Getting best value from a cloud requires a cloud-savvy approach to networking
Private cloud architectures promise breakthrough performance, scalability, and flexibility—all powerful selling points, and all directly applicable to the larger goal of creating the best possible infrastructure for fulfilling business strategies.
In many cases, though, a key element needed to fully realize the power of private clouds has gone undeveloped: networking, including network resources and network management.
Partly, this is because private clouds typically emerge as an evolutionary stage following virtualization and consolidation. That earlier process tends to focus largely on virtual servers, with a lesser focus on virtualized storage, and as this infrastructure is developed into a cloud, the original focus on servers and storage is retained.
It's certainly true that virtual servers and virtual storage are critical to cloud performance. Similarly critical, though, is the significance of the network and network resource management in various senses.
Consider how important bandwidth is to a cloud, for example. The number of virtual servers inside the cloud is in a constant state of flux; new servers are being created all the time. Each requires network bandwidth to fulfill its intended function. If that bandwidth is inadequate, virtual server performance will necessarily be inhibited, and so will be the performance of the applications and services it's responsible for. Yet ensuring that a cloud continually has all the necessary bandwidth is no easy task, partly because what is necessary changes from hour to hour.
Clouds also migrate virtual servers across hosts, in accordance with the terms of business policies. This implies many complexities from the standpoint of network resources. Among other concerns: Will a migrated server still have the necessary network resources to perform up to target levels, despite executing in a completely different physical environment? Can network requirements in such a case be handled automatically, without requiring direct oversight or manual configuration changes? Those are questions many IT teams will find difficult to answer.
In short, private clouds will require careful consideration of, and solutions pertaining to, network complexities to deliver their full business value.
IBM's cloud insights are a great resource and IBM cloud solutions leverage them fully
"IBM solutions, built on open standards, can interoperate easily both with each other and with your larger cloud architecture to orchestrate and automate cloud operations from one end of the service delivery chain to the other, and across the complete service lifecycle."
Fortunately for organizations interested in pursuing private cloud initiatives, IBM is in an excellent position to help—both with thought leadership and specific solutions. IBM's extensive history of successful customer engagement in both virtualization and cloud computing has, in fact, helped to guide the development of its product portfolio. And going forward, that portfolio will continue to be enhanced in networking capabilities.
For instance, consider the areas of network provisioning and network monitoring. Both are essential in a cloud context, just as they are for a traditional infrastructure, but in a cloud it's also essential that they be as automated as possible—and as interoperable as possible with other solutions that execute cloud operations.
The basic value proposition of the private cloud—cradle-to-grave service delivery with minimal action taken by IT—means that network issues, like other cloud functions, demand an automatic response. That, in turn, requires solutions that can interoperate with others in the cloud that perform logically related tasks (example: creating and provisioning virtual servers that will receive network bandwidth). All must be orchestrated seamlessly and automatically. The goal should be to automate not just services, but service management itself.
If you turn a microscope on these network tasks, you begin to see how cloud architectures, and the degree of automation they require, really do change the game for networking. Network provisioning is a good example. While resource allocation for computational tasks (such as processing or memory) is fairly mature, this is not usually the case for network tasks, such as provisioning. Virtual networking switches are typically the most advanced such technology available at most organizations, and they cannot handle fully-automated, cross-domain provisioning of the kind a private cloud will demand. Asking IT team members to handle such tasks manually will have the effect of defeating the basic value proposition of the cloud: that it automatically fulfill workloads and deliver services in an optimized way, despite changing conditions. Thus, smart solutions are needed to configure and provision network resources in a cloud—both physical and virtual—as well as integrate with the cloud platform as a whole, facilitating collaboration among both IT technologies and IT teams.
Network monitoring is similarly complex in a cloud context. Clouds are, by nature, incredibly dynamic—simultaneously delivering many services, each of which has fluctuating demand levels and resource requirements—and yet a cloud's overall health levels must continually remain inside confined limits, just as with any other architecture. Network monitoring solutions will be needed to assess cloud performance and resource levels by aggregating and analyzing data, weighing the results against predefined thresholds, and detecting patterns or changes that reflect emerging issues. When issues are detected, they'll have to be isolated rapidly to root causes for the fastest possible resolution and the smallest possible impact on the user experience. And, of course, all of these tasks should be as fully automated as possible.
A networking solution portfolio as smart as the cloud itself
If you're looking for solutions capable of addressing these and other related issues, you'll be glad to find that IBM can meet your needs. IBM solutions, built on open standards, can interoperate easily both with each other and with your larger cloud architecture to orchestrate and automate cloud operations from one end of the service delivery chain to the other, and across the complete service lifecycle.
In the area of network provisioning and configuration management, for instance, IBM Tivoli Netcool Configuration Manager is directly on point. This solution automates many key configuration tasks, controls network device access, and helps ensure that the terms of network policies are fulfilled despite the intrinsically dynamic nature of the cloud. Netcool Configuration Manager is also integrated with cloud provisioning and service orchestration solutions, delivering network resources in a fluid manner that fully realizes the cloud's deeply-virtualized design and capabilities for the highest performance via the most efficient resource allocation.
Similarly, IBM Tivoli Netcool OMNIbus and Network Manager delivers leading cloud network monitoring capabilities. It can discover all networks and network devices, aggregate information and events drawn from them, and thus provide cloud managers with a central point of visibility into the cloud's network performance and status levels. Thanks to event correlation capabilities, this solution can also help identify not just technical issues, but the root causes involved in each case, spurring a swift and effective resolution. Furthermore, the solution's capabilities aren't just limited to networking, but also address server, storage, and application resources as well—multiplying its value, while also minimizing the management complexity associated with those categories.
Finally, IBM's commitment to and leadership in the cloud computing space is also driving the ongoing development of cloud network features in these and other solutions. A case in point: IBM SmartCloud Foundation—a platform for incredibly rapid private cloud rollout and simplified management—will soon be enhanced, adding network management capabilities drawn from the IBM Tivoli Netcool family to its already impressive server provisioning and asset monitoring functionality.
Rethink IT, reinvent business, and recharge your professional network at IBM Pulse 2012
Join 8000+ peers, industry experts and IBM senior leaders for four days of top-notch education. Be inspired by guest speaker and Apple co-founder Steve Wozniak. And party the night away at Pulse Palooza, our annual networking event – featuring Grammy Award-winning Maroon 5!
Leverage and contribute to the collective wisdom around Tivoli