Cloud Network Application Cognition: Reality or fantasy

Before the advent of cloud technology, the idea that the network should differentiate between each application has aroused considerable interest in the application of cognitive applications because everyone has a unique need called applied cognitive or applied cognitive networks.

Cloud technology is particularly sensitive to application performance issues that are caused by network irregularities, along with potentially hosted applications on the network. The application of cognitive cloud computing requires a more complex application of cognitive methods, and the quality of the user experience will depend more on the availability of the applied energy itself.

In most cases, application awareness networks have been implemented through quality services (QoS) and understanding how changes in network performance affect the productivity of an enterprise user and the quality of consumer experience. Cloud computing may introduce complexity into the application cognitive network, but it will also lead us to explore potential solutions.

OpenStack Quantum: The beginning of cloud application cognition

OpenStack cloud software connects applications and clouds through a series of virtual interfaces, an interface for storage, an interface for computing services, and interfaces for network services. Virtual network Interface (quantum) is the first abstraction of network behavior, which is designed to allow applications to describe the desired content of network services. This makes it a useful starting point for discussion about cloud application awareness.

Currently, Quantumquantum defines a "virtual network" as a 2-tier structure such as a local area network (LAN) and a virtual local area network (VLAN). Extending the abstraction layer to layer 3 is no problem, however, quantum enables applications to run at the IP network level by combining features such as DNS and DHCP. Similarly, it is considered a quantum two-tier network, and the increased DHCP service enables providers to obtain IP addresses and gateway services for exiting the LAN and accessing IP domains such as the Internet. There are several recommendations regarding quantum expansion to include other network services, and it is likely that eventually quantum will fully support the 2nd and 3rd tier services.

The network structure defined by quantum is likely to be "virtual", which means that they are fragments of a complete network over Ethernet or IP connections. This is an important feature for cloud computing providers because users are tenants in a shared structure, but they must be isolated from each other for security and performance management reasons. This isolation must include all cloud resources, including networks. Quantum ' is important for applying cognitive networks, and this virtualization based isolation can also be applied to applications.

Software as a service and other higher-level cloud services are specific applications, and users will also tend to separate applications by further segmenting cloud computing. If an application uses an IP address through the API, it will be easily determined and its access traffic can theoretically be handled separately. Therefore, the virtual network in the cloud is the path to the network application cognition.

Indeed this is a method, but not a complete route. Virtual networks can be built in different ways, but they can all be categorized as overlays or embedding techniques. Superimposed virtual networks operate at the top of the network through standard application interfaces, and ultimately they will not affect the behavior of network devices or routes. Embedded technologies implemented by network nodes themselves, and because they can apply different, recognizable priorities and routing decisions to application traffic. Embedded virtual networks include VLANs and MPLS.

Applied cognitive networks: Advantages and disadvantages of SDN and virtual networks

There is another potential way to implement application awareness in cloud networks. A software-defined network (SDN) can be used to create embedded virtual networks that provide different QoS for each application as needed.

The problem is that as of September 2012 there was no uniform SDN definition and therefore no solid technical blueprint for creating a definition. SDN creates virtual networks for cloud data centers and IP cores, but these are often based on research prototypes and temporary software provided on traditional routers and switches. Future Web application awareness is likely to be increasingly based on SDN principles, but cloud providers are best to take their ideas to a small-scale test before submitting them to client applications.

Another problem with using cloud-based virtual networks as a way to apply cognitive approach is that the focus of network control is in the cloud computing datacenter, not the wide area networks (WANs). Applications can display their own local virtual network and user data, which is entered through a gateway, such as a router. But outside the path between the user and the gateway virtual network, the application of the traffic in the cloud era does not use the same technology, difficult to separate and manage, which means depth packet detection.

Security policies and access rights determine the level of access that users have to the application and the QoS attribute applications to the WAN. This suggests that one way to extend the application of knowledge from the cloud data center to the edge of the cloud is to integrate the application of performance management tools and firewall functionality with the branch or client side of the virtual network. This integration has been built into some products, with the development of cloud computing and cloud virtual networks, it is likely that edge devices will eventually become the "cloud edge devices" that include firewalls, SDN, and virtual network functions.

Another point that cloud service providers should consider is that extending application awareness to wide area networks may collide with the public policy goals of network neutrality. The Internet is the best service, including some countries in the United States, which have imposed regulations that limit the provider's access to Internet QoS, or limit vendor support for settlement premium processing among operators.

Enterprise Cloud services based on preconfigured IP virtual private networks (VPNs) are not bound by these regulations, so cloud services delivered via VPN are easier to develop quickly to application awareness. If the provider makes the edge of the cloud closer to the edge of the network, that is, closer to the user, then the business traverses the less external infrastructure to reach the application. Ultimately, QoS can be improved without violating network neutrality rules, and the extended cloud of this concept, in the long run, is the path to a truly applied cognitive network.