Rack-Scale Solution 

Supporting Memory Pooling and Acceleration 

MEMORY POOL 

Rack-Scale Solution Supporting 

Memory Pooling and Acceleration 

MEMORY

POOL

Disaggregated System Solution to Overcome Data Center Infrastructure Constraints 

Within a single server platform, computational tasks requiring large memory capacities are limited by expansion limits of DIMM-based memory in CPUs and capacity constraints of GPU memory (HBM).

To address this, a disaggregated system, which separates servers by function (CPU server, Accelerator server, Memory server) and connects them via CXL interface, has been proposed.

Acceleration Memory Pool serves as a memory server that performs its own computations within this Disaggregated system solution.

Increasing System Efficiency
with Flexible Memory Allocation 

Larger amounts of data to process requires more memory, which often leads to stranded memory within individual systems. 

By supporting memory pooling, which dynamically allocates or releases memory resources as needed,
the occurrence of stranded memory is reduced,
thus enhancing system efficiency.

 

Scaling Memory to Petabyte Levels by Connecting SSDs 

Rack-scale Memory Pool, which maximizes memory expansion by connecting SSDs, can significantly reduce the total cost of ownership (TCO) for data centers. 

Disaggregated System 

Solution to Overcome 

Data Center Infrastructure

Constraints 

Within a single server platform, computational tasks requiring large memory capacities are limited by expansion limits of DIMM-based memory in CPUs and capacity constraints of GPU memory (HBM). 

To address this, a disaggregated system, which separates servers by function (CPU server, Accelerator server, Memory server) and connects them via CXL interface, has been proposed. 

Acceleration Memory Pool serves as a memory server that performs its own computations within this Disaggregated system solution. 

Increasing
System Efficiency
with Flexible
Memory Allocation

Larger amounts of data to process requires more memory,
which often leads to stranded memory within individual systems. 

By supporting memory pooling, which dynamically allocates
or releases memory resources as needed,
the occurrence of stranded memory is reduced,
thus enhancing system efficiency. 

Scaling Memory to Petabyte 

Levels by Connecting SSDs 

Rack-scale Memory Pool, which maximizes memory expansion by connecting SSDs,
can significantly reduce the total cost of ownership (TCO) for data centers. 

w/o SSD-backed

w/ SSD-backed 

Xcelerate Your
Intelligence


Company Registration Number : 710-81-02837

Address : 20, Pangyoyeok-ro 241beon-gil, Bundang-gu,
                 Seongnam-si, Gyeonggi-do, Republic of Korea

CEO : Jin Kim

© 2024 MetisX | All Rights Reserved

Xcelerate
Your Intelligence


Company Registration Number : 710-81-02837

Address : 20, Pangyoyeok-ro 241beon-gil, Bundang-gu,
                 Seongnam-si, Gyeonggi-do, Republic of Korea

CEO : Jin Kim

© 2024 MetisX | All Rights Reserved