in another location.

A.G. wrote re: Site Collection Logical Architecture
on 08-06-2008 7:51 AM

Thanks Dave!  your comments do make sense and are a big help.  

One thing I forgot to mention in my original post is corporate calendars.  corp manages a couple of calendars.  BUs would need to display corporate and BU specific events.  Would search managed properties per your suggestion above do it as well?

dwollerman wrote re: Site Collection Logical Architecture
on 08-06-2008 8:16 AM

Sure, you can have search pull the specific information, the underlying calendar content are just list items, which can be structured with content types and site columns to assist search in returning those results.

The issue comes with rendering the information. You would have to do some amazing XSL work to get the information displayed in a calendar view.. :)

A.G. wrote re: Site Collection Logical Architecture
on 08-06-2008 12:26 PM

So I ran the scenario of having corp as another "BU" site where users would go to it to access and fill corp forms/workflow.  

Turns out the unique BU look and feel aspect is a hard requirement.  Users should not go outside of their BU brand not even to corp site.  My question is how do I handle access to crop forms/workflow, image library, training material and such?  Many thanks...

dwollerman wrote re: Site Collection Logical Architecture
on 08-06-2008 12:42 PM

I think that is a political issue. If corporate controls these items outside of the business units I don't believe this should be any different in SharePoint or any intranet.

The only way I can see each business unit being able to maintain their own corporate level sites, is the information for each business unit is unique and cannot be used by any other business unit anyway.

If you need to do this I would see how the "search" option works for whay you need to do. This will provide the consumption of the content from the BU site, but still host it in the proper location. Keep in mind duplicating data should not be an option. thats just poor business practice.

Cat_SJ wrote re: Site Collection Logical Architecture
on 08-27-2008 3:15 PM

Thank you for this post.  I have a question, please.  We have a farm that has one site collection.  In this site collection there are webpages for an application.  The database for the application resides on a remote SQL server.  The remote SQL server is continously sending data to our Sharepoint SQL servers.  We'd like to house both the web app database and the content database on the same server.  The problem is that we are unable to host the web app database on our Sharepoint SQL servers.  So, we are wondering if it is possible to do this:  Create a second site collection for just this web app, and during the creation of the new site collection, tell sharepoint to house the content databases for this new site collection on the remote SQL server.   We still want to be able to use our sharepoint environment for the web based interface for this web app, just want to have the content databases on a different SQL server other than the Sharepoint SQL servers.  We do not want to setup a complete second sharepoint environment for this web app.  Thanks so much for your time and expertise.

dwollerman wrote re: Site Collection Logical Architecture
on 08-28-2008 7:16 AM

Cat_SJ,

Based on best practices and recommendations I would recommend against creating a database for sharepoint outside your sharepoint environment. Also, you will have to make sure all the permissions are in place for SharePoint to create and continue to manage the content database.

with that said, there is nothing in the technology that will keep you from doing so. First you must create the content database before creating the site collection. The database can be created by going to central administration -> application management and clicking on "Manage content databases". Make sure to select the web application you want to host the database and click on the "Add Database" link in the toolbar. From there you have the option of specifiying the sql server to host the new database.

Once the database is created, you can create the site collection. First you have to make sure the databases are configured properly so the new site collection goes into the correct database. Under the same "Manage content databases" section, you have to specify the number of sites allowed per database. You do this by clicking on the database name and setting the site warning and limit settings.

What I usually do is set the limit to the number of current sites + 1. (you have to add one because the system won't take it if the limit and the current number match). then make the warning smaller then the limit. This is done on the database you do not want the site to go into. On the database that you want the site to go into you have to make sure there is room in the database based on the site limits specified. by default it specifies 15000 so there should be plenty of room.

When this is done you can create your site collection and it should go into the database that has room for it. Once the site collection is created in the new database and if that is the only site collection going into that database, then set the limits of the new database, based on the information above, so no new site collections can be added to it.

purvish wrote re: Site Collection Logical Architecture
on 09-03-2008 10:44 PM

excellent

Alias George wrote re: Site Collection Logical Architecture
on 09-15-2008 7:54 AM

This is a request for information about implementing site collections.

   We are planning to implement three site collections. One site collection for our intranet which will have all the departments as individual sites. One site collection is planned for management. Each of the departments will have project tracking of its own. But we want the management be able to go to each department project tracking info from the management site collection.

   There is a separate site collection for the customers. We need to give access for customers into an area where issues and orders are managed.

   How do we implement communication between site collections?

dwollerman wrote re: Site Collection Logical Architecture
on 09-15-2008 8:08 AM

Communications between site collections are primative. There are a few ways to do this.

Quick and easy route: Use Links

You can use a links list to provide access to different areas of your environment

Moderate route: RSS

You can utilize the RSS options for lists and libraries to share information across site collection boundaries

Advanced route (MOSS Option only): Search

You can utilize custom search web part as a rollup list. This requires the creation of custom search managed properties which tie to metadata in sharepoint lists. The search results web part can be configured to target these managed properties to return precisly what is required. Also search results are security trimmed, which is sometimes a requirement for inter-site collection communication.

V2 wrote re: Site Collection Logical Architecture
on 09-25-2008 6:02 AM

Dave, Thanks for the informative content here. Cleared a lot of doubts.

Do you have any post where you have summarized:

- best practices to follow while setting up SharePoint environment

- Design considerations for performance optimization (for e.g. in cases when you anticipate heavy usage, or in cases when some sections, like team sites, are expected to face most of the requests.)

Thanks

V2

dwollerman wrote re: Site Collection Logical Architecture
on 09-26-2008 8:35 AM

V2,

These are all good ideas for posts.. thank you. unfortunately I haven't done anything around these topics, at least publically. I do have some information related to these (especially setup and installation) and will be posting them shortly.

Samer Shennar wrote re: Site Collection Logical Architecture
on 10-21-2008 5:12 AM

I got aware of this early in my deployment. I'm creating site collections on the departmental level. Section level is planned to be subsites. How ever my -BIG- question is: How to aggregate several site collections' content when needed? I hope this doesn't require asking the development team to do .net digging for every simple requirement. I appreciate any help.

An example of what I need would be or instance:

Leaves Callendar:

That is, if every department includes a calendar to record their staff leaves, how could an upper level manager look at all of the organization's staff leaves in a single calendar by collecting information from each department's site collection?

dwollerman wrote re: Site Collection Logical Architecture
on 10-21-2008 7:19 AM

Samer,

Microsofts answer to aggregation is Content Query Web Part (for intra site collection content) and their only real good solution for inter-site collection aggregation is search. If you are running MOSS, you can configure search to return the information you are looking for.

With the use of content types and site columns in conjunction search managed properties will allow you to configure the search core results web part to pull only specific relevant information. The bonus is since it is useing search to return this information, it is automatically security trimmed as well.

Since search results are returned using XML, all you need is to configure the XSL to render the content any which way you need. Making it into a calendar might prove complicated with straight XSL, but I am sure it can be rendered in such a way to make it a quick win for your situation.

Other options of course rely on developing a solution using .net, especially if you choose to render into a calendar.

Samer Shennar wrote re: Site Collection Logical Architecture
on 10-22-2008 1:21 AM

Thanks for your reply dwollerman. You mentioned some cool ideas that I were'nt aware of. However, your reply is also telling me that I have a big homework ahead, learning advanced search configuration :)

Tasneem Nomani wrote re: Site Collection Logical Architecture
on 10-23-2008 10:32 AM

We are using site collections for our different departments. Accounting, HR, IT Marketing.

Now each of these departments will have public section which will be viwed by entire company and private psection which is a collaboration space within the department.

The logic that I am using to organize this structure is, have a site collection for each departmen which is public, then create a subsite to this site collection which will be private to taht deparment.

Pleas let me know if I am on the right parh, or should it be reversed.

dwollerman wrote re: Site Collection Logical Architecture
on 10-23-2008 11:02 AM

Tasneem,

My recommendation is to not include the private collaboration information within the public site collection.

What I usually recommend is a public intranet site collection, which has public departmental information as sub sites. Then the private information is contained in their own site collections.

example...

http://intranet.company.com/

this is the root level site collection in the web application which contains the companies intranet which is public "formal" tightly governed information.

http://intranet.company.com/projects/???

site collections can use managed paths to define and organize team sites. Such as a projects managed path where all the project based site collections can be created, or a /departments/ managed path which allows department team sites to be created, or /sites/ which can be used for virtual team sites in general.

Keeping the private sites organized by specific topics which allow for better administration, maintenance, and lifecycle management of the site. along with providing the owner a more loosely goveren workspace for them to build something which suits their needs vs. someone elses.

Rita Reynolds wrote re: Site Collection Logical Architecture
on 10-25-2008 12:41 PM

I am working on developing a user guide for our IT staff as we set up our New SharePoint Farm.  Would it be ok to use your diagram above in that user guide.  It's by far, the best visual that I have come across for explaining the logical architecture

dwollerman wrote re: Site Collection Logical Architecture
on 10-25-2008 12:53 PM

Rita,

Yes, feel free to use my image in your documentation.

Kevin wrote re: Site Collection Logical Architecture
on 10-29-2008 9:55 AM

I've been busy designing an architecture for SharePoint, and came up with the following idea:

I would like to create a portal named customers. This would be a Site Collection. Then I wish to create a site collection for each customer, with subsites for each project we do for a customer.

Would this be a good approach?

And what about URLs. I would like URLs such as:

http://myserver/customers

http://myserver/customers/customerA

http://myserver/customers/customerA/projectA

It seems that this is not possible. If I create a manged path named 'customers', it is either a root path, or a path for which I need to specify an underlying path.

I can't host a site collection on the root of a path, and site collections on 'sub' paths?

Thanks!

Kevin

dwollerman wrote re: Site Collection Logical Architecture
on 10-29-2008 10:20 AM

That is correct. You cannot specify 2 managed paths with the same name. Also you will want to stay away from having a wildcard managed path named the same as a sub-web under the root level site collection. This will cause sharepoint to get confused.

Take a look at using host header mode. You can esentially create "vanity" urls for each customer site collection. for example... http://customer.domain.com/ and creating sub webs for each project (http://customer.domain.com/projectA).

Keep in mind, having all projects for a customer as a sub-web is ok as long as 1) you don't need specific lifecycle management for each project 2) user permissions are the same (you want to stay away from too much unique permission management) 3) archive ability using backup methods or locking mechanisums provided by SharePoint. You might run into management issues in the long run by having projects as sub-webs under a customer.

Although sub-webs provide a better means for sharing content types, site columns, and rollup capabilities... I would make sure you weigh all the pros / cons on creating them in site collections vs. sub webs.

Kevin wrote re: Site Collection Logical Architecture
on 10-30-2008 7:01 AM

Thank you for your reply.

We do multiple projects for multiple customers.

With lifecycle management I guess you mean cleaning up sites, backup/restore sites, etc.?

I think that projects need to be archived at some moment, otherwise the database would grow to large. Further more it would make the view messy, if discontinued projects would still show up.

Would that still be possible in this situation?

Also, different projects have different members. But 1 employee can work on different projects. In this situation, is it difficult to set up the appropriate rights? What do you mean with the user permissions are the same? The same for each project?

If I create a collection for each project, I will have to make a page in which I provide some form of structure per customer, right? With links to each collection for a project.

Could you please clarify the management issues that can cause troubles on the long run.

Thanks!

Kevin

dwollerman wrote re: Site Collection Logical Architecture
on 10-30-2008 8:42 AM

Kevin,

Let me break down your response and clarify each section...

Lifecycle management?

quick answer... yes. You want to position yourself and your environment to be effective in the long term as well as the short term. This means you have to develop a plan for archiving sites, being able to move sites between databases, content databases utlization, disk utilization, site locking, site disposition, etc.

Keep in mind how Microsoft has designed SharePoint. A content database can contain multiple site collections, although a site collection CANNOT span multiple databases. This is a main point of concern. Putting projects as sub-sites in a single site collection for a custom means you can only have one database. Multiple projects and the data associated with these projects can grow out of control if not propertly planned out. Also you cannot offer storage quotas for sub-sites... they all use the site collection storage quota. This means you either not set one (bad idea) or set one and have any one of the projects potentially utilize all the space for themselves.

Permissions?

What I mean by the same permissions is you don't want to get into the habit of managing unique permissions for sub-sites on different projects because it can get messy. If all the projects have the same permissions then great inherit permissions from the parent and manage the permissions in a single location. I tend to think this is not the case at the customer level. Odds are the permissiosn are unique at the project level.

With this said, you need to develop a plan for user management. This goes back to thinking for the long term as well as the short term. Short term you throw a sub-site out there, manage permissions for a few people who may just need access to that project... then before you know it they have you managing permissions for all projects across all customers. You need to delegate this out to project managers / admins. can you do this in sub-sites... sure, but now there are multiple different admins intertwined into the site collection structure, which means more uncontrolled growth. With a site collection you can more easily manage this piece as well.

Ah the infamous question, how do I manage all my projects for a customer in one place?

The easy answer... use a site directory. If you have moss you can utilize the site directory template for each customer which can provide a list of projects for them on their site. If you don't have MOSS and have WSS, you don't have the site directory template. Since the site directory template uses lists to store its data, you can create a custom list for them to track their sites with a list.

Another direction... use search. Again, if you have MOSS you can easily incorporate search to provide a list of site collections. The best part, search is security trimmed... meaning customers will only see site collections in which they have access to in the list. Since search results are rendered with XSL, you can easily customize the output the customers see.

Bottom line...

I highly recommend taking at least a couple days to make sure you cover most your bases and develop a long term strategy for sustaining the environment. Yes I know you won't be able to think of everything, but if you look to the past in how the company has done business and their trends and then follow within the lines of how Microsoft has architected SharePoint to work, then odds are for you in the event an unknown comes up. A well thoughout plan makes life easier for you in the long term. If there is no plan, every issue is a fire drill and a band-aid fix, which lead to more intense fire drills. You want to reduce the occurances and intesity of the fire drills and effectively resolve any unforseen issues... the best way to do this is to have a plan of attack and stay one step ahead of the curve.

Frank Sun wrote re: Site Collection Logical Architecture
on 01-12-2009 3:44 AM

After reading all the questios and answers for this article, I still have one question about SharePoint architecture. You recommended us not to put public and private content in the same site collection, but what if we need to share part of private information on the public site. For example, we have a IT site collection containing all systems we support now with contacts for each system and project status and so on.  Now we'd like to share the contact info to all colleagues in our company. I came up with a way to share this information without duplicating the list is to create a public site and use CQWP to show the information.

Do you think that makes sense?

Another question, if we don't care about the Recycle Bin, site report, and administration issues you mentioned above, do you know the size limit for a site collection? In other words, if we want to build a portal and lots of sites under the portal with a size, say, 200GB, do you think it's feasible?

Thanks a lot in advance.

dwollerman wrote re: Site Collection Logical Architecture
on 01-12-2009 8:20 AM

Frank,

It is not a rule to keep private and public information separate. The separation allows you to govern the content differently. If you have people continuously working on private content in a public environment, this means the administration of this could become more cumbersome. Usually public content is goverened and secured to where there are only a few allowed content authors where as a private area would allow for more contributing parties to secure without affecting the security structure of the public information.

You could keep all the contracts in a private area to work on them and expose just the contract information in your private area for colleagues to consume. If the contracts are designed properly (content types, specific site columns) and are provisioned and mapped to search specific properties. You can utlize the search web parts to essentially roll up this information on your public site for all to consume. (example fixed keyword search syntax: contenttype:Contract)

As for the other question. There is no hard limit on the size of a site collection and it will continue to grow as long as your database hardware allows. With this said you increase risk into your site collection by doing so. I have seen companies put everything in a single site collection ranging to 100GB+ while trying to secure each level of a site collection just to be able to use something like the CQWP or the "manage site structure" functionality. But you will increase risk to backup / restore recovery times, plus the stsadm commands for backing up just a site collection will be unable to effectively backup and restore something that large. (10GB is usually the limit on this). Then you have to worry about the amount of activity and transactions occuring on a single database in SQL Server and whether your hardware can keep up with the data locks not conflicting. Keep in mind everything in a site collection is stored in a single table in the database. This means every item from and list or library (document is stored in another table, metadata and reference is still stored with all list information) is stored in a single table. If you have all sites in a single site collection all lists items in every site is stored in a single table... I have seen 500,000 items plus in this table before... you have to also make sure your database hardware can keep up with delivering the specific queries from this table, even though it is indexed the more items you contain in table the more it has to index and sort through to find items for a specific list or library to display in your browser.

Sorry for rambling... hopefully this infromation helps you out.

Site Collection Taxonomy Thoughts | Berazinsky User Profile wrote Site Collection Taxonomy Thoughts | Berazinsky User Profile
on 01-12-2009 2:06 PM

Pingback from  Site Collection Taxonomy Thoughts | Berazinsky User Profile

Frank Sun wrote re: Site Collection Logical Architecture
on 01-12-2009 7:31 PM

Not at all. Thanks for your sharing lots of insights.

It seems MS provides an architecture not workable for each case. You have to sacrifice something for every architecture.

In fact, I think that's a weakness of CQWP. If it supports quries across site collecitons, no one will be worries about the architecture anymore cause you can organize your sites in whatever way you like and able to public information to other site collections without duplicating data. I know search solves some problems, but you need to think of keywords before you search. For published information, it seems not intuitive enough.

MOSS: Too many site collections? | keyongtech wrote MOSS: Too many site collections? | keyongtech
on 01-18-2009 10:58 AM

Pingback from  MOSS: Too many site collections? | keyongtech

Wef wrote re: Site Collection Logical Architecture
on 01-19-2009 1:24 PM

I'm amazed at how many times I refer to this blog post and its comments.  It's really guided me in better directions.

=Wef

Newly Noted #10 | Patrick Verbruggen's Blog wrote Newly Noted #10 | Patrick Verbruggen's Blog
on 01-27-2009 3:19 AM

Pingback from  Newly Noted #10 | Patrick Verbruggen's Blog

Manish wrote re: Site Collection Logical Architecture
on 01-28-2009 4:50 AM

Hi,

I am working on a Application for HR department catering various needs in an organization.This is an Intranet based web application used accross the locations.

This application has modules like Leaves, Attendance,Training, Travel,Performance Evaluation etc.

In a typical organization with various departments and designations , how to structure the product with Sharepoint.

I am evaluating the various approach like

a) Single site per User (Employee)

b) Single site per Entity in the organization.

You suggesting , One user per site collection.

Pls expain in the above context. Also if you have any other way to handle this, will be appreciated.

alternatively mail me at mains.garg@gmail.com

Regards and Thanks in Advance,

Manish Garg

dwollerman wrote re: Site Collection Logical Architecture
on 01-28-2009 12:54 PM

Manish,

With development in SharePoint, you want to see where in your application does the features of SharePoint provide value and build modules to extend what SharePoint already gives you to complete the objective of your application. Meaning you don't want to reinvent the wheel developing an entire application from scratch essentially duplicating functionality of SharePoint, just to surface and integrate with SharePoint. Use SharePoint as a platform and extend where needed.

First I would identify the functionaly required for this application and where this functionality overlaps with SharePoint functionality. If your requirement is to have an individual be able to track training information and be advised of additional training which pertains to them. Evaluate where SharePoint already gives you this functionality... SharePoint already can do audiences targeting specific content, alerts, storing of a list of information, etc... then you are left with the pieces you need to develop to extend SharePoint to fill the void. such as to add an action to a list or item which will auto register the user based on information from their My Site user profile, then track the information back in their my site.

SharePoint provides the ability through it's API and web services plus the Office client application to produce an application which delivers the functionality you expect without having to develop databases, underlying principles for the storing and disposal of content, scalabiltiy, authentication, etc... Through the use of deployment through a sharepoint solution package and activation in the site's which require this functionality through the "Features" feature you should be able to design it in such away to deploy the solution granular enough to meed your business needs.

As far as the A and B listed in your comments... item A is essentially a personal site or what is often refered to as a "My Site". I wouldn't create additional site collection outside of the My Sites. Item B is more along the lines of creating team sites or departmental sites.

Manish wrote re: Site Collection Logical Architecture
on 01-29-2009 6:42 AM

Thanks for the feedback!!

Application are to be used by employees. These are nothing but a part of Organization chart spread into various Operations,Functions,Deparments and teams.

If One has to map any Organization in WSS, How one should go about it.

What I feel If one can map the organization In WSS, any application can be built on that.

Now thats where I have different thoughts.

My options A and B in my eariler blog was pointing to that only.

I just want to map that.

Regards,

Manish

dwollerman wrote re: Site Collection Logical Architecture
on 01-29-2009 8:18 AM

Manish,

Well there is no linkage or relationships between site collections in SharePoint. So linking up an individual site to a organization site would not be an option.

Since you mentioned WSS, using my sites and user profiles is not an option with WSS. If you have MOSS you would be able to utilzie My Sites and user profiles to at least provide relationships between people, but still not organizations. There is no organization relationship feature for SharePoint. You would have to custom develop this feature..

As far as A and B go... I would say use both. If you are using MOSS you can utlize My Sites and user profiles to provide the automation of creating personal sites for individuals. If not, you would have to develop a process or automated feature that would provision a site collection per individuals. Organizations deserve their own site collection as well. This will provide an area for each organization to collaborate.

I would not recommend creating an organization site collection then having individual sites per person underneath. This will make for an administration nightmare as people move or come/go from different orgs. Having them in site collections allow you to more easily archive, provision and move if needed.

Karim wrote re: Site Collection Logical Architecture
on 03-04-2009 3:24 AM

Hi Dave,

thanks a lot for this post and for your comments, it is very helpful.

I have following questions.

I m working for a company that needs to implement a multiple farm SharePoint infrastructure. They have two farms located in Europe and in the US. There will be one global site collection that will be defined as a portal. Then they will be several Site collections for different countries.

The European Server Farm will contain site collection for Germany, Austria, France…

The US Server Farm will contain site Collection for the US, Canada…

Further Site collection for new countries will be added later.

Now my question is: Is it more interesting to create for each Site collection a separate web application? Or are two application in the European Farm (one for the Portal and one for the Country site collections) and one for in the US Farm (for the US country Site collection) enough?

My second question is about Application pools. What does this is about? Is it preferable to create for each web application a new application pool?

Maybe it is important to mention that we need to implement both: local country search and global portal wide search (including the portal site collection and the country site collections from both server farms)

Thanks,

Karim

Karim wrote re: Site Collection Logical Architecture
on 03-04-2009 4:14 AM

I just wanted to add that the country sites are not personal sites but that these are sites where documents are processed through a lifecycle and tasks are created for different persons.

Andre wrote re: Site Collection Logical Architecture
on 03-18-2009 9:36 AM

We are trying to determine the best way to logically architect are MOSS environment.  Currently we have an intranet which resides in its own site collection.  We are trying to determine how to best accomodate the below numbered items into our existing  MOSS platform/architecture where our intranet resides. In particular, how many site collections should we be targeting?:

1. a Project Server 2007 installation which resides on another physically separate server.

2. about 30-50 (to be created) sharepoint departmental "team sites". These sites are to be used for individual teams potentially exclusively or have limited read access given to a small subset of employees.

carol wrote re: Site Collection Logical Architecture
on 04-19-2009 3:23 AM

Hi

 I have created a collaboration portal.This would be my site collection.I have created two sites within it.And I have created Documents .Now I need to copy this documents from one site to the other and include a workflow in between and create a Report on top of it...

Can  you tell me the way to copy those documents and how to incude a workflow???

Thanks In Advance

dwollerman wrote re: Site Collection Logical Architecture
on 05-10-2009 8:28 PM

Karam,

Sounds like you have a lot going on there. As far as web applications are concerned, You can include the portal site collection and the country site collections into the same web applicaton. You may want to think about creating managed paths for each country (http://webapp/country/site) in the event you will have more than one site collection per country.

Application pools allow you to seperate out isolated processes on a per application basis... be careful as you have to scale these based on your server specs. To many can eat up your memory if not configured properly. The recommendation is the web app with central administration is to be in its own application pool with its own identity. Other than that it is up to you and your server specs on how you configure the remaining web apps.

You can manage the local country search with the SSP on the regional farms, but for a global search you would need to index the other farms as well. this is usually done from a "central" farm and people are directed to the central farm to search globally. this is because the farm which provides the global search has to index the other farms which can be an issue because of your network.

dwollerman wrote re: Site Collection Logical Architecture
on 05-10-2009 8:32 PM

Andre,

The seperate project server will contain their own site collections on a per project basis. I don't think I can give you a suffiecent answer based on such a brief description. I would say the 30-50 department sites would each be their own site collection, but after that I am not sure.

dwollerman wrote re: Site Collection Logical Architecture
on 05-10-2009 8:35 PM

Carol,

Are you asking how to manually copy the information over while retaining the workflow, or how to have a workflow copy things from one document to another.

I assume the later. You can develop a workflow which copies information from one site to another. However, it would have to be a visual stuido workflow as SharePoint designer workflow wouldn't be able to cross site collections.

MiscBytes wrote re: Site Collection Logical Architecture
on 05-29-2009 4:07 PM

Excellent post and just what I needed, as a beginner.  There is a ton to read in the SharePoint world, but I have never seen anything that explained about why/when to create site collections.  And of course the container pics are great.  Thanks!

Mark Grimes wrote re: Site Collection Logical Architecture
on 06-26-2009 7:49 PM

OK...so the obvious question, probably with an obvious answer...based upon your diagrams, if you have 2 web applications, then you must have 2 contect databases?  My concern is over my current project which will have authentication and anonymous users.  This is a case where I have read that separating these two groups will help to prevent cross site scripting attacks.  Do you agree?

dwollerman wrote re: Site Collection Logical Architecture
on 06-26-2009 9:18 PM

technically you would have 2 content databases if you have 2 web apps. However, that is only if you "create new" web applications. if you "extend" a web application, then you would have 2 web apps (IIS Websites) pointing to the same content database. This allows for multiple authentication types to be configured to access the same data.

If your just talking sharepoint, you cannot have a pure anonymous web application since you have to pick Windows, Forms or WebSSO for the authentication type when creating or extending a web app. However, once it is created or extended, you can configure the corresponding IIS web stie to be just anonymous without integrated authentication configured.

As far as being more secure... it probably is, but at what expense. The additional effort for configuration and maintenance may out weigh the minimal gain of security.

Add a Comment

(required)  
(optional)
(required)  
Remember Me?
Need SharePoint Training? Attend a SharePoint Bootcamp!
Posts (c) their respective authors. Everything else (c) 2009 SharePoint Experts, Inc.