Maintaining network integrity

0
1406 views

Considering that blockchain is distributed technology the network is only as strong and secure as its weakest link. What is the best way to ensure that hardware glitches are dealt with independently while allowing local issues to be addressed without taking down the entire blockchain network to maintain its security and hardware integrity?  

Blockchain
Security
Maintenance
Integrity
Network Administration
Network Optimization
Healthcare
Healthcare Information Technology
Bryan Russell
68 months ago

4 answers

1

The best way is to have redundant devices set up with redundant cabling.

For example, say you have a device providing you bit mining. Having cabling going to two redundant switches, an A side and a B side switch is optimal. That way if the A side switch fails, the B side should sense the problem, and take over with minimal loss to your device.

Also, making sure you have two different networks is optimal as well. For example...A switch could be on the AT&T network, while B side network should be on Verizon, as an example. The reason for this is if one network experiences a loss or is knocked down due to a fiber cut or whatever reason, your traffic should still be functional on Verizon as that should be a separate network.

Finally, having one center as a production center and a 2nd center with all redundant equipment as your backup center would provide for optimal protection. If say a fire happened at your main data center facility, the backup data center could handle and pickup traffic in a short time, with minimal loss.

Joseph Zeug
68 months ago
These are all necessary and basic for redundancy. Tablestakes indeed. - Ron 68 months ago
1

Assuming I understand your question correctly, I would suggest first you only pursue blockchain through a virtualized network. If it is in a virtual network, then use micro-segmentation to isolate the technology on the network so that it will only associate with tools in the segment and communicate with systems you designate. if a problem is identified either within or outside the segment, you can manage its access.

When you are using blockchain, I would advise you have multiple service providers for the protection of your services and you negotiate with them service level agreements that match your operational target. You can put in load balance equipment that can handle the automatic switching should something arise. Just be sure the size of the connection can handle the load if you have to fail to one provider.

As Joseph said earlier, you should always have redundancy. A second center with independent feeds is an expense, but it is part of keeping your business connected.

Lynn Gibson
68 months ago
Adding virtualization on top of underlying redundant physical networking is key.. sometimes virtualization doesn't help becasue the underlying infrastructure is not redundant. - Ron 68 months ago
1

The question suggests an incorrect understanding of blockchains.

"the network is only as strong and secure as its weakest link"

This is a description of a brittle system. It seems to imply node failure ripples through the network and becomes a more widespread failure. This is not the case. These networks are indifferent to failure of specific nodes or even outright attack (in most implementations).

Try it this way:

"The network is as strong as the network tolerance for failed and malicious nodes and the eventual consensus assurances provided by the protocol itself."

All nodes will eventually reach consensus regarding the shared state at all given checkpoints, usually called a block height.

These are highly resilient systems that carry on in the face of adversity. I think it may help to separate concerns and also to recognize that there are many implementations of the blockchain principle, just as there are many implementations of databases.

Generally speaking, organization access to the blockchain doesn't require a fault-tolerant node, but may call for failover access to any node.

Disclaimer: There many varied blockchain implementations and each takes a different approach to consensus and topology. This question would be easier to answer with precision in the context of a specific implementation.

Hope it helps.

Rob Hitchens
68 months ago
Well said Rob.. this is a blockchain "network" response vs. the above comments are classic network responses.. unfortunately "network" means different things.. but in BC, only need to ensure node failover mechanisms work for BC network reliability. - Ron 68 months ago
Thanks, Ron. Agree completely. - Rob 67 months ago
1

If you view the BC "network" as an app running on top of the physical/virtual network underneath, than it is most important that the BC be independent of the underlying infrastructure. Node detect / fail-over, etc. mechanisms in BC designs help that.. quicker convergence of nodes can help as well.. adding resiliency at each layer and separation of the layers helps with overall system stability.

Ron da Silva
68 months ago

Have some input?