kentoh - Fotolia
SaaS and IaaS are fairly new delivery paradigms where the software resides in the cloud and is accessible from anywhere a user may be at any given time. Software consumers are no longer shackled to a corporate network via VPN connections. They're free to access what they need from wherever they may find themselves.
This environment certainly opens up a more flexible work paradigm, but it introduces several issues for network management teams that are trying to optimize application performance and guarantee a certain quality of experience.
When software resided on a corporate network, the variables in the quality of experience could be controlled easily -- or, at least, predictably. Quality of service mechanisms could prioritize traffic types, bandwidth could be tuned and the speed of the network could be increased through upgrades. Controlling access to the software was easy, and the access patterns were predictable, like when the network would be congested.
But, with the SaaS paradigm, many of the tools and methodologies we used on our networks have changed, and we must change with them.
SaaS optimization tools to the rescue
Many vendors specialize in tools to optimize the experience a user gets when accessing software in the cloud. Silver Peak's Unity EdgeConnect software-defined WAN platform is one such product. Many other enterprise SaaS optimization platforms are available, including products from Cisco, Riverbed and VMware, to name a few. But it might help to explore how these tools work through a single lens. This is not an endorsement of any product over another, but simply one example of the form.
Most SaaS optimization tools, like Silver Peak's Unity EdgeConnect, rely on a few tenets. DNS and traffic steering, for instance, help ensure users are always accessing the most efficient location of their SaaS software. The software is spread across multiple geographically diverse sites and load-balanced between even local points of presence. Consequently, the speed of one traffic path may be much higher than another because of latency, link saturation or simply a link could be out of service.
Additionally, the network team may want traffic steered over a particular path for cost reasons, preferring public internet transport rather than costlier MPLS or other private circuits. Most tools can classify packets by some mechanism, typically inspecting the first packet and making a rules-based determination as to the destination and optimal path. These decisions can be programmed manually. Or, in the case of more advanced tools, these decisions may be based on internal AI built into the optimization platform itself.
New tools provide network visibility
Network monitoring and visibility can be more challenging in an IaaS environment. SaaS applications are hard to optimize and troubleshoot if you can't see what's going on in the network. Several tools, however, address these issues, including LiveAction's LiveUX network monitoring and performance software. With these types of tools, you can get the same kind of visibility as tools built for traditional networks.
While networks are constantly changing, we should still have the same levels of visibility and optimized experience as we always have. Our tools and techniques need to adapt to the changing network. As long as we stick to the core tenets of network engineering, using new tool sets should not be an issue. We may even find that the tools have improved and provide more visibility and control than ever before.