Your mobile or branch office worker will access the Internet, the WAN-optimization-as-a-service will then accelerate that user as he traverses the Internet before handing it off to Microsoft's cloud data center. Here, only the "first mile" and "last mile" (the short connection from the user to the Internet and the Internet to the data center) are left un-optimized; the majority of the connection is accelerated by the cloud-based WAN optimization service.
Why it matters: Both the virtual and cloud-based form factors help lower the cost of WAN optimization by giving more flexible deployment options. You can deploy dedicated appliances when user and performance needs require it, while supplementing them with virtual and cloud form factors when economics or application types warrant it.
WAN optimization capability #2: Quality of Service
Network engineers should also consider quality of service (QoS) and decide which applications and services deserve specific amounts of bandwidth. You're prioritizing traffic to guarantee a high level of performance for specific mission-critical applications. Other non-essential (usually personal or recreational in nature) applications get a lower priority.
Traditionally, video and voice have topped the priority list, but as companies change the way they do business they should prioritize cloud-based services too, like Office 365 and Salesforce.com. You will need to decide which Internet-based applications and services drive the most value and assign those higher bandwidth. Turn to application monitoring tools to sort through all the data to make the right choices for your network and business.
Why it matters: QoS techniques help avoid unnecessary bandwidth upgrades by ensuring business-critical applications are not competing for valuable network resources. Moreover, QoS gives IT teams a way of aligning network and applications with business requirements, a critical component in providing a superior user experience.
WAN optimization capability #3: Path selection
Complementary to quality of service is the concept of path selection, which is essentially quality of path. Like QoS, path selection allows you to prioritize which traffic traverses which network link. You can send traffic out on a specific link based on performance, cost, security or availability criteria.
For example, traditionally we would backhaul Internet traffic across a private network, then send it off to the Internet, then route it back across the private network to the end user. Since these secure links are often quite expensive, the economics of this approach are poor. Plus, the user experience is terrible. Because so much Internet traffic is now mission-critical applications, this double trip of Internet traffic just clogs the central network when a direct-to-net link would be faster and more cost effective.
Let's examine another practical application of path selection. Consider a law firm paying for expensive 15Mbps MPLS connections from all of its offices back to its data centers. A WAN optimization solution with path selection found that a large amount of branch traffic is to the Internet for applications such as Office 365 or cloud-based document management. The firm realized if they cut those connections to 8Mbps, with the money saved they could then get 45Mpbs direct Internet connections at each location. Now they get 53Mbps aggregate bandwidth at zero additional cost. Equally important, the direct-to-Internet user experience can be made faster and more reliable with the deployment and QoS options detailed above.
Sign up for Computerworld eNewsletters.