The following article is part 2 of the Hybrid configuration checklist and pre-requirements article series.In…
Implementing Exchange Hybrid configuration in Office 365 environment can consider as a simple task or exhausting process.
The experience that we will have depended on the “readiness level” of the Exchange On-Premise infrastructure and, other related components.
In simple words – in case that you understand what are the pre-requirements that need to be configured and implemented, before starting the Exchange Hybrid Configuration Wizard, most of the chances that running the HCW (Hybrid Configuration Wizard) will be an easy task and, the process will complete successfully.
In the current article, we will review:
- The pre-requirements of Exchange hybrid environment
- Best practices and recommendation for the required preparations
- Tools and methods that will help us to check and verify if the on-Premises environment was configured correctly.
Hybrid deployment in Office 365 | Checklist and pre-requirements | The article series
The article series includes the following articles:
- Hybrid deployment in Office 365 | Checklist and pre requirements | Part 1/3
- Hybrid deployment in Office 365 | Checklist and pre requirements | Part 2/3
- Hybrid deployment in Office 365 | Checklist and pre requirements | Part 3/3
Table of content
Part 1: Hybrid deployment in Office 365 | Checklist and pre-requirements | Part 1/3
Part 2: Hybrid deployment in Office 365 | Checklist and pre-requirements | Part 2/3
Exchange Hybrid environment configuration in a Nutshell
The term Hybrid configuration or a Hybrid environment, describe a scenario in which two separated Exchange organizations that belong to different Active Directory forests are working as a “one unit”.
The term Hybrid configuration was created, for describing this type of relationship between the Exchange On-Premise infrastructure and the cloud (Exchange Online) infrastructure.
For example, in the following diagram, we see the logical concept of Hybrid environment.
The Public Domain name: o365info.com, configured as a “shared Domain”. The meaning is that two separate Exchange infrastructure “represent” this domain name or shared between them the same domain name.
When looking at the diagram, we can see two recipients: Bob@o365info.com and Alice@o365info.com
Technically, the recipient mailboxes must be configured on the Exchange on-Premises server or at the Exchange Online server, but logically, Bob and Alice don’t know where their mailboxes hosted.
In case that Bob mailbox is hosted on the Exchange on-Premises server and Alice’s mailbox is hosted on an Exchange Online server, Bob and Alice will have all the standard Exchange services such as: Free\Busy time, mail tips and more as if they hosted in the same Exchange organization. The reason is that the Hybrid environment, “connect” the two distinct Exchange environments and making them appear as one entity.
Hybrid configuration relationships and Trust concept
As mentioned before, the Hybrid configuration was designed for “connecting” two different Exchange environments and make them operate as one entity. A trust concept implements the “glow” between the two distinct environments.
- Federation trust – each of the Exchange environments (on-Premises + cloud) needs or must trust a “third element” named: MFG – Microsoft’s federation gateway (number 1 in the diagram).
- Exchange organization relationships – a trust model between two separate Exchange organizations. In Hybrid environment, the Exchange organization relationship is implemented between the Exchange on-Premises forest and the “Office 365 forest” (Exchange Online) (number 2 in the diagram).
The Hybrid configuration and the “Trust model” enable each of the “end points” (Exchange on-Premises and Exchange Online) to:
- Authenticate each other
- Verify the identity of each other
- Create a secure communication channel: Encrypt the information and implement data integrity by using a public certificate and by using a secure communication protocol such as SMTP\TLS and HTTPS.
Simple Exchange on-Premises environment versus complicated environment.
The term “Hybrid configuration” could use in describing a very simple scenario in which the organization has only one Exchange on-Premises server, who serves as a Hybrid server and is responsible for creating the “communication channel” between the on-Premises environment and the “cloud” (Exchange Online).
Another scenario of Hybrid configuration could be a more complicated situation, which is more common in enterprise environments that have complicated Exchange on-Premises infrastructure.
In this scenario, the “relationship” between the Exchange Online and the “on-Premises Exchange infrastructure” could be divided into many “communication channels” with different\separated Exchange on-Premises servers.
For example – the mail flow between Exchange on-Premises and Exchange Online could be implemented by using a “dedicated” Exchange on-Premises server which will be configured for sending mail to Exchange Online and, other Exchange on-Premises server which will be configured to “accept” mail from Exchange Online.
Another Exchange on-Premises server could assign to different roles\services such as dedicated Exchange on-Premises server who will provide AutoDiscover services, dedicated Exchange on-Premises server who will provide EWS services and so on.
Pre-requirements for Hybrid deployment in Office 365
1. Exchange Hybrid server version
The term “Exchange Hybrid server” is just a logical term that describes Microsoft Exchange server which can be a part of a Hybrid environment.
Note – the Hybrid environment based on two different “end point” such as Exchange on-Premises environment and the “cloud” (Exchange Online) environment. At the current time, the “cloud side” of the Hybrid configuration is based on Exchange 2013 SP1 technology.
The Exchange on-Premises server “Hybrid server” could be implemented by using:
- Exchange 2010 SP3
- Exchange 2013
Exchange 2010 SP3 as Hybrid server
In case that we want to use an Exchange 2010 as a Hybrid server, the minimum requirement is service pack 3. Besides of the requirement for Service Pack 3, the best practice is: to install the most updated Exchange Rollup versions because each of the software updates (Exchange Rollup) includes a solution to issues\problem that was discovered and the fixed is included in the Rollups.
Many times the customer or the organization IT will “resist” to the recommendation of “installing the most updated Exchange rollup“ but, it’s important to emphasize that installing the most updated Rollups can prevent many of the future problems and consider as an important factor in the process of building the Hybrid environment.
The following quotation relates to Rollups 4 for Exchange 2010 SP3, but you get the idea.
[Source of information: Understanding Hybrid Servers in Exchange 2010 Hybrid Deployments ]
Additionally, we recommend installing future Update Rollups 4 for Exchange 2010 SP3 on all your hybrid servers. Microsoft releases update rollup packages approximately every six to eight weeks. The rollup packages are available via Microsoft Update and the Microsoft Download Center. In the Search box on the Microsoft Download Center, type “Exchange 2010 SP3 updates rollup” to find links to the rollup packages for Exchange 2010 SP3.
Download link for the required Exchange on-Premises server software updates
|Exchange service pack and software updates|
|Exchange 2010 service Pack 3: Microsoft Exchange Server 2010 Service Pack 3 (SP3)|
|Update Rollup 32 For Exchange 2010 SP3: Update Rollup 32 For Exchange 2010 SP3 (KB5000978)|
|Exchange 2013 Cumulative Update: Exchange Server 2013 CU23|
Exchange 2010 | How can I know what is the current Exchange Rollup?
In case that you want to get information about the existing status of the Exchange 2010 on-Premises server, you can view the current version by using the Help menu and click on the About Exchange server 2010.
In the following screenshot, we can see that the Exchange on-Premises server version is: 14.03.0.195.001
So the next question could be: how can I know what is the Exchange on-Premises server service pack or rollup version based on this number? To be able to “translate” the value to a clearer information, we can use the article: Exchange Server and Update Rollups Build Numbers
In the following screenshot, we can see that the version number: 220.127.116.11 is “telling” us that the Exchange 2010 on-Premises server includes an installation of Service Pack 3 + Rollup 6 for Exchange 2010 SP3.
In this case, we will need to download and install to most updated Rollup (for example Rollup 9)
[Source of information – Keep Exchange up to date]
The Cumulative Update (CU), Rollup, and Service Packs you have running on the on-premises server should also not be overlooked. Under normal circumstances we support you being no more than two updates behind the currently released update for Exchange; however, for hybrid environments, we are stricter, and you should not be more than one build behind.
If the latest update is Exchange 2013 CU9, then you must have either Exchange 2013 CU9 or CU8 to be considered in a supported state. We are stricter with our hybrid requirements because of how tightly the on-premises and Exchange Online environments will be coupled together. For more information on our available updates please go here.
New Hybrid server versus existing Exchange On-Premise
In case that the organization Exchange infrastructure based on older versions of Exchange such as Exchange 2003, 2007, we will need to “add” or install a new Exchange on-Premises server (2010 SP3 or 2013) that will serve as the “Hybrid server.”
The “New” Exchange On-Premise Hybrid server can implement as Exchange 2010 or Exchange 2013 but, the best practice is to install Exchange 2013 server instead Exchange 2010 because Exchange 2013 includes improving features that relate to the Hybrid environment.
You can read more information about the improvement in Exchange On-Premise 2013 in the following articles:
2. Exchange On-Premise Hybrid server | Public IP address and Public name (FQDN)
Hybrid configuration is all about enabling Exchange On-Premise server which is configured as “Hybrid server” to create a communication channel with the Exchange Online infrastructure that exists in a public network.
To be able to communicate hosts or “endpoint” on a Public network, the Exchange Hybrid server must have:
- Public Name – The public name of the Exchange Hybrid server should be published in the Public DNS and should resolve to the Public IP of the Exchange Hybrid server.
- Public IP address – A Public IP address that “Point” to the Exchange Hybrid server should be assigned. Most of the time, the Public IP address will not directly attach to the Exchange on-Premises server, but instead, the Public IP will be allocated to a Firewall server. Th eFirewall will accept the communication requests to the Exchange on-Premises server and forward the request to the internal IP address of the Exchange on-Premises server.
Note – In case that we use more complicated scenario in which the on-Premises environment is “represented” by more than one Exchange on-Premises server, each of the Exchange On-Premise servers will need to have a dedicated Public IP.
For example, in case that the Outbound mail flow based on two Exchange on-Premises servers who can send mail to the Exchange Online server, each of this server will need to have a dedicated Public IP address.
For example, in case that the Outbound mail flow based on two Exchange on-Premises servers who can send mail to the Exchange Online server, each of this server will need to have a dedicated Public IP address.
How can I know what is the Public name of the Exchange On-Premise?
The simple answer is that if you are the Exchange On-Premise Administrator, you supposed to know what the Exchange On-Premise public name is but, in some scenarios, we will have to configure and hybrid deployment in an environment which we are not familiar with.
One option to get information about the “Public name” of the Exchange On-Premise server is by looking at the “External URL” that appears in the “client access” section under server configuration in the Exchange MMC (when we use Exchange 2010 MMC).
In the following example, we will look under the “Server configuration\Client access\EX01”
Exchange server “publish a ” couple of services. In our example, we look at the ECP tab (the ECP tab includes the Internal + External URL of the Exchange server using the web management interface).
We can see that the “pubic name” (External URL) of the Exchange On-Premise is: mail.o365info.com
Note – the External URL information includes parts that are only relevant for the URL syntax. Part of the URL is the host FQDN (Fully Qualified Domain Name). In our scenario, we are looking only for the Public Exchange server name (mail.o365info.com).
Verify that the Exchange Hybrid Server Public name (FQDN) is mapped to his Public IP
The verification process of the Exchange On-Premise Public IP is very simple. Open the command prompt and Ping the Public name of the Exchange On-Premise server. In our example, the Exchange On-Premise public name is: mail.o365info.com
In the following screenshot, we can see that we got as an “answer” the public IP of the Exchange On-Premise server.
One of the most common misconceptions is – that there is a problem because we got a “Request timed out”.
This response is not a sign of a problem because, the host whom we “ping” (Exchange On-Premise in our scenario), was not supposed to reply to the ping request. This is a foreseeable result because most of the time the organization Firewall blocks the ICMP protocol (that used for the Ping reply).
To recap: the fact that we got a response the Public IP is the required results, meaning the Exchange on-Premises server have a public name + Public IP address.
Note – besides of verifying the Exchange Hybrid server public name and Public IP; we will need to check additional parameters such as the ability to access the Exchange Hybrid server using a particular protocol and so on. In the next sections, we will review these other requirements
3. Exchange On-Premise Hybrid server | Port number and protocols
Hybrid configuration based on sharing data and services between Exchange Online and Exchange on-Premises server. The communication channel implemented by using two communications protocols:
- HTTPS – access to the Exchange services (from Exchange on-Premises server to Exchange Online and vice versa) implemented by using the HTTPS protocol.
- SMTP – the SMTP protocol used for implementing mail flow, and the data is encrypted using TLS (TLS over SMTP).
The underlying assumption is that – the Exchange on-Premises server protected by a Firewall. To be able to implement the communication channel, between the Exchange on-Premises server and the Exchange Online successfully, we will need to verify that the Firewall includes the following inbound and outbound rules:
- The inbound rule that enables to access the Exchange on-Premises server using port 25 (SMTP) and 443 (HTTPS).
- The outbound rule that enables the Exchange on-Premises server to access Exchange Online using the port 25 (SMTP) and 443 (HTTPS).
Reference from Microsoft article
The following screenshot is taken from a Microsoft article and includes a table the describe the port number and the services that need to enabled for Hybrid configuration
Verify access from the public network to the Exchange Hybrid server
To check that the required rule configured on the organization Firewall and, the Exchange on-Premises server can be accessed from the public network, we can use the option of “online scanning.”
There are a couple of websites that offer these services.
Personally, I love to use the mxtoolbox website. In the next step, we use the mxtoolbox for check what are the available ports on the Exchange on-Premises server.
In our scenario, the Exchange on-Premises server name is – mail.o365info.com
To get all the available options, we will choose the more menu and then the scan option.
We will provide the hostname (or the IP address) of the host that we want to scan.
Note – in case that there is an implementation of an inbound rule that enables access to the Exchange on-Premises server only from a particular Public IP range such as the Public IP range of the Exchange Online servers, the results of the online scan will not be relevant.
Verify that the Exchange Hybrid server can access external network using the required ports.
To be able to verify that the Exchange on-Premises server has access to public hosts using the SMTP (Port 25) protocol and can initialize a TLS session, we can use the Telnet command from the Exchange on-Premises server command prompt.
We can see that the destination server “reply” and we get a response that the target server is ready.
To start the SMTP\TLS session, we will type the command: helo, and after we get the “response” from the destination server, we can type the command: starttls
In the following screenshot, we can see that the destination mail server responds with the message: “220 2.0.0 SMTP server ready.”
[Source of information – Troubleshoot a hybrid deployment]
Availability – correctly publishing the on-premises Client Access servers to the Internet is vital to features working correctly in your hybrid deployment.
For hybrid features to work correctly, you must configure your on-premises firewall or other security appliances to allow inbound access from the Internet to the Autodiscover and Exchange Web Services (EWS) endpoints on the on-premises Client Access servers. Additionally, the Client Access servers must also be configured to accept inbound SMTP mail.
If the Microsoft Exchange Online Protection (EOP) service included in your Office 365 tenant organization can’t reach the on-premises Client Access servers, secure mail transport from the Exchange Online organization to the on-premises organization will not function correctly
4. Exchange On-Premise Hybrid server| Public IP address and Static NAT
An important factor that we need to verify is that the Exchange on-Premises server is using the public IP address that was assigned to him when he responds to a communication request of external hosts or when he initializes a connection to an external host.
The technical term for this scenario could be as a “two-way static NAT”.
For example: when using the Exchange 2010 Hybrid configuration wizard, we need to provide the public IP of the Exchange on-Premises server which is “allowed” to send an E-mail to the Exchange Online server.
When the Exchange on-Premises server communicates the Exchange Online, it is important that the Exchange on-Premises server use the public IP that configured in the Exchange hybrid wizard.
Other examples could be when the Exchange Online starts a communication process to the public IP of the Exchange on-Premises server. In this case, the Exchange Online server is “waiting” for a response from the IP address that used for starting the communication channel.
An example for the static NAT rule could be:
In the following diagram, we can see an example of static NAT rule. When external hosts such as Exchange Online try to communicate with the Public IP of the Exchange On-Premise server, the “response” from the Exchange on-Premises server implemented by using the same public IP address that we use for “publishing” the Exchange on-Premises server.
Verify what is the Public IP address that the Exchange on-Premises server uses
In case that we want to be sure that the Exchange on-Premises server “respond” or, access external network by using the Public IP address that was assigned to him (to verify that the Static NAT rule configured correctly), we can use a free service that is offered by many websites, that will enable us to get the Public IP that is used by specific internal hosts.
All we need to be open a browser search page from the Exchange On-Premise server and type: “what is my IP.”
In the following screenshot, we can see “what is the Public IP” of the Exchange on-Premises server.
5. ISA-TMG server and a Firewall server
When using ISA\TMG server to publish an Exchange On-Premise server, the configurations are a little bit different compares to a “standard Firewall” because, ISA\TMG is a Proxy server and additionally, Firewall server.
When using a “Standard Firewall,” we redirect the communication to the internal Exchange On-Premise server by using a simple “access rule”.
When using ISA\TMG Firewall, redirection to the internal Exchange On-Premise server is implemented by using a: Web publishing rule.
The ISA\TMG web publishing rule relates to a particular or pre-configured Exchange On-Premise “path” such as -OWA, EWS an additional component that used in the ISA\TMG environment is the Authentication settings.
Because ISA\TMG is a proxy server, many times the configuration of the authentication process implemented in the following way: external host authenticates (provide his credential) to the ISA\TMG, the ISA\TMG server approves or disapproves the credentials and if the complete successfully ISA\TMG will “forward in” the communication request of the external hosts.
In the hybrid environment, this configuration will cause problem and errors. In simple words: when we publish Exchange On-Premise server using ISA\TMG server, we need to cancel or disable the option in which ISA\TMG server is authentication external host’s communication request.
1. ISA\TMG & Authentication settings
The communication between Exchange Online and Exchange on-Premises server (and vice versa) is implemented by creating an encrypted communication channel and, by enabling each of the sides to authenticate and identify each other.
We need to enable this “channel” to exist without any interruption or Intervention.
ISA\TMG server is a proxy server that is sometimes is configured to enforce authentication process for the end point or, the host who wants to communicate with “internal host.” When relating to Hybrid environment and to a scenario in which the ISA\TMG publishes the Exchange on-Premises server, we need to cancel the setting that enforces the external endpoint to “authenticate to the ISA\TMG server.
When using the ISA\TMG rule for publishing Exchange on-Premises server AutoDiscover and EWS services, we need to verify that the TMG web publishing rule is configured with the setting: No delegation, but the client may authenticate directly.
2. ISA\TMG | Exchange On-Premise | Web Publishing rule and Path
When using ISA\TMG Firewall to publish internal hosts such as Exchange On-Premise, the “access rule” is implemented by a Web Publishing Rule.
The ISA\TMG Web publishing rule includes parameters such as – the name\IP of the internal host (Web Server) that we want to publish and additionally, a particular path of the web server.
In a Hybrid environment, it’s important to verify that the ISA publishing rule includes the following path:
Note – A major part of Exchange services is implemented as Web services. Exchange uses the IIS server (Web Server) as an infrastructure for the different web services and each of the different Exchange services, is represented by a dedicated IIS virtual folder.
The ISA\TMG web publishing rule should be configured to enable access to additional Exchange On-Premise web services such as OWA, Active Sync and so on.
When relating to Hybrid configuration scenario, the relevant virtual folder that need to be configured in the ISA\TMG rules are the EWS and the AutoDiscover folders.
Reference from Microsoft article
The following screenshot was taken from a Microsoft article.
In the following table, we can see reference for the:
- Path name that need to be added to the ISA\TMG publishing rule
- Authentication – the need to disable or cancel the ISA\TMG feature that require
Authentication (and instead, let the client to authenticate directly before the Exchange On-Premise server)
6. Firewall inbound and Outbound access policy | Office 365 and Exchange Online Public IP range
In many organizations, because of a regulation or other security requirements, there is an implementation of outbound and inbound policy that restricts access only to a dedicated or a predefined IP range.
For example: when we say:” Exchange on-Premises server is creating a communication channel with Exchange Online,” what does is mean from the “IP range” point of view?
Does Exchange Online infrastructure represented by a particular or a predefined public IP range?
The answer to this question is: “Yes.”
All the Office 365 environment such as – the Windows Azure Active Directory, Exchange Online, SharePoint Online and so on, based on a “publish” or well-known public IP range.
The implementation of Outbound and inbound firewall rules that restrict the access only to a specific or a predefined IP range consider as “good practice” from the security point of view. But, at the same time, can complicate and interrupt the process of the “first-time configuration” that we use for building the “Hybrid communication channel” between the Exchange On-Premise and the Exchange Online server.
Before we start to implement a specific organizational requirement for implementing outbound\inbound policy with a restricted IP range, my recommendations are as follows:
1. TEST\Pilot Phase
In case that you are only in the Test\Pilot phase, the main purpose is to verify that we can successfully use the HCW (Hybrid Configuration Wizard) for establishing the communication channel between the Exchange On-Premise and the Exchange Online (and vice versa).
In this phase, my recommendation is not to implement the outbound\inbound policy with a restricted IP range because, in case that this range not configured correctly, the Hybrid configuration wizard will fail, and it’s not so easy to find the cause of the failure.
After we successfully create all the Hybrid required settings (by using the HCW or by using PowerShell command) and verify that the communication path between the Exchange on-Premises server and the Exchange Online is “working,” you are most welcome to implement any outbound\inbound policy with a restricted IP range.
2. Learn and get familiar with the Office 365 Public IP range.
We will not get into many details about the subject of:” Office 365 Public IP range” but, it is crucial to understand that the term “Office 365,” is just a logical term that defines a very complicated set of infrastructures, services, Public IP ranges and so on.
My recommendation is to read thoroughly the information about the different Office 365 infrastructures and the dedicated Public IP range that is assigned by Microsoft for each of the different Office 365 infrastructures before starting with the creation of Firewall rules, Network range and so on.
3. The dynamic nature of the Public IP range
Office 365 is a dynamic environment. The basic assumption is that from time to time, there will be updated that relate to the Office 365 public IP range. To be able to stay updated and to be notified about features or current update that relate to the Office 365 Public IP ranges, you can read Office 365 URLs and IP address ranges.