Remote office/branch office technology trends in 2019 and beyond include more widespread use of software-defined WANs, hyper-converged infrastructure and improved backup processes.
Remote offices/branch offices (ROBOs) have always been a bit overlooked compared to the full data center or the cloud. But IT administrators can't migrate everything to the cloud; some workloads have issues with data locality, data sensitivity, or even performance and latency problems that don't make them good candidates for the cloud. Hypervisor vendors have realized this and have been quick to capitalize on this market.
The emergence of SD-WANs
One of the major remote office/branch office technology trends is software-defined WAN (SD-WAN). Hardware-defined WAN management devices have been around for a while, but virtualizing them offers a whole new world of opportunity; being software-defined makes these devices easier to deploy, manage and upgrade. More physical devices mean more complex patching and the need for engineers to visit the site.
In addition, hardware failure is less of an issue with SD-WANs. In a properly configured infrastructure, a device failure should only affect the SD-WAN link for a few seconds. Plus, there's one less hardware asset to manage, cable and troubleshoot with an SD-WAN.
These reasons are more than likely part of the thinking behind VMware's purchase of SD-WAN vendor VeloCloud. Other vendors, such as Cisco and Citrix, also offer SD-WANs.
SD-WAN appliances enhance the functionality of ROBO sites, and if admins already work with a vendor that offers the technology, then they have a better chance of retaining customization than they would with a third-party. SD-WANs also enable increased performance and intelligent routing, depending on where the packets must go.
For example, when a user at a ROBO site wants to connect to the internet for a lunchtime browse, the data doesn't have to go all the way back to the data center and then through its firewall, which would cause slow performance and use more bandwidth than required. Instead, the SD-WAN knows how to route the packets.
In some cases, however, security might dictate that all web traffic goes through the data center. SD-WANs can also use the ROBO site's infrastructure to do WAN compression on the fly, increasing data throughput without causing drops in speed.
How HCI supports ROBO sites
Remote office/branch office technology trends also include hyper-converged infrastructure (HCI), though HCI benefits ROBOs indirectly. Historically, ROBOs have had more single points of failure than they should, in part due to using single servers with directly attached storage. The choice between more personnel and extra hardware comes down to the lower cost option. HCI is expensive upfront, but staffing more people to manage old infrastructure ends up being more expensive.
HCI is nearly a fit-and-forget scenario. The complexity is hidden from users, and there is no need for external storage systems that consume power or on-site infrastructure. HCI shrinks it down to a smaller footprint that only has one pane of glass when it comes to management. Support also comes from only one vendor. This means that less maintenance is usually required, so using HCI reduces costs on that front, as well.
In addition, HCI offers better performance, with resiliency built in from the outset. Resiliency means that admins can carry on working even after a failure. Some businesses only have one server, which translates to lower costs, licensing and complexity. But in the event of a serious hardware failure -- the motherboard, for example -- having only a single host can be a bad thing because there is no resiliency. With a properly configured HCI server, a single-node failure would still allow the remaining nodes to work while the other is remediated.
The ROBO site needs at least three nodes for HCI to work because one node must act as a witness and decide which node is the master node. Vendors have listened and now it's quite common to be able to place the witness node in the cloud.
Putting a witness node in the cloud reduces the cost significantly; otherwise, it would be several thousand dollars for a third hardware node that admins probably wouldn't use much because the two other nodes would service requests. It ultimately depends on utilization but, most likely, it would be cheaper for a cloud-based witness disk than for a whole new hardware node.
Improved backups ease management
The way backup works also contributes to remote office/branch office technology trends. It used to be that a secretary was responsible for changing tapes and, at some point, a failure would occur.
Such failures are no longer acceptable, especially with the modern risk profile including hackers looking for data and malware authors. A single executable could render the entire office cryptolockered, with a backup being the only way forward.
Fortunately, hypervisors have eliminated the need for a secretary thanks to virtual backup and disaster recovery devices. Such virtual backup systems are centrally administered by technical staff, and there is no danger of tape loss. Overnight, these devices replicate data to the cloud or to a centralized backup infrastructure, keeping track of changes -- the Deltas -- so admins can upload the data quickly without large amounts of infrastructure to support it.
Once data is safe at the remote site, admins can conduct restores with minimal effort. Removing the human from the manual side means companies can do restores on demand, not just when the person who knows how to load a tape is present.