A few random thoughts from ONS and Networking Field Day 15 this week:
- Intel is really, really, really pushing their 5 generation (5G) wireless network. Note this is not Gen5 fibre channel or 5G 802.11 networking. This is the successor to LTE and capable of pushing a ridiculous amount of data to a very small handset. This is one of those “sure thing” technologies that is going to have a huge impact on our networks. Carriers and service providers are already trying to cope with the client rates we have now. What happens when they are two or three times faster?
- PNDA has some huge potential for networking a data analytics. Their presentation had some of the most technical discussion during the event. They’re also the basis for a lot of other projects that are in the pipeline. Make sure you check them out. The project organizers suggest that you get started with the documentation and perhaps even help contribute some writing to get more people on board.
- VMware hosted a dinner for us that had some pretty luminary speakers like Bruce Davie and James Watters. They talked about the journey from traditional networking to a new paradigm filled with microservices and intelligence in the application layer. While I think this is the golden standard that everyone is looking toward for the future, I also think there is still quite a bit of technical debt to unpack before we can get there.
- Another fun thought kicking around: When we look at these new agile, paradigm shifting deployments, why are they always on new hardware? Would you see the similar improvement of existing processes on new hardware? What would these new processes look like on existing things? I think this one is worth investigating.
“When we look at these new agile, paradigm shifting deployments, why are they always on new hardware?”
That seems a “infrastructure on premises” perspective. Lots of new deployments are spun up on existing public IaaS, often because it allows agility while on premises infra typically is a constraint, not an enabler..
“Would you see the similar improvement of existing processes on new hardware?”
The point often isn’t to retain the existing process, since it’s often part of the company’s process debt (like technical debt, but process oriented.) The objective is to improve the end to end from business request to business result, which typically means changing the existing process, often fundamentally.
“What would these new processes look like on existing things?”
Really depends on how much technical debt is in the existing infra. Often the new process is in conflict with some constraint caused by the existing infra. Simple example, a firewall appliance at the edge causing tromboning from an app in a hypervisor when a firewall function could be under the management of the same hypervisor,
PNDA looks Great, and it is the first time I come across such complete and network oriented Tool or, rather, set of tools put together. Very ambitious. Is there any competitor in the market or do we have a “first” here?
Pingback: Short Take – The Present Future of the Net - Tech Field Day