06 Oct Remote Office/Branch Office
Remote Office and Branch Office (ROBO) IT can create some pretty tough situations for IT pros to conquer. Perhaps the most significant problem is one of scale. Many ROBO environments have the need to grow very large, but need to do so by remaining very small. Consider a fictional company that has 500 locations. This is a big company, so the overall aggregate technology needs of the organization will be significant. At the same time, though, each branch location in this company supports 20 employees and is also a sales outlet. The individual sizing needs of each branch are relatively modest.
Traditional ROBO Challenges
At first look, it might seem like a simple solution. Throw a couple of servers into each location and call it a day. Unfortunately, it’s not that easy. There’s a lot more to the scenario than meets the eye.
First and foremost, just a couple of servers may not meet the needs of the branch office. Each branch is probably sized a little differently, so some may be able to operate with just a couple of servers while others may need more substantial storage capabilities. You’ll probably want two or more servers just so you can have some level of availability. If one server fails, the other one can pick up the load. Getting high availability with just two servers, while solvable, isn’t always straightforward.
At the same time, you have to keep an eye on performance to make sure that poorly performing local applications don’t negatively impact the branch’s business. You can’t forget about data protection, either. If this was a single-site company, data protection would be relatively easy; you just back data up to a tape, disk, or a second location. But if you have many sites and some have slow Internet links, it can be tough to protect data in a way that makes sense. You don’t want to have local IT staff that needs to change tapes, or watch backup appliances. You also don’t want to have non-technical people trying to do this as a part of their jobs. It doesn’t always work out well.
Plus, there’s ongoing support. Stuff happens. You need to be able to keep every site operational. However, with each site you add (each with its own unique needs), the overall complexity level can become overwhelming. As complexity increases, efficiency decreases and it becomes more difficult to correct problems that might occur. Figure 1 provides a demonstrative overview of today’s data center. In many ROBOs, centralized IT delivers services to the remote sites from a centralized location over a WAN. By centralizing IT, the company eliminates the cost of skilled IT staff on site at remote sites and reduces the risk to business continuity since IT handles data protection. However, the major drawbacks are often poor application performance, scattered management, and difficulty correcting issues that arise in remote sites.
To summarize the challenges faced in ROBO environments:
- There is a need for a lot of decentralized systems to support individual branch offices, and there is often lack of a cohesive management platform.
- Bandwidth to branch offices can often be limited and may not be reliable. Most ROBO sites lack the full breadth of data center services (high-performance storage, WAN accelerators, etc.) enjoyed by headquarters and by single-site companies.
- Data generated at branch offices needs first-class citizen protection, but often can’t get it using legacy tools.
- Hardware at branch offices might run the gamut from just a server or two to a full cluster with a SAN, but most companies want to be able to have minimal hardware at branch locations when possible and need to be able to centrally manage solutions.
- There is a lack of technical personnel at most ROBO locations and companies don’t want to have to hire dedicated technical staff for each one.
Figure 1: Chaos is the “Norm” in Many ROBO Environments
Without some kind of change, the dystopian future for the ROBO will be so challenging that even Katniss Everdeen would call it quits and hang up her bow and arrow.
Transforming ROBO Operations with Hyperconverged Infrastructure
ROBO operations is an area in which the right hyperconverged infrastructure solution has the potential to completely transform the environment and how that environment is managed. The overall results can be lower costs, improved efficiency, and better overall disaster recovery capabilities.
So what does it take to achieve this ROBO utopia and how does hyperconverged infrastructure fit into the equation?
Keeping IT Simple
Hyperconverged infrastructure brings simplicity to chaotic IT organizations and nothing says “chaotic” like dozens of different sites running disparate hardware managed as individual entities. By moving to a common hyperconverged infrastructure platform, you instantly gain centralized administrative capabilities that encompass every site. Moreover, when it comes to hardware support, every site becomes a mirror of the others, thereby streamlining your support efforts. Such an architecture eliminates the need for dedicated technical staff at each branch.
The need to keep management simple cannot be overstated. Companies are no longer willing to scale IT staff at the same rate that they add sites and services, but they expect consistent performance. To solve this seeming paradox of intentions, IT has to look at the ROBO environment much more discerningly and deploy solutions. They need to choose solutions that unify management across all ROBO sites in a way that makes them appear as if they’re a single entity even while they support a dispersed organization.
Some sites need very little hardware while others need more. Some sites traditionally needed dedicated storage while others didn’t. It’s chaos. With the right hyperconverged infrastructure solution, you can have sites that operate on just one or two appliance-based nodes without having to compromise on storage capacity and performance. You simply deploy the two nodes, and they form their own cluster that imbues the branch with highly efficient storage capabilities that include comprehensive data deduplication and reduction. For larger sites, you simply add more nodes. No SAN is needed and all of the hardware across all of the sites is common, easy to support, and provides enterprise-level capabilities, even in a single-node or two-node cluster.
The data reduction features available in some hyperconverged infrastructure solutions mean that you don’t need to constantly add storage. With reduction, you get to cram more data into the same amount of overall capacity at the branch site. Reduction also has other benefits. Read on.
Comprehensive Data Protection
Data generated or managed at branch sites needs to be treated just like data generated at HQ. In many cases, the data at branch sites is even more important because it’s the information that’s created as the result of sales or other customer-facing efforts. With a hyperconverged infrastructure solution that has the ability to fully compress and deduplicate data and that can work with data in its reduced form, you can get data protection capabilities that allow you to replicate branch office data to other branches or to headquarters even over slow WAN connections. Better yet, you don’t need WAN accelerators to accomplish this feat. With the right solution, reduced data is transferred over the wire and, even then, only the blocks that don’t already exist at the target site are transferred, resulting in an incredibly efficient process. This kind of data protection infrastructure also eliminates the need for on-site staff to perform tasks such as changing tapes and increases the potential for successful recovery in the event of a disaster. In Figure 2, you see a nice, neat, and streamlined infrastructure.
Figure 2: Hyperconverged Infrastructure Tames the ROBO Chaos
As you’re deploying ROBO solutions using hyperconverged infrastructure, you need to take a look at how you want your ROBO sites configured. There are two typical models available:
- Hub and spoke (Figure 3) With this architecture, there is a centralized hub in the center and each remote site is at the end of a spoke. With this model, the various remote sites will generally talk to the hub spoke, but not often with each other. Backups and other data transfer operations will generally flow from the end of one of the spokes back the hub.
Figure 3: The Hub and Spoke ROBO Model
- Mesh. In a mesh environment, all of the sites can talk directly to the other sites in the mesh. Under this model, it’s possible to have individual sites back up to each other and the organization can, theoretically, operate without a centralized hub, although one of the nodes often acts in this capacity.
Figure 3: The Mesh-Based ROBO Model
As you’re deploying hyperconvergence throughout your organization, it’s important to ensure that the intended solution can easily support whichever deployment model you use, even if it happens to be a combination of the two. Most importantly, regardless of which model you use, you should be able to centrally manage everything and have the ability to implement data protection in whatever way makes the most sense for you. Finally, adding new sites – scaling the environment – should be a basic feature of the solution and not a complicated afterthought.
ROBO environments can be considered an “application” that requires some specialized attention. This is a perfect use case for hyperconvergence.