AXIS Recommended Equipment

This is a quick summary of the basic AXIS Hardware requirements. For Software and System Requirements please visit Knowledge Base article 1927. This page is intended to provide a general guideline for provisioning the elements within your AXIS ecosystem (AXIS EnterpriseLink Solo, Server, and GridLink). For more specific guidance or additional details, we encourage you to contact client support so that we can address your specific case.

Monitors

Recommended Monitor Specs

Processors

AXIS makes strong demands on the floating-point capabilities of your processor, so we will always advise you to buy the best floating-point performance you can afford. The leading factors in determining the processors' performance, as it relates to establishing our recommendations, for AXIS computational workloads would be:

• Floating Point Calculation per Processor Cycle (FLOP/s)
• L2 / L3 Cache Configurations
• I/O Bus specification

Servers

Intel's Processors leveraging the Skylake architecture, marketed as the Intel Xeon "Bronze", "Silver", "Gold", and "Platinum", are currently our top recommendations.

Recommended Models

AXIS Server Tech Specs »

AXIS Role / Function Sizing Considerations

EnterpriseLink

The EnterpriseLink Server/AWS EC2/Azure Instance specifications would be dependent upon the number of concurrent users expected to be using the service; as a general rule of thumb, you should size it in accordance with any Terminal Service Server (Citrix or Microsoft Remote Desktop), based on having the limited Application Profile for EnterpriseLink exclusive use. In sizing the EnterpriseLink Server, the Primary Considerations would be: RAM, Network Latency, and GPU.

Our sizing considerations for RAM and Storage on the EnterpriseLink Host(s) follow this formula:

• RAM: 4GB per Core (eg. for 24 Cores, the minimum requirement would be 96GB)
• Hard Drive: SSD, Azure Premium SSD, AWS NVMe SSD Preferred

Formula for Core Capacity Estimates to Size AXIS EnterpriseLink Hosts:

• EnterpriseLink Server: (Round (No. of User / No. of Cores per Host)) = No. of Hosts
  For example: (Round (50 Users / 24 Core Server)) = 3 EnterpriseLink Hosts

GridLink

The GridLink Server/AWS EC2/Azure Instance specifications would be dependent on whether you planned to use AWS Auto Scale Groups / Azure Virtual Machine Scale Sets or Permanent Grid Infrastructure; with regards to sizing for these instances, that would be determined by the types of batch jobs your Actuarial users run, baselining the current workload, etc.; in sizing the GridLink hosts, the Primary Considerations would be: Floating Point Calculation, L2 / L3 Cache, Disk Access Speed.

Our sizing considerations for RAM and Storage on the GridLink Compute Nodes follow this formula:

• RAM: 4GB for 32-bit AXIS and 8GB for 64-bit AXIS per Core (eg. for 24 Cores, the minimum requirement would be 96GB for 32-bit and 192 for 64-bit).
• Hard Drive: SSD, Azure Premium SSD, AWS NVMe SSD Preferred. We recommend a minimum of 3 x SSD drives in raid 0. If a disk failure occurs on a helper server, the work is automatically redistributed to other nodes, so disk redundancy on helper servers is optional. On master servers, however, it is recommended to provision disks in Raid 10 so that disk failure does not cause running jobs to fail. You may also provide separate drives (RAID 1) for the O/S on both Master and Helpers servers although this is not a GridLink requirement or a GGY recommendation. You can save money and drive bays by introducing an O/S partition of say 100GB on the main RAID array.
• We recommend that this storage be provided via local disks. For master servers only, you can alternately provision the drives via a direct fiber connection to a suitable SAN. The SAN, however, needs to be able to sustain performance equivalent to the local disk configurations noted above.
• Formula for Minimum Storage Capacity Requirements for AXIS GridLink Hosts:
  • GridLink Server Storage: (No. of Cores * Max Dataset Size * 1.25) = Minimum Available Storage (excluding OS Volume)
  • For example: 24 Cores * 50GB * 1.25 = 1.5 TB of Storage

Due to the extreme workload placed on the drive systems in multi-processor servers and the availability of higher clock rates in smaller servers, we strongly recommend 2-processor servers over 4-processor or 8-processor servers slated for use as part of a GridLink farm. Racks of dual-processor servers are a popular choice for scaling AXIS under Distributed Processing and Grid Computing. Where space is at a premium, Blade servers are an excellent option.

Please see a matrix of sample recommended models from the major vendors (IBM, HP, Dell). This is by no means an exhaustive or a restricted listing, just a representative selection of models, leveraging one of our recommended processors.

Recommended Models

Desktops/Laptops

Intel recently announced their latest processors based on the "Kaby Lake" architecture, but these have not hit the mainstream yet. So our recommendations for Intel-based workstations, desktops, and notebooks are based on the "Skylake" architecture, branded as "Core i7" and "Core i9" processors.

AMD recently released second-generation Threadripper processors are based on the ZEN+ microarchitecture, featuring frequency and memory improvements, as compared to their previous family of processors. The biggest change second-generation brought is introduction of higher core count models. 2900-series doubled the core count to as much as 32 cores and introduced a new automatic overclocking feature called Precision Boost Overdrive. As of this writing, the 2900-series processor family has not been adopted by the major workstation, desktop, or notebook manufacturers (eg. HP, Dell, Lenovo); this is reflected in our sample workstation/desktop/laptop listing.

Recommended Models

Please see a matrix of sample recommended models from the major vendors (Lenovo, HP, Dell). This is by no means an exhaustive or a restricted listing, just a representative selection of models, leveraging one of our recommended processors.

AXIS Laptop and Desktop Tech Specs »

Recommended Models