OPTIMISATION RELIES ON UNDERSTANDING
Optimisation relies on understanding
18th November 2019
Fifteen years ago, Michael Hicks, now Director of Product Development at Ikara, published his second book: “Optimizing Applications on Cisco Networks”. It was a practical look at how network and application performance management can proactively support organisational objectives. To his credit the book is still selling online. We asked Mike about the book, about what has and hasn’t changed, and about what business pressures are driving Ikara forward.
Mike, it has been 15 years since you wrote “Optimizing Applications on Cisco Networks”. It was a very different time. Can you talk about the networks of 2004?
It was not so different from what we are still trying to achieve today with intent-based networking. That is to make network traffic management decisions based on policy, or business rules. Imagine your CEO is visiting a remote office. You might want to give her traffic the VIP treatment. Give it priority over most traffic, but not over mission-critical traffic, for example.
That might seem extreme and in 2004 it was probably beyond us. We were trying to optimise network traffic to ensure business outcomes, but it was all manually configured. Today with Software Defined Networking and other automations, this sort of dynamic control is coming within reach.
So, the book was for network technicians, for people that manage the network?
That’s right. However, managing network traffic in isolation effectively ignores the other half of the equation, which is application performance. The network engineers were only concerned with passing packets. If the network was forwarding data, that was all we cared about. The application guys were just the same. If the application worked at the server, it was fine and anything that happened beyond that wasn’t their problem.
It ignored the potential for business disruptions caused by the interplay of systems in the real world. Meanwhile, the enterprise was increasingly reliant on these applications working as promised.
Tell me about the applications. They would have been quite different from what we see today with the emergence of cloud and pervasive SaaS applications that we see in use today.
Back then the environment was already becoming complex and that was the reason for the book. We were already moving away from hub and spoke application architectures toward meshed network environments. We hadn’t graduated to micro-services or serverless type architectures, but we were starting to split functions up. We’d distribute functions and use CDNs to reduce WAN reliance, but we still had big data centres, which were serving, doing replication and so on.
So, even though applications were becoming distributed and the concepts of QoS (Quality of Service) were around, we just didn’t have the tools to manage this dynamically. It was still very much ‘guess, set, and forget’.
Since the book, you have stayed quite close to the industry, including as a Product Management lead at application monitoring developer Dynatrace. How has the technology evolved beyond the networks of 15 years ago?
I’ve been in monitoring, performance monitoring and network monitoring, for a long, long time now, but we’ve always talked in terms of technical metrics. We’ve talked about things like packet loss rate, jitter, latency. Monitoring has improved enormously, but we’ve lacked a way to correlate that into impacts on real business metrics.
The industry is starting to talk about business insights, but all too often there’s still a disconnect between the technical metrics and the business metrics.
You recently joined Ikara as Director of Product Development. What attracted you to the organization?
There are two aspects to this. I’ve had an ongoing interest in self-healing networks, where you proactively, or even automatically rectify problems based on feedback from the environment. This still lacks an understanding of the business rules. We could tell you about network downtime, but not about how that affected the business.
If an issue falls outside that core technical metric, I can’t properly interpret it. All these outliers can directly impact operations. If you can understand them, you could ultimately improve that business KPI.
This is what Ikara does with Echelon. It lets us build connections between the business rules and requirements and the complex array of systems and providers that enable them. It effectively allows us to ingest that disparate technical data and interpret real-world business consequences from them.
So, users are getting the benefits? How do these changes impact technology professionals?
Back in 2004 technology was such a ‘black box’. There was a disconnect between user expectations at work and their experience in their personal lives. Users would complain that the network was down, or slow and the network guys would check and say, “No, it’s fine. The packets are being passed. We’ve got connectivity.”
Today the users have 4G on their mobile, they have NBN in their home, and there is an expectation of performance. They just aren’t willing to accept less in the workplace.
They’re saying things like, “I want the same experience at the mine site as I would if I was back at head office”, or “I want consistent performance regardless of my location or the time of day”.
What does that mean for business leaders? They are looking for productivity improvements. Is that sort of performance their focus?
Business decision-makers are more likely to ask what they can we do to improve their KPI’s. To achieve this, they need insights that cover the end-to-end business process. As we progress with digital transformation the technology is positioned right at the core of business processes. That means it has to work. You have to get a consistent experience, but most importantly you have to strive to get continuous improvement.
The only way to do that through technology is to have insights across all the different vendors involved, all the different applications, and aspects of the network.
Where do you see the short-term future of technology moving?
Beyond that. With Echelon we also bring in data that is sourced from vendors outside of the corporate technology team., Like electricity, for example. Because business owners are accountable for the ROI on their technology spend, they increasingly demand better metrics to demonstrate the system is delivering the business results they need.
Senior managers, even Board members, are taking more of an interest. It is still a ‘black box’ to them, but they want to understand how the technology affects service delivery. They will ask; “How will this change affect that KPI?” So, the justification processes for technology change are more stringent, more evidence based. That’s a good thing. If you can deliver the evidence.