Tool sprawl is a daunting problem that plagues enterprise IT everywhere you look, from data centre operations to cybersecurity to network reliability to application performance.
Tool sprawl occurs when organisations acquire licenses (or not, in the case of open source) for multiple tools that tackle related, but not completely overlapping, issues. Layered security is the most obvious example of this, but the problem bedevils networking, DevOps, cloud teams, etc.
“It doesn’t matter how many tools you throw at network and infrastructure management, there will still be gaps,” says Mark Leary, Director of Network Analytics and Automation at IDC. “Observability is a massive task. For monitoring and measurement alone, the majority of organisations have somewhere between six to 20-plus tools because they all do something a little bit differently than the others.”
Why is tool sprawl so hard to solve?
A survey conducted by 451 Research found that 39 per cent of respondents were juggling 11 to 30 monitoring tools in an effort to keep an eye on their application, infrastructure, and cloud environments – with eight per cent using between 21 and 30 tools.
For network monitoring and management alone, IT teams may have some combination of Sumo Logic, Perimeter 81, TrendMicro, Opsview, Nagios, ManageEngine, SolarWinds, and ExtraHop, to name only a few, as well as vendor-specific tools and narrow point products that address niche or emerging problems.
“From on operational standpoint, this creates inefficiencies in every aspect of managing IT and managing networks,” said Mike Fratto, Senior Research Analyst, Applied Infrastructure and DevOps for S&P Global Market Intelligence (451 Research’s parent company). “The operator now is the glue among all of those tools, which leads to swivel-chair management. The operator must constantly pivot from console to console to console.”
The connective tissue among those consoles is whatever logic holds the patchwork together in the operator’s head, which makes that person not only a single point of failure, but also someone who is essentially indispensable and extremely hard to replace or even relieve in a tight job market.
Even with multiple tools and an operator who can juggle the many dashboards effortlessly, modern, complex IT environments require still more. “Even with the latest tools, you will have gaps in your visibility, and this can be very disabling. If you have all of these tools, but there’s no unified architecture, then IT is always stuck in a rut. Inevitably, highly skilled, highly talented IT workers are trapped doing mundane work,” Fratto said.
Leary agrees. “Observability alone takes so many tools, and that’s before factoring in cloud, security, containers, etc. For networking, the problem is even worse. Networks are comprised of a bunch of different subdomains designed for different purposes, so you end up needing tools to monitor your data center, your campus network, cloud network, corporate Wi-Fi, 5G, and it never stops,” Leary said.
Each new subdomain requires greater specialisation both in the tools used to monitor and manage it, as well as in the staff that must receive specialised training to run those tools.
“This drives up cost, and reduces productivity,” Leary said. “But even with all of those tools, you still have to cope with gaps.”
What causes tool sprawl?
Tool sprawl occurs for many legitimate reasons, including M&A activities, staff turnover (with new staffers bringing their preferred tools with them), and the constant push and pull between best-of-breed vs. consolidated products.
Added to those causes are such common issues as vendor-specific tools that don’t integrate well with others, different internal teams purchasing different products for similar tasks, the prevalence of open-source options, and the need for networks and infrastructure to be nimble and dynamic to adapt to ever-evolving technologies and constantly shifting market conditions.
Another issue is day-to-day firefighting, which keeps IT trapped in a reactive mode. “Often IT buys tools in order to solve a very specific problem, which can lead to a glut of niche products that don’t integrate well with other monitoring and management tools,” Leary said.
Over time some of those niche products may be consolidated into larger platforms or integrated into vendor-specific ecosystems. Yet, even as tools consolidate, networks and infrastructure keep evolving in ways that create new opportunities for startups to roll out new point products that plug new gaps.
And the cycle starts all over again.
How much does tool sprawl cost?
The high cost of tool sprawl is obvious, but opaque and hard to quantify. The escalating cost of overlapping licenses and use-it-once shelfware are easy enough to add up, but many of the costs of tools sprawl, such as lost productivity, various opportunity costs, and lower job satisfaction for highly skilled and difficult to replace senior IT leaders are far more difficult to measure.
According to an August 2020 report in which 451 Research surveyed 700 IT professionals, the market-research firm found that only “11 per cent of decision makers are satisfied with their monitoring tools.”
Dissatisfaction with tools can easily transform into job dissatisfaction. “Tool sprawl can take on a life of its own, miring network operators in mundane chores that burn out highly skilled professionals,” Fratto said.
Leary agrees. “Should your senior staff have to figure out which alarms mean something?” he asked. “Because that’s what they constantly have to do today.” Most organisations and most senior IT leaders would rather have these highly trained specialists working on architectural, development, and strategic projects, rather than hunting down the root causes of whatever just broke.
How to tame tool sprawl: 7 steps
1. Don't be afraid of new tools
While this may sound contradictory, adding yet more tools could help tame tool sprawl. Network automation tools from the likes of NetBrain, Infoblox, SolarWinds, and Cisco apply AI, ML, and analytics to the problem of network sprawl and help organisations begin automating various tasks.
For certain use cases, such as connectivity to branch offices, many enterprises have already embraced managed network services, such as NaaS and SD-WAN, which have analytics, intelligence, and automation built into the core offerings.
2. Embrace outsourcing
Outsourcing can help lift some of the burden off internal IT staffers, but don’t expect a managed service to solve all of your problems.
“If you outsource pieces of your network, such as 5G, you will get the advantage of specialisation, and your MSP will handle various tasks for you,” Leary said. “But it’s like the early days of cloud. People didn’t close down their data centers. Rather, they passed along certain workloads to the cloud, and if that went well, they added more.” Don’t be surprised if networking and network automation follows a similar path.
3. Consolidate and integrate
Over time, larger vendors will acquire some of the point products you rely on, and in other cases, those point solutions will simply become features in competing products that do something new that you need to address.
Consolidation can certainly help tame sprawl, but even with consolidated products, integration with all of your other tools is key. Vendor ecosystems are a great place to find products that work well together, but what happens when a specialised point product that you rely on doesn’t fit into the ecosystem?
Holding vendors’ feet to the fire about integration and open APIs can help you avoid more sprawl down the road.
4. Take the easy wins
There’s no need to make cutting back sprawl more difficult than it has to be. “Where are IT people spending the bulk of their time? If it’s managing data streams and managing multiple tools, start there. Integrate and consolidate,” Fratto recommends. “Find patterns that are continually recurring and solve those.”
5. Improve processes and workflows, not just tools
Not all fixes need to be technological ones. Having clearly defined processes and workflows for IT teams to follow can deliver big dividends simply by protecting senior IT leaders from being pulled into every mundane issue because less experienced staffers fear making a decision. Moreover, clearly defined processes and workflows will cut back on the ad-hoc nature of IT problem solving.
6. Break down both data and organisational silos
One reason network tool sprawl is such an issue is that the many devices connected to the network churn out a constant stream of data, but the data tends to get locked into application- or tool-specific silos. Finding a way to break down data silos will help you get more value out of your data, but many of the silos impeding progress aren’t data silos, but organisational ones.
“From an organisational perspective, enterprises need to find a structure that lends itself to working together, using common tools, and sharing data,” Leary said.
Many IT teams take a proprietary view of their tools and their data, often reluctant to share. “Collaboration across teams doesn’t happen naturally, so leaders must actively encourage it,” Leary said. “If they value the kind of collaboration that reduces sprawl, organisations are going to have to force the issue.”
7. Automate, automate, automate
“Automate” is such common advice for solving IT problems that it’s become the equivalent of a fitness instructor reminding you to stay hydrated. Yes. We know. However, if you’re prepping to run your first marathon, your trainer will provide you with actionable advice that will help get you to the finish line.
As you seek to consolidate, streamline, and modernise, don’t forget common-sense guidance. Commit to finding tools that are easy to integrate and that deliver not just an avalanche of data, but actionable advice.
“Automation lends itself to better reliability. It standardizes things,” Leary pointed out. “But the big gains today don’t come from basic automation, such as auto-configurations, but by making your networks more dynamic, more robust, and better able to adapt to changing loads.”
If you hope to finish the marathon of containing tool sprawl, seek out tools that leverage analytics to automatically understand how network conditions are changing, where new demand is coming from, and how to adapt on the fly, while eliminating manual, error-prone processes.