Minimum Requirements

Server (Node) Specifications
CPU	Intel Xeon x86 processor Ivy Bridge (v2) AMD Opteron x86 processor Opteron (62xx) Consistent processor family per cluster Note: A cluster is a minimum of 2 servers (nodes) with like architecture, multiple clusters can live within an environment Virtualization support (VT-x, AMD-V/AMD-SVM) Hyperthreading Support (Intel-HT, AMD-CMT, AMD-SMT) Four (4) 64 bit cores
Memory	A minimum of 64GB of ECC capable RAM Note: Memory is a variable that depends on Guest OS requirements, VergeOS, and Verge vSAN consumption One (1) GB of RAM will be allocated to caching for the Verge vSAN per one (1) TB of storage in each server (node) participating in storage Note: If storage space is increased, RAM consumption used by the vSAN will also increase by the one (1) GB per one (1) TB ratio
Storage	A RAID controller that supports JBOD mode (NO RAID) or Host Bus Adapter (HBA) certified to work with the latest LTS Linux Kernel One (1) High Endurance (≥ 3DWPD) NVME of 400GB or greater to be used as dedicated storage for environment metadata Note: This is only required by the first two servers (nodes) which act as the controllers of the environment Two (2) NVME or SATA/SAS SSD’s of the same size per server (node) for guest usage (more are recommended for performance purposes) Note: Two (2) HDD’s can be used in place of NVME or SATA/SAS SSD’s with an acknowledged performance impact Note: Individual HDD’s larger than 8TB are not recommended. If you have a special use case please discuss with Verge.io Sales.
Networking	Two (2) 10gbe NICs per server (node) are required for node communication One (1) NIC per server (node) for “external” communication (UI access and guest traffic leaving the environment) IPMI Port for remote management (RMM) of the physical servers Note: Intel, Mellanox, and Broadcom NICs are the most widely used and supported
GPU (Optional)	Processor capable of virtualization of I/O resources (IOMMU (Intel VT-d, AMD-Vi/VT)) NVIDIA GPU compatible with NVIDIA GRID drivers 13.2 or 14.0.

Cluster / Environment Specifications
Additional Considerations	For environments larger than 2 servers (nodes), a set of “core” switches must be used in the design for node communication Jumbo Frames are required for the “core” networking. Therefore any switch ports that participate in the “core” networking must have a MTU of 9216 or greater It is recommended that the switches be self contained, but currently in use equipment is acceptable providing vlans can be created Note: native vlans/pvids are required for node communications Example: Server (node)1 nic1 links to switch 1 port 1 with a pvid/native vlan of 900 with a mtu of 9216 Server (node)2 nic1 links to switch 1 port 2 with a pvid/native vlan of 900 with a mtu of 9216 Server (node)3 nic1 links to switch 1 port 3 with a pvid/native vlan of 900 with a mtu of 9216 Server (node)1 nic2 links to switch 2 port 1 with a pvid/native vlan of 901 with a mtu of 9216 Server (node)2 nic2 links to switch 2 port 2 with a pvid/native vlan of 901 with a mtu of 9216 Server (node)3 nic2 links to switch 2 port 3 with a pvid/native vlan of 901 with a mtu of 9216 For backup and/or DR capability TCP ports 14201 and 443 must be accessible by the source and destination environment

¶ Minimum Requirements - 2 Nodes