Market exuberance for Office 365 has inspired business mandates to adopt the cloud-hosted collaboration and productivity suite without regards to the underlying chaos. While multi-location organizations are virtualizing, operating models haven’t necessarily changed. This partial transformation that excludes automation and simplification of the network puts Office 365 deployments (and other software-as-a-service offerings) in danger of failing.
Take banking as an example — IT must design an architecture that can support the varying needs of thousands of branch locations and multiple divisions. Where the consumers at the branches might require online banking and informational video streaming, the trading division (a much smaller subset of the organization that typically makes the bulk of the revenue) will need their applications prioritized.
To achieve this, IT may start to build an overlay network, relaying this blend of application services across the plain old data network and continuing to rely on backhauling. In these situations, when a customer wants to complete a large money transfer or open an account, the data packets of this application may cross an entire continent before going back to a cloud provisioning center that is next-door. Unnecessary travel isn’t just inefficient in cost, bandwidth, and response time — it’s where chaos is born, and impacts application performance and end-user response time through increased latency during transfer.
Chaos Takes Many Forms, But Always Creates the Same Challenges
Chaos isn’t limited to backhauling or using generic devices to route traffic. Instead, it can include:
Stacking physical hardware appliances with different network functions in an attempt to enforce business policies at the branch edge of the network
Running a separate infrastructure for important business functions
Using connectivity that was chosen based on price, rather than performance and agility
A network platform lacking orchestration or automation capabilities
Regardless of the exact network configuration, chaos consistently creates the following challenges when attempting to deploy Office 365:
Poor Performance and End-User Experience - Long and inefficient traffic patterns add latency and diminish the user experience. Additionally, without any orchestration or automation there is no distinction between transactional or bulk application types for appropriate performance and availability controls.
Increased Costs - While on the surface, replacing legacy desktop applications appears to be a cost reduction, deploying Office 365 on existing architecture without fine-grained application policy and bandwidth controls leads to increasing connectivity costs to manage the application demands.
Diminished Visibility - The complexity of chaos can lead to confusion or clouding of current bandwidth usage, unimportant traffic traveling across an expensive MPLS connection, and over-provisioning to combat poor performance that is unrelated to bandwidth.
[previously posted on APM Digest, https://bit.ly/2QqZ5Yf]