The recent Microsoft outage was a jolting reminder of what happens when a technology that has a track record of being reliable and consistent breaks. Flights were grounded, hospitals went into a scramble, retail stopped working, the world came to a screeching halt.
When network issues strike, the consequences are just as disastrous. The network works very well on good days, and that is most days, leading to the expectation that it will always work the same. But potential problems can manifest any day, degrading user experience across the board.
What No One Sees
Outages and disruptions knocking systems out in the middle of busy workdays have become more and more frequent in the news. To help understand what causes these impactful incidents, one needs to look behind the scenes.
The way users perceive connectivity is misleadingly simple. It is not just a device and a resource communicating over the network. It’s a little more complicated than that.
Connectivity encompasses a breadth of things that include people, devices, network services, infrastructures, and applications.
“The complexity of how a user accesses an application or a resource today is very complex,” remarks Marko Tisler, group product manager at Cisco ThousandEyes. “It breaks in so many fabulous ways.”
Tisler explains that when a user connects to a resource, the request traverses multiple networks via various paths, going through several hops and handshakes, before it is able to consume the service. Interruptions can happen at any point in this complex and dynamic path.
A problem can originate in the client device, itself for example. Or an authentication error or service problem in the middle mile can break or slowdown the communication. App downtime too is a leading cause of poor user experience.
“Users don’t see any of these complexities that lie between the two points of connectivity,” he says.
DXA, a Three-Legged Stool
On the provider’s side, it’s a complex diagram of IoTs and ISPs, gateways and VPNs, cloud and datacenters networks. Some of these networks they control, but the ones outside their perimeters, are not in their control.
In an ideal world, a provider should have visibility of everything that’s happening in any of these networks at any given time of the day, like their own networks. But in reality, broken observability of intermediate networks and nodes make service assurance hard.
More than one network visibility solution has failed to provide fine-grained visibility across the nodes because the key principles of digital experience assurance were not identified.
Consider a three-legged stool. If DXA is this stool, the three legs it stands on are visibility of the end-to-end path, AI-augmented intelligence, and closed-loop operations.
If a solution has these three capabilities, digital service assurance can be rightfully expected.
More Data, Clearer Visibility
Since its inception, ThousandEyes has sought to understand all the interesting ways the network breaks.
Cisco, for years, has done visibility by means of synthetic testing where different favors of agents are leveraged to perform tests on certain resources. There are two approaches to this – an “outside looking in” methodology where cloud-hosted agents test resources hosted in client environments.
“We have multiple points of presence across the globe, and our customers can use those as proxy metric of find out how the connectivity looks, their applications, from customers that are distributed across the globe,” he informs.
The other is the “inside looking out” approach where agents hosted in customer environments are used to test SaaS services used by employees.
But the “end-to-end visibility from the network interface of the user device to where the application is hosted – be it datacenter, physical server or a hyperscaler environment,” that is the ultimate goal demands more datapoints.
Besides synthetics, “we need to consume other datapoints to be able to be more precise, to give more context, and context is what we really are after,” Tisler points out. It is the cornerstone for establishing root cause.
New DXA Innovations from the House of ThousandEyes
At Tech Field Day Extra at Cisco Live US 2024, ThousandEyes debuted three new solutions – Cloud Insights, Traffic Insights, and Shared Cross-Domain Context.
The trifecta of these solutions will give customers highly specific context and the right datapoints to expedite troubleshooting and reduce mean time to resolution (MTTR). This, they say, will assure top-notch digital experience end-to-end over any network.
ThousandEyes Cloud Insights is a cloud topology visualization tool designed to provide enhanced visibility into public clouds.
Cloud Insights creates topological mappings of customers’ cloud infrastructures and environments. It can auto-discover cloud provider resources, drill into service dependencies individually for that extended visibility, and show traffic patterns and characteristics in the cloud.
Additionally, it can track down infrastructure changes, and correlate them to the user experience. This tells operators how a change affects the digital experience for a group of users.
Knowing what’s going on “beyond the front door of the hyperscalers” enables operators to identify issues faster, triage quickly and perform a speedy root cause analysis.
ThousandEyes Traffic Insights is a traffic analytics system. Its key capability is unifying views of the internal and external network conditions.
Traffic Insights gathers flow data from the routers, and surface contexts allowing operators to cross-relate them with synthetics.
“We’re seeing the demand for this from two different perspectives,” Tisler told.
Synthetic tests point to a problem, but they do not offer any insights into experience degradation.
“In some cases, there’s nobody on-site using an application at a point in time which probably means it’s a less important issue compared to the same problem manifesting itself when thousands of users are trying to access that resource from that same site.”
ThousandEyes can determine the urgency of an issue based on the number of users trying to access the resource.
Traffic Insights’ advanced analytics also prove helpful for capacity planning. The holistic view of usage makes planning precise while providing basis for future forecasts.
Shared Cross-Domain Context is a joint implementation of ThousandEyes and Meraki. When a problem occurs, the team jumps in to find the proverbial needles in the haystack.
Shared Cross-Domain Context enables a bidirectional exchange of data between Meraki and ThousandEyes. This gives it the ability to correlate data from across the digital supply chain. With information about infrastructure health, client health and app health in one place, operators can promptly isolate service degradations and dig into the root cause.
For more, watch Cisco ThousandEyes’ presentation from Tech Field Day Extra at Cisco Live US 2024 at the Tech Field Day website.
