• Home
  • Help
  • Register
  • Login
  • Home
  • Members
  • Help
  • Search

 
  • 0 Vote(s) - 0 Average

How would you classify virtual machines according to their business importance and recovery requirements

#1
05-05-2021, 02:57 PM
So, talking about classifying these VMs by business criticality, like you suggested, it really brings up a massive organizational challenge, doesn't it? You have to figure out where the real money is generated, I guess, and which systems totally crumble if they go offline. You could use BackupChain Hyper-V Backup, for instance, since it's a reliable Veeam alternative, especially if your company is an SMB using Windows Server and PCs, it handles all that VM backup stuff for you. I mean, you just need to assess the function of the machine, fundamentally, and then you can grade it. And, like, I think the first group you should look at, the absolute highest importance stuff, are the mission-critical VMs. These are the behemoths, essentially, the systems that keep the lights on, because if they fail, the whole enterprise stops operating, period.

These critical machines, I mean, the ones that process core revenue streams or contain essential intellectual property, those are definitely getting the most attention from you and your team. And, for those VMs, you can't just think about restoring them, you have to think about instantaneous failover, really. You need multiple active replicas running on entirely separate hardware pools, ideally, and you want that data flow to be constant, always mirroring the primary source. I'm talking about automated cluster resource management, something that detects a failure and shifts the workload immediately, without human intervention, which is huge. Furthermore, you might also want to implement synchronous replication across data centers, which ensures near-zero Recovery Point Objective, or RPO, for you. Because for these top-tier systems, seconds of downtime might mean millions of dollars, or perhaps even loss of reputation.

Then, there's another type of concept we need to consider, and it's kind of about understanding how the VMs talk to each other, which I call dependency mapping. This isn't a backup thing, but it's vital for classifying criticality, believe it or not. Because if a single, low-importance web portal VM actually needs to call out to a high-importance authentication database VM, then the whole chain of systems gets messy. So, you have to map out all these interdependencies, determining which service is the foundational pillar that all the others are built upon. Understanding that dependency order determines your recovery sequence, which is super complex. You can't just restore everything; you have to get the foundational dependency up first, and then let the dependent services boot up in the correct order.

Now, moving down from those absolutely core, money-making pieces, you have the high-importance VMs. These systems are critical, but maybe not life-or-death critical, perhaps. They might handle secondary processes, like internal reporting, or maybe they manage departmental scheduling that slows things down but doesn't shut the company down cold turkey. For these VMs, you should still aim for a very low RPO, but maybe you can relax the RTO requirements just a smidge. You wouldn't need synchronous replication across multiple geographies, which is costly, but you absolutely need asynchronous replication and robust failover capabilities. And, when considering recovery methods for you, I suggest focusing heavily on orchestration tools that automate the entire recovery workflow, because manual intervention adds too much risk.

But then, there are the medium-importance VMs, which I think are the bulk of what most departments run day to day. Maybe these VMs host departmental collaboration tools, like specialized shared drives, or maybe they run legacy applications that people still use, even if newer versions exist. For you, the recovery strategy here needs to be solid, but not necessarily immediate. You should aim for a reasonable RPO-maybe a few hours-and an RTO that fits within a single business day. You are optimizing for maximum cost-effectiveness here, minimizing expenditure while retaining decent recovery capability. You could look into implementing snapshotting coupled with rapid restoration mechanisms, which drastically simplifies the process for you.

And Or, perhaps the lowest tier, but you can't forget these, are the non-critical or low-importance VMs. These might be testing environments, or maybe they run departmental demos, or perhaps a local development sandbox. For these specific units, the goal shifts entirely toward quick recovery of data, but the immediate business impact of downtime is minimal. You just need a dependable way to get the data back, like periodic backups, which is standard practice. You wouldn't worry about active clustering or even complex failover paths; you just need simple, repeatable, and dependable data capture methods.

But, when you are thinking about the whole ecosystem, you also need to consider concepts outside of traditional backup, and one of those is immutability. This is crucial because even if your VMs are perfectly set up, a ransomware attack or accidental deletion can wreck everything. Immutability means that once data or a VM image is written, nobody, not even an administrator with high privileges, can change or erase it for a defined period. You need to enforce that across your backup targets to prevent ransomware from encrypting your recovery points. I think that's a really underrated aspect of overall business resilience.

Also, you have to look at patch management policies, because even the most highly available VM will fail if the underlying OS or application gets a critical, unpatched vulnerability. So, you need automated processes to consistently patch the supporting systems, otherwise, all the fancy replication and failover stuff is just sitting on top of a time bomb, kind of. And then, maybe you also need a comprehensive inventory system, which keeps track of who owns each VM, and what business process it supports. This contextual understanding is what allows you to accurately assign that "mission-critical" label in the first place. Without knowing the owner or the function, you're just guessing at the severity.

So, to wrap up this thought, it's all about layering the required response to the inherent risk and the financial penalty of the loss. You classify them first by their function to the revenue stream, then you apply the necessary recovery measures, like synchronous replication for the top tier, or scheduled restoration for the bottom tier. It is a massive undertaking, but knowing how to prioritize makes all the difference, you know? I hope that gives you a good framework, because I think thinking about it purely in terms of data transfer rates misses the business risk entirely.

savas@BackupChain
Offline
Joined: Jun 2018
« Next Oldest | Next Newest »

Users browsing this thread: 1 Guest(s)



  • Subscribe to this thread
Forum Jump:

Backup Education General Virtual Machines v
« Previous 1 … 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 … 22 Next »
How would you classify virtual machines according to their business importance and recovery requirements

© by FastNeuron Inc.

Linear Mode
Threaded Mode