Home
Overview and Installation
Getting started with MemVerge Memory Machine requires that you understand some background information about how persistent memory works. You can then install Memory Machine on a server and begin monitoring memory usage and capturing application snapshots.
Introduction to MemVerge Memory Machine
MemVerge Memory Machine (Memory Machine) enables applications to expand from volatile memory (DRAM) into persistent memory (PMem).
Introduction to PMem
Using volatile memory (DRAM) and persistent memory (PMem) in combination enables applications to expand memory capacity relatively cheaply and to persist memory data objects, a capability that previously was only possible by writing to storage devices.

Overview and Installation
Getting started with MemVerge Memory Machine requires that you understand some background information about how persistent memory works. You can then install Memory Machine on a server and begin monitoring memory usage and capturing application snapshots.
- Introduction to MemVerge Memory Machine
  MemVerge Memory Machine (Memory Machine) enables applications to expand from volatile memory (DRAM) into persistent memory (PMem).
  - Introduction to PMem
    Using volatile memory (DRAM) and persistent memory (PMem) in combination enables applications to expand memory capacity relatively cheaply and to persist memory data objects, a capability that previously was only possible by writing to storage devices.
    - Memory Performance by Type
      PMem offers access times superior to those of any storage media. DRAM is faster, but it is considerably more costly.
  - Transparent Memory Service
    The Transparent Memory Service (TMS) intercepts Linux glibc.so function calls and reroutes them to the MemVerge Memory Allocation Service (MVMAS), which manages the MemVerge PMem store.
  - The MemVerge Memory Allocation Service
    MemVerge MemVerge Memory Allocation Service (MVMAS) is at the heart of Memory Machine. MVMAS manages memory allocation to volatile memory (DRAM) and persistent memory (PMem) for all applications run under Memory Machine.
  - Memory Machine Management Service
    Memory Machine Management Service (Management Service) runs alongside the Memory Machine data service to handle tasks not having to do with application data.
  - Understanding Snapshots
    A snapshot is a static image of a running application's process tree. You can use Memory Machine ZeroIO Snapshot to almost instantly capture a snapshot, then later just as quickly restore the running process to the state captured in the snapshot.
- Getting Started with Memory Machine
  You run Memory Machine on a host equipped with one or more PMem devices.
Using the Memory Machine Management Center
Memory Machine Management Center (M3C) is a web UI application available on the host that is running the Memory Machine Management Service. You can view system resources and performance metrics; create and restore snapshots; schedule jobs; and more in M3C.
Using the Command Line Interface
You can administer and use MemVerge Memory Machine using the Command Line Interface (CLI) on any installed host.
Working with Applications
MemVerge provides additional instructions for some popular applications supported by Memory Machine, either to assist with complex setup or to support additional functionality.
All Documentation
Links to all technical documentation for MemVerge Memory Machine including previous releases.

Introduction to PMem

Using volatile memory (DRAM) and persistent memory (PMem) in combination enables applications to expand memory capacity relatively cheaply and to persist memory data objects, a capability that previously was only possible by writing to storage devices.

Nonvolatile, or persistent, memory (PMem) is a type of computer memory that does not lose its data when power is turned off. PMem is cheaper than system dynamic RAM (DRAM), which is faster; and faster than solid state storage devices (SSD), which are cheaper. PMem therefore occupies an entirely new role in server-side computing data systems that traditionally has been split into storage (SSD or hard disk storage) and memory (DRAM and various caching schemes).

Memory Machine can be integrated with existing applications without code changes and recompilations. This is done using a custom preload library and a memory allocation service that are transparent to running applications. See Transparent Memory Service.

For more precise control of memory allocation in volatile vs. non-volatile memory, applications can be modified using an available API. The API enables developers to customize applications to select which memory items to place into DRAM or PMem to lower latency and system load.

Exploiting the properties of PMem enables Memory Machine to offer previously impossible memory-based data services.

Restriction: ZeroIO Snapshot based services are available only in Memory Machine Advanced Edition.

ZeroIO Snapshot enables images of application instances to be almost instantly captured and restored to local or remote PMem.
Snapshots can be cloned to run several versions of the same application instance at once.
With Memory Machine Publication-Subscription Service (MM Pub-Sub), an application can publish time-series data, such as stock ticks, to serve multiple client subscribers with low-variability, near-zero latency.