Black ID

In this section

• 4GT Architecture
• 4GT Processes and Interactions

The technology called 4-gigabyte tuning (4GT) is a scale up component in the Windows Server 2003 family. The Microsoft operating environment delivers scalability via two methods: scale up and scale out. Scaling up refers to running a single application or image on a single server and having the ability to incrementally add system hardware resources (processors and memory) to increase overall system performance. Scaling out refers to distributing the computing workload among multiple servers with the ability to add or subtract servers to increase or decrease capacity.

Enterprises are required to scale flexibly to respond to spikes in business and to support the rapid roll out of new products and services. It is therefore increasingly important that an organization’s computing infrastructure provides the ability to increase or decrease computing capacity almost instantly, while continuing to deliver value. 4GT helps provide this scalability.

There are few restrictions regarding the environment in which 4GT can be enabled. 4GT has the following hardware and software requirements:

• x86-based processor
• 2 gigabytes (GB) or more of RAM
• Windows Server 2003, Enterprise Edition, or Windows Server 2003, Datacenter Edition
• Windows Server 2003, Standard Edition, but only in non-production environments.

4GT should not be used if either of the following is true:

• Greater than 16 GB of physical memory are available and Physical Address Extension (PAE) X86 is enabled on the server.
• Important performance characteristics of the application are adversely affected by limiting the kernel to 1 GB of RAM, for example, more than 1 GB of RAM might be needed to support the maximum number of concurrent connections.

Note

• 4GT is not required on the 64-bit versions of the Windows Server 2003 family.

4GT alters the way the memory manager allocates the virtual address space available on the computer. The following diagram and corresponding table illustrate this process.

4GT Architecture

When virtual memory space is requested by the user mode applications or the kernel, the memory manager determines where that memory will physically reside. The virtual address to physical address translation information is kept in the page table, which is part of the kernel’s virtual address space. Using this translation data, the memory manager accesses the data

The following table describes the components related to the 4GT architecture.

Components of the 4GT Architecture

In some cases, restricting the kernel to only 1 GB of memory might not be desirable, but the kernel might not need 2 GB either. In order to further refine memory management options, a new parameter, /Userva, has been introduced in Windows Server 2003, Standard Edition; Windows Server 2003, Enterprise Edition; and Windows Server 2003, Datacenter Edition. This parameter can only be used in conjunction with the /3GB switch. The /Userva parameter allows an administrator to determine the specific size of the user mode virtual address space between 2 GB and 3 GB. The value specified by the /Userva parameter in the Boot.ini files (expressed in MB) determines the amount of memory the memory manager will return to the kernel. The amount of memory returned is 3072 MB less the value specified by the /Userva parameter. For example, to increase the kernel’s memory space by 172 MB, the following parameter combination would be used in the Boot.ini file:

3GB /Userva=2900

This additional kernel address space is held in reserve and used as additional address space for Page Table Entries (PTEs) in the event that the system runs out of free PTE space. This address space is not allocated to PTEs until the system runs low on PTE space. For more information about PTEs, see “Page Table Entries” later in this section.

Page files store data that cannot be kept in RAM. The default page file size is equal to 1.5 times the amount of physical memory on the system. On a 4 GB computer, this can result in a page file larger than 4 GB. Under such circumstances, the effectiveness of this large of page file could be minimal, based on the usage of the system, such as when hosting database management applications. However, there are also times when even larger page files are needed, such as when multiple applications are consolidated on to one system. Therefore, on servers with 4 GB or more of memory, it is often preferable to change the default size of the page file based on the usage of the server. Regardless, on x86-based computers with 4 GB or more of physical memory there should be a page file sufficiently large to capture a kernel memory dump.

A page file size of at least 2,050 MB is the minimum size necessary for computers with 4 GB or more of physical memory to capture a complete kernel memory dump. The page file can be much larger (up to three times the physical RAM is typical) if the system is being used for application consolidation.

The page table is used by the memory manager to track the physical location of the page files. When using 4GT, because of the smaller amount of virtual memory allocated to the kernel, the number of Page Table Entries (PTEs) that can be recorded by the memory manager in the page table is also significantly reduced. In a standard configuration (without 4GT), there are typically 80,000 to 140,000 PTEs available. This is reduced to only 40,000 PTEs when using 4GT. This can lead to problems when the amount of physical memory increases and the kernel is required to keep track of larger and larger amounts of PTEs.