By François de Repentigny, Vice President of Guavus
In order for long-term adoption of to be fully realised, big data analytics needs to become an essential part of the telco IT infrastructure. CSPs need to be data-driven.
Traditionally, telcos have turned to Do-It-Yourself IT solutions for business intelligence and data warehouse projects; yet in the era of big data, a new kind of expertise is required, beyond just technology and tools. As such, a holistic operational intelligence platform that can deliver faster time to value and greater business insight is necessary; yet the question remains, should you build or buy?
Every CSP is unique, with different ticketing systems and Element Management Systems (EMSs), but essentially they are all running the same type of technology they have been using for 15 years. As such, IT teams have the experience to know the hot spots to check if something goes wrong, how to optimise configurations to ensure efficiency, and have the skills to build and integrate new technologies.
It stands to reason, therefore, that in-house teams are often tasked with building and delivering any transformative IT projects; including big data. Yet while CSP IT teams are highly skilled, this knowledge is redundant in a big data world; the rules have all changed. Traditional approaches are no longer valid and new skills are required.
The acquisition of these new skills is one of the key drivers behind the shift towards pre-built big data solutions.
As with any new technology, there is always a period of trial and error; as a result, many big data projects experience severe teething problems with over-running timescales and billowing support costs.
The most common misconception when it comes to deploying a big data solution is that there’s a need to store everything, as it’s unknown as to what data is valuable and what isn’t. However, just because you can store everything, doesn’t mean you should.
Hadoop clusters still require CPU, electricity, server space, and most importantly people to manage them. This means that taking a “store everything” approach will drive up the total cost of ownership for a big data implementation.
Additionally, the more data that is stored, the longer it will take for queries to be resolved, as there is more information to sift through. These delays prevent IT from delivering the next level of real-time processing that is needed for timely decision making; as such, taking a build approach can often slow down speed to insights.
Big data is about a lot more than just technology; it is about applying technology to business problems, which requires a combination of business acumen and technical expertise. Business insights cannot be gleaned by simply throwing all your data into Hadoop.
In order to support business decisions with up-to-the-minute, relevant information, you need to do the work up front to understand what you are trying to achieve and how data can support that. By understanding the business challenge, IT is able to set the parameters of what data is and isn’t important; this ensures the business is relying on accurate and reliable information.
Yet many IT departments are focusing on the technology first, without taking the time to think about how the data will be applied in a business context. In reality, business objectives should be set before any code has been written or clusters spun-up.
This is why we have seen such a surge in demand for data scientists – a new breed of engineer that not only has the technical capabilities to build big data platforms, but also the business know-how to interpret the data and deliver value to the organisation.
While the promise of Hadoop may be appealing, it is important that CSPs consider the total cost of ownership when making a decision. Specialist solutions, pre-built with data intelligence, are available and can provide an effective short cut to returns.
By taking a holistic approach, CSPs can ensure that projects are delivered on time, that they meet their stated objectives, and that resources are used in the most efficient and effective way. The key for CSPs looking to successfully make use of big data will be in assessing their internal skills set, both from a business and technical standpoint, and deciding what will work best in their environment; as doing things the way they always have may not work.
Click here to watch videos of European Communications’ big data seminar
Click here to read more opinion pieces