When workloads are distributed across cloud, data centre and on-premises environments, application performance and resilience is a make-or-break issue.
Every IT team is well aware that, if access to an app goes down, it can stop employees working, cause mission-critical parts of the organisation to grind to a halt and impact badly on the customer experience.
Although huge app crashes like this are a rare occurrence, there’s an equally critical – but more hidden – effect happening. As networks creak under the strain of supporting apps / workloads in a distributed environment, app performance is often hit in ways that are hard to identify – but have a cumulative, serious effect on operations.
Are today’s tactics for tackling app performance issues working?
Within this scenario, IT teams can solve what they think are individual problems cropping up in different areas, all their traditional Service Level Agreements (SLAs) can be met, and every Red-Amber-Green indicator can be a steady green. However, this surface fix isn’t always an absolute fix. There can be a larger, underlying problem causing these apparently individual issues – but the current level of performance insight just can’t identify it, and so the ‘individual’ issues rumble on. And because Experience SLAs (XSLAs) are rare, there’s no incentive for providers to approach services on a more holistic basis.
Some organisations look to SD-WAN to improve the flexibility of routing and path selections to support app performance. But many find SD-WAN isn’t fully meeting their expectations of cost savings and transformed performance because the underlay it’s paired with isn’t performing as well as it could. Variable internet latency from multiple hops across different ISPs plus last mile makes a consistent, high quality end-to-end app performance difficult to ensure, and reduced visibility and control over app performance make it harder to diagnose the root cause of the problem.
Similarly, when considering cloud connectivity, you need to think about how cloud provider SLAs align with networking SLAs, and how you create a solution that is aligned on outcomes. This is further compounded by increasingly complex multi-cloud solutions, that in turn require a much deeper understanding of the connections between the cloud and how to assure performance when unexpected situations occur.
At the bottom of all these application performance issues is one simple fact – the underlying network wasn’t designed to support the emerging, more distributed, workload environment.
Three essential network developments to protect app performance
When organisations move completely into a distributed environment, they need to be confident they can rely on the performance and resilience of their applications. What capabilities should IT teams be looking for from any network evolution?
1. Wider, deeper, real-time monitoring
A DIY approach combined with basic metric reporting on latency, jitter and packet loss isn’t enough if you’re supporting big data and distributed workloads. Leveraging the right skills, resources and tooling is critical to ensuring end-to-end visibility of app performance. Having the right partner is essential to understanding where and what to invest in.
Organisations know this, so they’re looking to shift their monitoring from passive to active, using availability and performance information that can only come from seeing the app’s complete delivery pathway. This will involve combining hyperscaler insight data, more detailed monitoring data, as well as machine learning and anomaly detection techniques. It will also require automated methods to easily discover and map network and app structures, and any flows across the complete app path. Incorporating diagnostics and repair processes to enable near real-time fault detection before there’s a noticeable impact will be vital, too. On a more macro scale, this insight will enable IT teams to spot apps hosted in places that aren’t the best choice for the organisation, and even to identify purchased capacity that’s no longer needed.
2. Easier app migration processes
Organisations are operating in an environment where business priorities can change fast, and many face the task of migrating a huge volume of legacy apps to the cloud. IT teams want a rapid, easy process that will allow them to site apps wherever they best fit with business goals and outcomes – and then action the necessary shifts quickly. Optimising operational performance can also involve adopting multiple clouds in multiple regions, and then moving apps into or within them. Genuinely easy app migration services need to be integrated into the organisation’s own workflow tools via effective APIs.
3. The ability to flex with regulatory developments
However, this greater flexibility in where apps are hosted can be hampered by regulations for data storage and processing that vary from country to country. For example, a banking organisation might not want their data leaving Switzerland, and would need the capability to site apps quickly in that domain. It also takes time and expertise to navigate the implications of moving an app from one jurisdiction to another, particularly when the regulatory landscape that governs moving apps is continually shifting. Organisations want a network that can work well in this scenario.
Global Fabric is designed for an app-centric world
Our new global network is a programmable platform, delivering networking services to the cloud and between the clouds, whether public or private. It will bring increased speed and flexibility to network changes, opening up new levels of responsiveness. Now, larger organisations that were traditionally slower to adapt than smaller competitors will have a network that’s no longer a drag on their speed of change.
Global Fabric will offer enhanced end-to-end monitoring, giving you the application visibility and control you need to make better decisions, and address issues on your network quickly and intelligently. It’ll provide a step change in analysis and monitoring, backed up by robust SLAs to the cloud and between clouds, whether public or private.
Importantly, it will include support for experience SLAs. Just as Global Fabric is the next level of networking, XSLAs are a significant advance in accountability within a contract. Aspirational and outcomes-based, they add context and emotion into performance analysis. Offering XSLAs is crucial to Global Fabric’s intent to transform cloud networking, because they measure what’s core to Global Fabric’s success: supporting reliable app performance.
To keep app performance strong, if there’s an issue, traffic can be automatically re-routed to an equally efficient non-faulty path. This will keep apps running, employees working effectively, and end customers served and satisfied – all without degradation in service.
And when the organisation wants to adapt and respond to market changes, Global Fabric will be right there, with the ability to stand up services in minutes so apps can go wherever they’re needed.
For the first time, you'll be able to use embedded monitoring capabilities to look at predictive performance before you select routes or buy services. Global Fabric will also offer optional ways to build end-to-end observability across multi-clouds, from the end-user device, all the way through the LAN, WAN and within the hyperscaler environment. It’s easy to split hosting geographically to achieve best performance – but keep that overall monitoring picture. And it’s a network that will make it possible to move your workloads from one hyperscaler to another at the press of a button if your app migration process supports that.
We also recognise that organisations will want different levels of management support with Global Fabric so, again, we’ve built flexibility into this area. Hand management over to us completely, use our tools to manage your usage in-house, or opt for a hybrid approach – it’s up to you.
Visit our dedicated webpage for more information, including our whitepaper and expert blog posts exploring the five transformative benefits of Global Fabric.