The three components of data centers are compute, storage, and network.  Arranged in racks, towers, or silos, the servers (or compute) work with an array of hard or solid state drives (storage) and connect to the outside through rack-mounted switches (network).  The data center customer orders each component from selected hardware vendors and integrates them into a functional unit.  Vendors conform to open standards, so integration is as seamless as ever.  The units scale up and out to large sizes, typically to the extent of a large building. 

Because traditional data center architecture must converge to become smaller and simpler, rack servers get replaced with the blade-type, and hard drives get replaced with all flash arrays.  Software virtualization is often brought in.  To converge further, some hardware vendors collaborate, build all three components under a single platform, and sell to customer.  In a Hyper Converged Infrastructure (HCI), a single hardware vendor does the entire platform, or a single software vendor services a hardware-agnostic “Bring-Your-Own-Technology” platform.  Either way, several of these platforms work in tandem to scale in or out as needed. HCI is a purpose-built data center architecture that adapts easier than the others to the assigned purpose.  It simplifies operational and management tasks, promotes fast orchestration, and brings efficient scalability for the customer. 

Amphenol ICC’s solutions meet or exceed data center convergence expectations.  HCI’s non-volatile memory specs call for connectors like U.3 based out of SFF-TA-1001 and Mini Cool Edge based out of SFF-TA-1002 standards.  Enterprise-grade SAS starts from basic up to the 24 Gb/s fourth generation.  Standard PCI Express® is offered from basic to 5th gen.  Volatile system memory is served by connectors in all variants up to DDR5M.2 connectors for system cache storage are both PCIe and SATA compliant. 

Check out our data center offering at https://www.amphenol-icc.com/storage-server-system.html