Yesterday’s announcement that Microsoft had released Service Pack 1 (SP1) for the Office Filter Pack 2010 provoked some thoughts about the importance of this component to Exchange 2010. In fact, the Office Filter Pack is a prerequisite for both mailbox and transport server roles and you cannot install these roles on a server if you haven’t first installed the Office Filter Pack.
According to Microsoft’s blog entry, the “updated filter pack doesn’t have any fixes or updates that impact Exchange scenarios or functionality”, a somewhat curious state of affairs as you’d assume that a service pack contains something of interest. The conclusion that I have is that the fixes and updates incorporated into this service pack are of interest to other products – or something like that. Some of the backchat from Exchange engineers about the update indicates that Exchange customers don’t have to rush to get the update. They can simply wait for Office Filter Pack 2010 SP1 to appear through Windows Update and then apply it.
Exchange first started to use the Office Filter Pack in Exchange 2007 when they introduced content indexing as part of the mailbox server role. As items are added to a mailbox database, Exchange indexes the content into a set of catalog files maintained for the database. Users who work in online mode can use the indexes to perform searches. Usually these are OWA users as Outlook users typically work in cached Exchange mode and use local indexes for searches, only resorting to content indexes when they need to find something in an archive mailbox. Interestingly, the Mailbox Replication Service (MRS) in Exchange 2010 updates the content index for the target database as it moves content from source to target mailbox to ensure that the content index is fully up-to-date when MRS eventually switches the Active Directory pointers to allow a user to access their moved mailbox.
Exchange 2010 introduced discovery search functionality as part of its initiative to help organizations control information better. Discovery searches depend on content indexes and can only find information if the content indexes are updated and fully populated. If the content of an item cannot be understood by Exchange its content will not be fully indexed and might therefore be undiscoverable by a search.
Apart from searches, Exchange uses content filters on transport servers to allow the transport engine to examine content in messages as they pass through its pipeline. Transport rules that need to block messages that contain certain words or phrases make heavy use of the filters to locate suspect messages.
Exchange content tends to come in two forms: the MAPI properties for an item (subject, title, date created, and so on) and any attachments. The body part of a message is stored as a MAPI property and is either plain text, RTF, or HTML. The MAPI properties can be indexed using Exchange’s access to the Information Store but attachments pose a different challenge because an attachment can be in a huge variety of formats from the common Office document types (Word, Excel, PowerPoint) to binary ZIP files to Adobe PDF. The Office Filter Pack contains a set of IFilters to enable Exchange to access and index information held in common Office formats. IFilters are code that allow search services to index content of specific file types. In the case of the Office Filter Pack 2010, those formats are:
A useful change made to Exchange 2010 SP1 registers the IFilters automatically so that you don’t have to run scripts to make the IFilters known to Exchange.
As you can see from the list, the set of IFilters bundled in the Office Filter Pack gives Exchange mailbox and transport servers reasonable access to content. In most companies, the most obvious missing filter is one to allow Exchange to access content in Adobe PDF documents. You can fix this problem by downloading and installing the free PDF IFilter from Adobe’s web site. This IFilter certainly works and it can be used with other Microsoft products such as SharePoint 2010 but some commentators report that they get better results (in other words, documents are indexed more reliably) with Foxit Software’s version of a PDF IFilter. You pay your money and make your choice…
A warning to finish. Don’t assume that you can simply install SP1 on a server to get a fully functional Office Pack. Instead, you first install the RTM version of the Office Filter Pack 2010 and then you apply SP1. As noted above, the installation of the RTM version of the Filter Pack is sufficient to meet the prerequisite for Exchange 2010 mailbox and hub transport servers and you can install SP1 anytime afterwards. Maybe on a wet Wednesday afternoon.
The only other idle thought going through my mind is whether Microsoft has already updated its Exchange 2010 servers running inwith the new filters. I’m sure that they have. After all, isn’t software maintenance a big part of why we pay Microsoft the big bucks when we take out an Office 365 subscription?