It’s hard to fathom 100 terabytes. I mean, it’s 100,000Gb but that is the amount of storage capacity recently purchased and installed at LiveBackup. It brings the entire storage at the primary Equinix data centre to 160 terabytes with a further 60 terabytes located at the Fujitsu IX replication data centre in Brisbane.
When we first established LiveBackup, we knew that we would have to architect a storage infrastructure that would be both highly reliable and almost infinitely scalable to any level of storage capacity. Additionally, with the criticality of the information that is stored, it was important for LiveBackup to operate an infrastructure that had no single point of failure.
We chose BlueArc as our storage infrastructure. They are of course not the only storage vendor on the market. Nettapp, IBM, HP and EMC all have great storage offers too. We have some international relationships that we held long discussions with and in the end we became comfortable with the BlueArc offering. We have been using the infrastructure for nearly 24 months and during that time the company was acquired by Hitachi.
Enterprise storage is different than standard server storage. Whilst the principles of hard disks still exist, other factors come into play. One of those is the number of read/write operations and the throughput that can be achieved. We purchased 72 2Tb NLSAS (near line, serial attached SCSI) hard disks that are enclosed in high-density disk trays.
One of the primary reasons for selecting enterprise storage is that it allows us to meet our service level agreements across the entire storage offering. Consistency in disks, controllers and a comprehensive software management platform make the management of the storage far easier.
With the Hitachi solution we could achieve 99.999% data availability and no single point of failure.
One of the primary differentiating factors of enterprise storage is the controller. It is effectively advanced software operating on dedicated hardware provided by the vendor. The controller is directing traffic in, out and within the storage. It is also communicating with the other controllers in the pool.
With that quantity of storage and enterprise storage capacity, it is near impossible to provide the necessary throughput over Ethernet and therefore we needed to install a switching architecture that is both redundant and could process the vast volumes of data that needs to be processed.
To complement the Hitachi storage, LiveBackup invested in the Brocade 300 series SAN (storage area network) switches. They feature a non-blocking architecture with as many as 24 ports concurrently active at 8 Gbps (full duplex) with no over-subscription—providing an overall bandwidth of 192 Gbps.
The entire aim of the LiveBackup service is to deliver a world-leading backup, storage and recovery service that meets the changing needs of business data protection requirements while maintaining a huge storage capacity. The Hitachi storage and Brocade switching relationship are key ingredients in the overall solution. I will continue to update with posts as new and significant changes occur in our solution.♦ End