These three examples highlight some decision factors for using a public, private, or hybrid cloud deployment. Table 17-3
summarizes some of these decision factors.
Table 17-3 Public versus Private Cloud
Most decisions about whether to use a public or private cloud will be heavily influenced by culture rather than costs and data ownership.
The Impact of Cloud Computing on Network
Engineering
Some network engineers believe cloud computing will eventually end the practice of network engineering. While cloud computing impacts the scale and scope of network engineering, the real-world results are more nuanced than destroying or displacing an entire career field.
Early cloud services emphasized turning information processing into an easily accessed commodity. For instance, you might hear statements like: “The network should be like water: it should deliver packets with no muss or fuss…it should just work.” Cloud has indeed helped commodify networks, but this has not been, and will probably not be, the disaster many pundits prophecy.
Commodification has occurred in forwarding engines (FEs or ASICs), and hardware has become less important in networking. Network hardware design is mainly well- understood: routers and switches are sold based on their port speeds, overall bandwidth, buffer depth, and the balance between processing speed and features.
While individual routers and switches may move closer to being commodities, the network itself is not. If a router, switch, or another network middlebox is like an individual rail car, the entire network is closer to a complete transportation system.
Networks are becoming more complex, and moving data is becoming more critical to businesses daily.
Individual network devices are treated less like pets and more like cattle…but this is not bad. Instead of focusing on individual devices, network engineers can think about how businesses and applications use the network. Moving up the stack adds value.
Automation
Moving from pets to cattle also enables network automation.
Automation (done correctly) can reduce network outages, increase security, and make the network more consumable for developers and application owners.
There is a second-order effect from the large-scale environments built by cloud providers. These large-scale networks require automation to operate efficiently. The skills and lessons created by developing these tools inevitably “leak down” into smaller networks over time, improving every organization’s ability to build and deploy private clouds efficiently.
Cloud-Based Network Management
Public cloud solutions are also being directly adapted to managing and troubleshooting networks. Many private cloud systems rely on a public cloud component to monitor the network, predict outages, and optimize resources.
For instance, Cisco Systems offers the Cisco Cloud Network Controller for managing on-premises networks. These tools allow network engineers to shift their focus to design and application support rather than monitoring and managing the network.
Remote and Hybrid Work
Before cloud computing, accessing corporate resources while traveling to a conference, taking a cruise, going to a coffee shop, or even working in your backyard was—at its easiest—difficult.
In the early days of network engineering, accessing the corporate network involved racks full of computers acting as hosts for dial-in clients, wired to racks of MODEMs. Remote access was slow, clunky, and error-prone.
Cloud computing caused a complete revolution in remote access.
Any applications requiring widespread remote access can use the public cloud native or hybrid deployment models to reach tens of thousands of users no matter where they are. Access to applications in the private cloud is still easier because of Security as a Service options and the zero trust models created in the cloud computing world.
In the cloud era, it is much easier to work from anywhere all the time (fully remote) or part of the time (hybrid work) and to support users who prefer fully remote or hybrid work.
Employment Shifts
Even if the worst possible predictions were accurate, and every company moved their data into some public cloud, the world would still need network engineering skills.
First, vendors and public cloud operators need people with network engineering expertise to build their products.
Second, every modern application developed runs over a network. Application developers who understand how networks work and how to make an application run effectively and efficiently over a network are not very common. Security engineers with network engineering skills will always be more successful.
Even in the worst-case scenario, network engineering skills translate to many other kinds of work in information technology.