3
Deploying Oracle Web Cache

Note: Oracle Web Cache is compatible with Oracle HTTP Server or any other HTTP-compliant application Web server.

This chapter presents several scenarios for deploying Oracle Web Cache. This chapter contains these topics:

• Caching Content for One Application Web Server

• Forwarding HTTPS Requests

• Load Balancing Requests Among Application Web Servers

• Accelerating Portions of a Web Site

• Using Oracle Web Cache Servers in a Failover Pair

• Working with Firewalls

• Deploying Oracle Web Cache Servers in a Distributed Network

Caching Content for One Application Web Server

Oracle Web Cache can be deployed on the same computer as the application Web server or on a separate computer.

Figure 3-1 shows Oracle Web Cache deployed on the same computer as the application Web server.

Figure 3-1 Oracle Web Cache On Same Computer As the Application Web Server

Text description of the illustration deploy9.gif

For this deployment, configure Oracle Web Cache with the host name of the application Web server.

Figure 3-2 shows Oracle Web Cache deployed on a different computer from the application Web server.

Figure 3-2 Oracle Web Cache On a Different Computer From the Application Web Server

Text description of the illustration deploy10.gif

To configure this deployment:

1. Register the IP address of Oracle Web Cache server with the Web site's domain name.

2. Rename the application Web server, and assign the computer running Oracle Web Cache with the name that was previously assigned to the application Web server. In Figure 3-2, Oracle Web Cache is named www.server.com, which was the name of the application Web server. The application Web server is renamed to www-internal.server.com.

3. Configure Oracle Web Cache with the host name of the application Web server.

Forwarding HTTPS Requests

Note: In future releases, you can configure Oracle Web Cache to process both HTTP and HTTPS requests.

In this release, Oracle Web Cache can cache documents for HTTP protocol requests. For HTTPS protocol requests, you can use a Layer 4 (L4) switch to pass all HTTP requests, typically port 80 traffic, to Oracle Web Cache, and forward all HTTPS requests for secure pages to a particular application Web server. Figure 3-3 shows an L4 switch passing HTTP requests to Oracle Web Cache server www-internal.server1.com and HTTPS requests to application Web server www-internal.server3.com. Please note that HTTPS requests could also be passed to www-internal.server2.com.

An L4 switch operates at Layer 4, the Transport (or protocol) layer, of the Open Systems Interconnection (OSI) model. L4 switches determine where to send requests based on the protocol and port number.

See Also: http://www.ietf.org/ for information about the OSI stack

Figure 3-3 Using an L4 Switch to Forward HTTPS Requests

Text description of the illustration deploy2.gif

To configure this deployment:

1. Configure the L4 switch with the host names of Oracle Web Cache server for HTTP requests and the application Web server for HTTPS requests.

2. Configure Oracle Web Cache with the host names of the application Web servers.

Load Balancing Requests Among Application Web Servers

Many of today's Web sites use a Load Balancer to balance the incoming requests among multiple Web servers. Instead, as shown in Figure 3-4, you can use Oracle Web Cache to distribute HTTP requests among two or more application Web servers.

Figure 3-4 Load Balancing with Oracle Web Cache

Text description of the illustration deploya.gif

To configure this deployment:

1. Assign the name of the Load Balancer to Oracle Web Cache.

2. Configure Oracle Web Cache with the host names of the application Web servers.

If you prefer to use a third-party Load Balancer, you can use the deployment depicted in Figure 3-5. In this deployment, an L4 switch sends HTTP requests to Oracle Web Cache server www-internal.server1.com and HTTPS requests to application Web server www-internal.server4.com. The Load Balancer distributes requests from Oracle Web Cache among the pool of application Web servers.

Figure 3-5 Load Balancing with a Third-Party Load Balancer

Text description of the illustration deploy8.gif

To configure this deployment:

1. Configure the L4 switch with the host name of Oracle Web Cache server for HTTP requests and the application Web server for HTTPS requests.

2. Configure Oracle Web Cache with the host name of the Load Balancer.

3. Configure the second Load Balancer with the host names of the application Web servers.

Accelerating Portions of a Web Site

Many Web sites contain cacheable catalog content and non noncacheable secure and search content. For these Web sites, you can use Oracle Web Cache servers to cache content for just the portions of the Web site with the cacheable content. Figure 3-6 shows an Layer 7 (L7) switch passing catalog requests to Oracle Web Cache server www-internal.server1.com and order entry and account requests to application Web servers www-internal.server2.com and www-internal.server4.com.

An L7 switch operates at Layer 7, the Application Layer layer, of the OSI model. L7 switches determine where to send requests based on URL content.

See Also: http://www.ietf.org/ for information about the OSI stack

Figure 3-6 Accelerating Portions of a Web Site

Text description of the illustration calad012.gif

To configure this deployment:

1. Configure the L7 switch with the host name of the Oracle Web Cache server.

2. Configure Oracle Web Cache with the host names of the application Web servers for which it is caching documents. In this example, Oracle Web Cache server www-internal.server1.com is configured to cache for application Web server www-internal.server3.com.

Using Oracle Web Cache Servers in a Failover Pair

To maintain performance during an application Web server failure, you can configure two Oracle Web Cache servers as a failover pair. Both Oracle Web Cache servers are configured to cache the same content. When both Oracle Web Cache servers are running, a Load Balancer distributes the load among both servers. If one server fails, the other server receives and processes all incoming requests. This deployment is depicted in Figure 3-7.

Figure 3-7 Configuring Multiple Oracle Web Caches as a Failover Pair

Text description of the illustration deploy5.gif

To configure this deployment:

1. Configure the Load Balancer with the host names of the Oracle Web Cache servers.

2. Configure each Oracle Web Cache server with the host names of the application Web servers.

Working with Firewalls

You can deploy Oracle Web Cache inside or outside a firewall.

Figure 3-8 shows Oracle Web Cache positioned inside a firewall. Deploying Oracle Web Cache inside a firewall ensures that HTTP traffic enters the Demilitarized Zone (DMZ), but only authorized traffic from the application Web servers can directly interact with the database.

Figure 3-8 Configuring Oracle Web Cache Inside a Firewall

Text description of the illustration deploy7.gif

Figure 3-9 shows Oracle Web Cache positioned outside a firewall. With this deployment, the throughput burden is placed on Oracle Web Cache rather than the firewall. The firewall receives only requests that must go to the application Web servers. This deployment requires securing Oracle Web Cache from intruders.

Security experts disagree about whether caches should be placed outside the DMZ. Oracle Corporation recommends that you check your company's policy before deploying Oracle Web Cache outside the DMZ.

Figure 3-9 Configuring Oracle Web Cache Outside a Firewall

Text description of the illustration deploy6.gif

Deploying Oracle Web Cache Servers in a Distributed Network

Many Web sites have several data centers. For networks with a distributed topology, you can deploy Oracle Web Cache at each of the data centers.

Figure 3-10 shows a distributed topology in which Oracle Web Cache is distributed in offices in the United Kingdom, Japan, and the United States. Browsers make a request to local DNS servers to resolve www.server.com. The local DNS server is routed to the authoritative DNS server www.server.com. The authoritative DNS server uses the IP address of the browser, the topology model, and the current load to pick an Oracle Web Cache server to satisfy the request. It then returns the IP address of the appropriate Oracle Web Cache server to the browser.

Figure 3-10 Distributed Caching

Text description of the illustration deploy3.gif

To configure this deployment:

1. Configure the local DNS servers with the location of the authoritative DNS server.

2. Configure the authoritative DNS server with the host names and IP addresses of the Oracle Web Cache servers throughout the distributed network.

3. Configure each Oracle Web Cache server with the host names of the application Web servers it is caching documents for.

Another distributed deployment solution is depicted in Figure 3-11. In this deployment, an Oracle Web Cache server is located in the United States office and another server is located in the Japan office. The application Web servers for both offices are located in the United States office, centralizing the data source to one geographic location.

Figure 3-11 Centralizing the Data Source

Text description of the illustration deploy11.gif

To configure this deployment, configure each Oracle Web Cache server with the host names of the application Web servers for which it is caching documents. In this example:

• Oracle Web Cache server www.us.server.com is configured to cache for application Web servers www-internal.us.server1.com and www-internal.us.server2.com

• Oracle Web Cache server www.jp.server.com is configured to cache for application Web servers www-internal.jp.server.com.