This technical white paper discusses Dell EMC’s solution for 100,000 Exchange mailboxes with 70 GB size per mailbox, supporting up to 150 messages per day in a four-copy DAG. The solution uses the Dell EMC DSS 7000/7500 chassis/sleds for the Exchange mailbox server role and uses the internal storage of DSS7000 for storing the Exchange mailbox databases and transaction logs.
DSS 7000/DSS 7500 is a 4U 90x drives ultra-dense storage server. These systems consist of DSS 7000, which is the Chassis and 2x DSS 7500 sleds with a highly expandable memory and impressive I/O capabilities. This standard width, 4U system offers the lowest cost per GB in the industry today.
DSS 7000/DSS 7500 can readily handle data-intensive applications that require a large storage capacity such as email and archive solutions. It delivers the performance and availability required for mission-critical email and is a great hardware building block for super-large organizations with huge mailboxes.
The Dell EMC DSS 7000/DSS 7500 has the following key areas of differentiation:
In this solution, Dell EMC DSS 7000/DSS 7500 with 3.5-inch LFF drives are used as Mailbox Servers. The DSS 7000 / DSS 7500 combination provides SAS-based internal storage with RAID. The solution uses 90x 3.5-inch LFF 7.2KRPM NL-SAS disks across one DSS 7000 chassis with two DSS 7500 sleds in the chassis. The solution also uses four back-accessible 1.8-inch SSDs—two each across DSS 7500 sleds. The disk layout is listed below:
The solution has a 4-copy DAG Layout (Active/Active) with Exchange Servers distributed between two sites: Site A and Site B. Each server node has 42-RAID 0 LUNs hosting one active and three passive databases per LUN during normal operating conditions. Each of these databases hosts 23 users with 70 GB mailbox size. Thus, a single server can accommodate 475 mailboxes in normal runtime. Sixteen such servers in one DAG (8 each in Site A and Site B) provide Exchange Mailbox Services for 7,600 users. The solution can be scaled out for 100,000 users as outlined in this white paper by implementing additional DAGs.