Advanced Computing in the Age of AI | Thursday, March 28, 2024

The Hybrid Network: Monitoring and Management Beyond Your Own Four Walls 

saicle/Shutterstock

More organizations are migrating applications to the cloud, along with more basic infrastructure, such as storage and application development platforms. A recent survey of IT professionals found that 92 percent say adopting cloud technologies is important to their organizations’ long-term business success, and 27 percent call it extremely important.

As a result, today’s IT professionals are increasingly judged by their ability to manage cloud services. And while the cloud aims to deliver greater operational simplicity, network administrators remain wary. The IT professional’s mandate to manage cloud services brings with it the inherent challenge of being held accountable for networks the company doesn’t even own — i.e., those of the cloud provider — which are beyond the reach of traditional enterprise network management tools.

The broad adoption of cloud services is transforming the nature of the corporate network, forcing it beyond the four walls of an organization and creating a complex management scenario. On a good day, cloud services may seem to make everyone’s life easier. What does hybrid network troubleshooting look like on a bad day?

The Hybrid IT Network Manager’s Dilemma

Picture this: a network administrator is monitoring and optimizing the corporate network when the sales team flags slow response times and frequent connection time-outs when connecting to Salesforce.

The company’s Salesforce representative insists there are no issues on their end, yet the network administrator’s enterprise monitoring dashboard shows a healthy enterprise network. After a deeper investigation, the network administrator determines the problem must be related to the traffic moving between the enterprise network and the cloud provider.

Unfortunately, understanding what happens to the internet traffic once it leaves the company’s network is a convoluted, often unachievable task with traditional network monitoring tools. Piecing together network traffic from the sales desk to the cloud service using basic utilities, such as Ping and Traceroute (which tracks the series of hops network packets take to reach a given network address), are often foiled by company firewalls that block outbound Traceroute connections. Even probes initiated outside the firewall are typically insufficient. The result is a partial and confusing picture.

Destiny Bertucci

Destiny Bertucci

In the midst of troubleshooting efforts, the network administrator is informed everything has “fixed itself” and end users are back online. Despite the sudden resolution, the network administrator is still unsure of where the problem originated, and how to fix it in the future. (Side note: It is always a best practice to reference the detection-prevention-analysis-response cycle to create strategic and tactical ways to address any network issues now and in the future.)

So how does network management strategy need to change in order to meet these new challenges?

Network Management 2.0  

While network administrators have made enterprise networks more manageable with visual tools – such as interactive network maps that can be supplemented with probes for monitoring databases, application servers and other key elements of the network architecture – until recently that level of surveillance has ended at the boundaries of the enterprise network. Cloud providers may offer service status dashboards where major outages are announced, but a green light doesn’t mean an underlying problem doesn’t exist.

To effectively manage borderless networks, IT professionals need more than just a dashboard—they need complete visualization of the entire network, both on-premises and in the cloud, and basic tools like Traceroute and DNS lookups are becoming obsolete in the face of cloud and hybrid IT.

The key is to break down the invisible wall between the physical data center and the cloud to realize greater visibility, meaning network managers should be able to see the pathways of applications and the quality of the service. Visual path monitoring allows IT to regain much of the authority lost in the move to hybrid IT. It enables not just the simplified detection of issues in internal networks, but also extends troubleshooting through the internet and into service provider’s networks.

This is made possible with modern network path monitoring tools modeled after native protocols used by cloud applications, rather than a specialized diagnostics protocol like Traceroute, and can simulate application-specific traffic that passes through firewalls in the same way as user traffic. For example, these tools can let network administrators see that a request to search the records within a web-based application is reaching the load balancer at the border of the cloud provider’s network, but is getting hung up on a secondary node that exhibits high latency and packet loss. Now, if they need to call or email the cloud service provider to resolve a problem, they can present a snapshot of exactly where the problem lies, down to the IP address of the problem network node.

Alternatively, a network manager might find that the problem lies within the telecommunications infrastructure between an organization’s network and the cloud service. Rather than wasting time attempting to get a resolution from the cloud service provider, network path visibility tools allow administrators to determine that the problem lies elsewhere and drill down into the specific node for contact information.

The data from these tools can also be crucial for capacity planning—historical data can be used to anticipate when the organization will need more on-premises resources to alleviate bottlenecks. And it can enable IT to secure a service level agreement with its cloud vendor that ensures certain capacity at certain times.

Gaining Modern Network Visibility  

Network managers and IT organizations should adopt the following best practices to better visualize and control the networks they don’t own:

  • Monitor the entire network—on-premises and cloud—from a single platform that can visualize the complete network landscape. Through a single platform, IT professionals should be able to see when application performance is slowing down or underperforming, whether in the cloud or on-premises, and compare relative performance to make informed decisions. A single platform provides a holistic view of the organization’s activity, enabling IT to turn data points into valuable, actionable insights. If IT professionals—especially network administrators—have to go between multiple platforms to manage these, they risk losing continuity.
  • Quality of Service (QoS) and end user experience is key. To ensure quality of experience, an important metric for an IT department’s success, IT professionals need to be able to trace how end users are actually using any one application, and see the service quality firsthand. Key metrics should be quickly identified and monitored to generate better visibility. Monitoring must include actionable insight, such as details on utilization, saturation and errors, which are all critical for speed, collaboration and QoS.
  • Remember that cloud providers are not the enemy. Network administrators and other IT professionals often find it difficult to trust cloud providers with their data because it travels across networks they don’t own. To combat cloud migration jitters, network administrators should have a fundamental understanding of each provider and its services. Many services provide information and reporting that can be integrated into an IT department’s own applications to create a more comprehensive approach to monitoring and managing on-premises networks and the cloud.

Destiny Bertucci is Head Geek at SolarWinds.

EnterpriseAI