Scaling is Still Scary
Meeting Expectations and Managing Risks in an On Demand World
Looking back even a short while, there’s no doubting we are in a new world when it comes to scaling technical applications. Those fluffy white servers in the sky have brought about enormous change. Be it navigating one-off traffic spikes, ramping up to tackle those all consuming data migrations or just meeting the day to day needs of a growing small business, the speed and agility we can summon to meet these challenges of scale would have been previously unimaginable. Additionally, there are companies like Kanerika Software that are offering seriously creative designs that can bring real computational heft to even the smallest enterprises.
Keeping Up With the Joneses 3.0
To be clear, this capacity is by no means evenly distributed. Some times it’s the tech; many enterprises and even entire industries remain stuck in a land of legacy applications that aren’t quite so cooperative when the need arises to do Big Things. Other times, it’s the organizational knowledge and wherewithal. More often than not, it’s simply a matter of time and resources.
While it’s easy to fall behind the times for any number of reasons, market expectations absolutely do not. We may have tolerated an over-capacity outage in an early Web 2.0 world, but no longer. End user patience has evaporated. If the payment doesn’t process or that metric gets missed, we’re moving on to the next application ready in waiting. There’s too many options and too little time.
While the demand for scale is evident, the instances that drive it can come from a very broad array of often difficult to predict sources. Minor changes to standard operational applications yield much greater amounts of data than predicted. Social media driven virality can be a blessing or a curse depending on how well prepared you are (content not withstanding). Entities need to have a multi-tiered approach to an increasing array of threats from bad actors, like a DDOS attack.
Perhaps the most important demand for scale is the one-off market changing initiative. It could be a bold new public-facing feature that’s the first of it’s kind to market. It might look like a recursive analytic project for a research institution on the cusp of a major breakthrough. We generally don’t know what it will look like, but when it comes to fruition, we know we’ll need it then and there.
The Fear Factor
All of this belies a reality many organizations grapple with. Despite all our advances, scaling is still pretty scary. What approach should the organization take? In world of increasingly complex systems, there aren’t too many settings where the on premises binary of horizontal or vertical makes sense anymore.
Part of what drives the fear is the sheer volume of options presented. What does a hybridized and multi-site scaling operation look like? How do we control costs while ensuring optimal efficiency and maximum capacity? We know that what the up/down & right/left paradigm offers in control and capacity it demands in equal parts upfront capital. While organizations increasingly look towards API driven microservices to manage the incremental demands, the sheer cost of operational monitoring and maintenance itself can be vast and difficult to predict. The potential in the omnipresent cloud is difficult to overstate, but is another layer of vendor management worth the flexibility? What guarantees can we make to our customers and stakeholders when we find ourselves at the mercy of a third party cloud provider? This is where companies like Kanerika come in.
Tech to the Rescue
It appears help is on the horizon. There is a class of services providers who specialize in complicated scaling endeavors. They leverage innovate software solutions, often in conjunction with the leading cloud providers to take the legwork out of scaling instances and operations. This can be enormously beneficial, particularly for the types of institutions that may lack the resources to undergo the often taxing efforts themselves.
Take for instance the emergent Kanerika. The Hyderbad-based company specializes in data applications and analytics, particularly at volume. They help institutions move beyond the staid conceptions of aggregating raw data into genuine meter moving analysis.
Kanerika recently underwent an effort to expand the data analytics speed and capacity of a top-tier research institution. The goal was to cull the top 3 green building designs from over a million plus simulations with a broad parameter input for each simulation. The results were produced in an hour, delivering what used to take weeks in terms of research and generating a substantial cost savings.
It’s these types of the scenarios that are the real drivers of scalability requirements in our on-demand world. The ability to summon capacity in real time and collapse just as quickly will distinguish the more intrepid enterprises from its peers.
