American actor Lee Van Cleef (1925 - 1989) plays the villainous Setenza in Sergio Leone's spaghetti ... [+]
Technology is both good and bad. In general terms we can define ‘good’ technology as software products, services and solutions that fulfil the purpose for which they are designed at a cost that supports (rather than inhibits) profitability… and with functionality that boosts productivity.
Defining ‘bad technology is more difficult, principally because it usually was at some point considered to be good technology. But as time goes on, software platforms evolve, standards and form factors progress, creative innovations come about and shinier newer software services (and indeed hardware units) are brought to market by enterprise IT vendors who all push to create the next big thing every year.
Why bad IT pervades
The problem is, it’s often quite tough to get rid of bad (or at least less performant than other) technology. Customers often buy into platform-level enterprise technology (such as a database, a document and/or process management system, a CRM or ERP suite or some other weight piece of tech) with a view to using it ad infinitum, or at least beyond the working lifespan of the current IT manager.
This is not to suggest that it’s impossible to get rid of clunky software; every technology platform company worth its salt will have a SWAT team of migration engineering specialists who may know more about their competitors’ products than their own in order to help customers move-off of their incumbent IT suppliers.
But even with the opportunity to migrate, some businesses sit with bad software because it’s just too expensive to move. The result is that the vendor behemoths just keep reaping in licence fees and the customer grins and bears it. So what should businesses think about this predicament and how should they counter the challenge?
In his role as president & global head for modernization at automation-led IT services company Hexaware, Milan Bhatt says that where businesses are still sitting on bad technology, it’s often because many have no idea where to start and mistakenly think everything needs to be modernized. New York City-based Bhatt has seen a lot of organizations convince themselves it’s simply too complex or expensive to modernize making them a kind of tech-purchasing sitting ducks… but he has three steps for modernisation.
“Map the current environment,” asserts Bhatt. “Too many businesses don’t have an overview of their IT environment – without this, how can they realistically make decisions on whether/how to modernize? Key details to understand range from mapping the underlying infrastructure, whether there are apps sitting abandoned and whether it’s even possible to modernize the legacy applications they have.”
Step two he says is to avoid going a mile wide and an inch deep. Organizations should not assume they need to modernize absolutely everything. They need to identify clear business use cases to ensure any modernization will have a definite beneficial impact. They need to interrogate the whole IT landscape: spending three years modernising the entire IT estate ‘for the sake of it’ would be a huge waste of time and money.
“In the fight against bad IT, firms should use a modernization platform for speed and precision,” claims Bhatt.” The actual process of modernization is extremely complex: businesses have tens of thousands of individual decisions to make, which would take years if reviewed manually. Platforms can help understand the current situation, build a modernization strategy and then modernize relevant apps, drawing on the power of automation to keep the process as quick and light-touch as possible.”
The gravity of debt
Where this brings us to is the notion of technical debt i.e. the increasing cost of compensating for a simple-but-limited approach taken at some point in the past. The weight of this debt is one that’s long concerned Ravi Mayuram in his capacity as CTO at open source document-orientated database company Couchbase.
“Organizations will have collected technical debt over a long period of time. Paying this off isn’t a simple act of replacing one piece of technology in the stack. Instead, it’s like pulling at a single thread and ending up unravelling an entire sweater. The problem is so big that spending the time and effort to modernize becomes too intimidating, perpetuating a legacy stack that costs the organization much more in the long run,” said Mayuram.
But at the same time, he cautions, organizations need to understand why they’re stuck with this debt. A lot of it comes down to skillsets: people who aren’t trained in new technologies will want to retain the status quo. At the same time a ‘grass is greener, something better’ approach of an all-new stack or skill set, or a different approach altogether, won’t necessarily do the right thing.
For instance, says Mayuram, moving to the cloud should be a slam dunk in terms of making services more agile and controlling costs, but the shine can easily come off when businesses realize they’re stuck with their data in the wrong world region; or with pricing options that force them into a higher bracket just to get what they need.
“The cost of modernizing away from this legacy tech can seem excessive, so organizations tend to adopt an ‘if it ain’t broke dont’ fix it’ policy. But this typically has a cascading effect; leading to legacy stacks and ever-increasing technical debt,” said Mayuram. “Looking at ways to smooth the journey – for instance using established standards, skills and processes – can help organizations finally break free of debt.”
Yet moving to the cloud is not always a ‘slam dunk’ in terms of cost savings and efficiency gains, argues Matt Barker, president of cloud native services at Venafi. “Companies have been desperate to move to the cloud, yet they have failed to understand what an undertaking it is to get it right. Many organizations have just done a ‘lift and shift’ of their traditional on-premises patterns into a public cloud, expecting the cloud to solve their efficiency issues and make it cheaper. But it doesn’t work like that. Sure, they may see some gains in the short term, but over time they end up dragging all their old problems along with them.
A reckoning on cloud
Barker argues that we are now heading for a reckoning on cloud which will force companies to look much more closely at how applications are designed.
“Cloud gets much more expensive relatively as you scale, especially if you’ve not optimized your infrastructure. Now we're working in a tech winter, companies are taking a proper look at what they are spending, and how they are spending it. Finance and operations teams are unwilling to overlook the big cloud bills anymore, which is why we are seeing such an uptick in interest in FinOps. Unless companies start to plan their migrations properly, refactoring and designing apps for the cloud, using cloud native technologies, then these problems will only multiply,” asserted Barker.
Roman Spitzbart, VP EMEA for solutions engineering at Dynatrace agrees that the way in which organizations approach modernization and cloud migration is often at the root of their continued reliance on bad IT. He says that very often, modernization drives become a challenge because organizations haven’t actually modernized their applications.
“What happens here is that these firms have just been picked up and moved to the cloud, rather than being rebuilt on a modern architecture. As a result, the applications end up performing even worse than they did on-premises, because they simply weren’t built with the cloud in mind. This erodes the Return on Investment (ROI) that organizations see from migrating to modern cloud-based services and encourages them to stick with what they’ve got, hence they end up maintaining their reliance on ‘bad IT', although it’s far from an industry-standard term,” said Spitzbart.
Spitzbart believes that to combat this scenario, organizations need to shift their focus to ‘doing cloud right’ rather than adopting half measures. To get cloud migration right, he suggests that observability is key. Without it, the complexity and ever-changing nature of cloud can become overwhelming and it’s impossible to optimize applications to perform at their best.
“By combining observability with deterministic AI and automation, organizations can gain the insight they need to rearchitect their applications successfully without putting their operations or customer experience at risk. This will enable them to modernize existing workloads in a way that drives greater speed and agility in the cloud, boosting confidence that the rewards of moving away from ‘bad IT’ will make the effort worthwhile,” surmised Spitzbart.
Will all future tech be good?
Can we look forward to a future where more technology is good, or at least where there is a lower overall percentage of bad clunky brittle technology in place?
The rise of cloud-native technology and the composable nature of modern IT stacks suggest that the prospects are positive, but the devil will still be in the implementation detail and the nuances of every individual deployment. With major influencing factor technologies threatening (in a good way) to shake up our use of business IT every day - such as the move to real-time data or new advances in AI - we need to keep the thoughts here in mind to make sure we focus on the good, reduce the bad and if at all possible make sure we avoid the ugly.
