I'm looking at our GP for Office 2010. I think, for 2007 enabling Outlook Anywhere allowed it to be configured and, I think, for Office 2010 if it's enabled in GP, it checks the box to turn it on. Are there reasons to not enable Outlook Anywhere internally?

No not really. It is only on Citrix I have Outlook running in Online Mode. Cashed mode doesn't require as much I/O on the Exchange server since it will look at its own cashed files instead of always querying the Exchange server for the files. /MartinExchange is a passion not just a collaboration software.

Thanks for the reply. I'm not asking about Cached Mode. I'm asking about Outlook Anywhere / RPC over HTTP.

If I'm correct, Outlook Anywhere isn't used internally. Well, not unless you configure your outlook to use RPC over HTTP. I've seen some setups before where RPC over HTTP was used because it would only need 1 port to be opened on the internal network. Personally, I don't see the immediate advantage of it :-)

Yeah, that's where my quetsion comes from. To get laptop users globally configured, Outlook Anywhere was set in GP. No issues have come up from this, as you say it isn't needed internally, so I was just wondering if there is a reason to change the setting in GP. In this case AD guys will have to create another GP for laptops - that's not my issue but I was wanting to know if there is anything negative from the setting internally.

Hi, The setting in the ADMX is so that you can setup the Outlook Anywhere centrally so you don't have to go to each workstation to do it. But as Michael wrote it isn't used internally normally but it can be setup but no reason to do that at all. /MartinExchange is a passion not just a collaboration software.

Hi, Actually, Outlook can get good performance by using TCP/IP compare to the HTTPS (RPC Over HTTS). For the TCP/IP, the Outlook connects the Exchange server directly. For the Outlook Anywhere, Outlook sends the request to the CAS (RPC Proxy unmarshal the RPC packet from the HTTP which comes from the Outlook client), then forward the request to the endpoint which the client wants to talk to. For more information about how does Outlook Anywhere work, you can refer to the below article: http://msexchangeteam.com/archive/2008/06/20/449053.aspx Thanks AllenAllen Song

There are currently a number of pcs on the LAN and out on the WAN that are configured to use rpc/http. Some of these pcs are also currently configured to use cached mode. When we roll out offices 2010 we will have to disable cached mode to prevent large mailboxes from crippling the WAN to their office. Later I will rollout cached mode a little at a time. At the same time I want to get the rpc/http setting right and remove this setting from non-mobile systems. I’m wondering what effect this will have on the Mailbox servers when wan connected pcs that will be in online mode connect directly to the server. When configured to use RPC/HTTP on a local network, will outlook try to use tpc and directly connect to Exchange first? -there is no noticeable slow down with the RPC/Http setting. Now, what about when cached mode/online mode is thrown into the situation? If a pc is on WAN, say a 1mb pipe, and configured to disable cached mode, would this be slower than RPC/Http? I started to read this article before, then cached mode got thrown in the equation. I had to stop and determine what policy/script is controlling cached mode and Rpc over http. We have no network or slow outlook issues currently but I need to be sure when we make these changes we still have no issues.

If you configure Outlook Anywhere in Exchange, then Outlook 2007 and higher will be configured to use it on all clients by Autodiscover. The configuration that is pushed out by Outlook Anywhere is the configuration where Outlook attempts to connect via TCP/IP first, and then fails over to HTTP. Therefore in an office situation, while clients are configured to use it, they would rarely, if ever would actually do so. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

Thanks Sembee and Alan! This is what I was looking for. This is where my question was coming from, I thought that Outlook 2010 would connect directly to Exchange internally, even if OA was configured. So, I don't need to worry about trying to disable OA on internal clients! And, mobile clients will connect directly with TCP/IP when they are plugged in to the network here and will use RPC/HTTP when they connect outside the network. Do you know where / how the failover is determined? Say the user is connected to the corp network over a vpn. DNS would route them internally to the Exchange server for a TCP/IP connection, but is there a bandwidth threshold that pushes the failover to RPC? Along with this, how can I tell what method a paticular client is connecting with? I want to be able to demonstrate this.

Thanks Sembee and Alan! This is what I was looking for. This is where my question was coming from, I thought that Outlook 2010 would connect directly to Exchange internally, even if OA was configured. So, I don't need to worry about trying to disable OA on internal clients! And, mobile clients will connect directly with TCP/IP when they are plugged in to the network here and will use RPC/HTTP when they connect outside the network. Do you know where / how the failover is determined? Say the user is connected to the corp network over a vpn. DNS would route them internally to the Exchange server for a TCP/IP connection, but is there a bandwidth threshold that pushes the failover to RPC?

thanks for the good discussion! I was able to document and demonstrate for the team. 128k is the slow/fast failover point? Now I need the same info for cached mode.

Hi, The default settings for Exchange Proxy Settings are just that: with the Connect to option "On Fast networks" not set Outlook will attempt RPC connections first. if the connection fails, then Outlook attempt to use the "On slow networks" if set which is by default. Note: The 128kbps connection speed is a carryover from a long time ago. This is setting is derived from the network adapter speed and not based on any bandwidth calculation. So for the most part, it is irrevelant. If connection is made via RPC, the connection is considered fast otherwise, it is slow. For the Cached Mode, it's not related to Outlook Anywhere. Refer to the advantage of the Cached Mode, you can view the KB link: http://support.microsoft.com/kb/870926 Thanks Allen Allen Song

Thanks Allen When users connect to the company network via vpn, then open outlook, HTTPs shows in the connection status. Of course the speed is dependent on many things. With a connection that appears to be very good, HTTPS still shows as the connection method. Outlook anywhere and cached mode are blamed by users for just about any issue, password prompts, "my email is slow", I need to be able to demonstrate that each is not the problem. Have there been any changes to cached mode since Outlook 2003?

Take them out of cached mode and then they will see slow email. Cached mode means that all of their email is stored locally. If it is slow, then the problem is with their machine, not Exchange or the functionality. If you stop using cached mode then everything goes across the wire. Unless you have a very fast WAN connection (so not xDSL) then it is going to be very slow. The biggest cause of performance problems in Outlook is simply the number of items in a folder. Keep it below around 5000 and Outlook is fine. If you have 10,000 or more items then it is going to be slow. Nothing you can do will change that, other than putting items in to sub folders. Password prompts can be largely eliminated if you enabled Integrated/NTLM authentication on Outlook Anywhere AND the client machines are members of the domain. You will need to test it though, as some firewalls break that authentication method. Simon.Simon Butler, Exchange MVP Blog | Exchange Resources | In the UK? Hire Me.

Thanks Sembee, You know there are perceived issues and real issues, for some reason these 2 features have been perceived as a problem. I need to gather and demonstrate everything I can to help correct this. For cached mode I do have one real issue and that is pushing mailboxes out to slow wan sites. We can’t push Office out with cached mode turned on to everyone at every site, this would cripple the network at some of those slow sites. And that would definitely not help the way cached mode is perceived. My next step is going to be looking at the headers only setting and see if it will be possible to turned cached mode on a little at a time with headers only going to “slow connections” This is also 128k by default but can be set in GP.