Hi,

I am working on what will eventually be a large scale document repository. (Tens of millions of docs, mostly PDF.)  Obviously I am considering using Document Centers or Records Centers, but I am still a little unclear as to which direction I should go.  This is the kind of system where the documents themselves will not be edited or changed at all, the only changes might be changes to metadata values associated with the docs, and that, only on rare occasions.  Everything going in is considered to be finalized.  So my initial thought was that a Records Center would be appropriate.  

Part of the functionality of this system is that, although there will be limited addition of content, there is going to be a lot of search-and-retrieve type use, and I have read conflicting things about whether or not a Records Center is appropriate for that.  Can anyone clarify?

Also, I have read various suggestions for content database limits for Records Centers (namely 200GB, 1TB, and 4TB).  Which of those limits would be best for a system that will have write once, read many type functionality?

Thanks for any clarification / advice / suggestions!

Hi Dylan,

Records Centre sounds about right for that use.

However, there are supported rules for site collection and content database sizing.

In general, site collections should be under 100gb, and content databases under 200gb.

If you have only one site collection in the content database, it can grow to 200gb.

If there is only one site collection in the content database, and disk IO performance is at least 0.25 IOPS per gb, and you have not more than 20 columns per list, then it can grow to 4tb.

If there is only one site collection in the content database, and disk IO performance is at least 0.25 IOPS per gb, and it's based upon the Document Centre or Records Centre templates, less than 5% of the content in the content database is accessed each month on average, and less than 1% of content is modified or written each month on average, then it can grow indefinitely.

I have a summary of SharePoint boundaries, thresholds and supported limits here: http://joelblogs.co.uk/2010/10/19/sharepoint-2010-boundaries-thresholds-and-supported-limits/  

And there's a really, really useful TechNet article you need to read here: http://technet.microsoft.com/en-us/library/cc262787.aspx (especially the bit on content databases.

Hope this helps!

joel

Thanks Joel,  by those guidelines I see that I am going to be able to use the 4TB limit.

The article I read that made me question whether a Records Center or Document Center would be more appropriate is this one:

http://technet.microsoft.com/en-us/library/9994b57f-fef8-44e7-9bf9-ca620ce20734#plan_ent_typlarge

In the section on "Large-scale content archive", it describes a knowledge base type archive, with many users accessing content primarily to read it, as being based on a Document Center.  It also talks about a Records Center, but says "Many more users submit content to a records center than view or read it," which is not what's going to be happening with my system.  I wasn't sure if that distinction was made because there was something going on behind the scenes that would affect search and retrieval performance from a Records Center.

Hi Dylan,

Although you said this in your first post, I didnt pick up on the knowledge base aspect of your requirements, and the scale of reads vs writes.

I would agree that a document centre is probably more appropriate.

Regards,

joel

Ok, so there is a performance difference between the two, when it comes to heavy search and retrieval use?

Good to know, because I definitely want to design a system that is going to perform well in the way my customer wants and needs.

But, it's a bit disappointing because:

A) by going with a Document Center, I have to go with the 1TB limit, and will be reaching the point of having a distributed archive much sooner, and

B) another part of the functionality of the system is that it is supposed to also have the typical information policy / records management functionality that a Records Center provides. (i.e. automatic application of disposition rules, etc.)  So this means I'm going to have to use in-place records management in a Document Center to make it behave like a Records Center?

[I'm fishing for either confirmation that that's a good strategy, or that it's not because the better records management functionality inherent in a Records Center (vs. in-place records management) outweighs the search-and-retrieve performance gains from a Document Center.]

- Dylan

I know this thread is old but it certainly sounds exactly like what I am about to embark on, although the percent of data that will be read each month is not going to approach knowledgebase levels.

If you did finally decide to do this project in SharePoint, can you please share a little of your experience with what happened and what choices you made? I would love to know how it went. Also, are the "boundaries" for document and record centers actually coded into the product (hard limits) or are they recommendations (soft limits)?

Thanks!

Limits on database size are soft limits. Theoretically, it can grow up to a maximum size supported by SQL server.