Taking data and analytics to the cloud is a winning strategy. Many people, particularly data scientists, mistake data for wisdom. They are distinguished by their maturity. Data tells you that the red thing is a tomato. Knowledge tells you that it’s a fruit. But only wisdom tells you not to put it into a fruit salad. So data is raw, unrefined source material for knowledge. It needs the treatments of Comparison and Experience in order to be turned into wisdom.
By nature, data is malleable, which is why it can be found in data lakes and, with the application of energy, it enables a telecom data warehouse migration to the public cloud. Should data, like the tomato, be put into the Fruit Salad that we call The Public Cloud?
What is data? Data is raw intelligence stripped of any context. Data types are as varied as the fruit family and they all contain fertile seeds that can germinate ideas.
While fruits have three layers, the corporate data lake tends to have four layers, as the chart above illustrates and two categories of data pervading them. The data elements are the Relational Data, which tends to be deep and complex, and fast moving Streaming Data, which can easily wash over you.
Both will be, by turns, subject to the stages of Ingestion, Distillation and Processing before finally (hopefully), emerging into Insights. It is worth outlining the differences between the two types of data.
What happens when you are moving a telco’s data to a public cloud?
Relational data comprises types like database tables and spreadsheets, which would typically be used in business departments and applications. Streaming data comes at you fast from sensor output, application logs and mobile interactions. This is the domain of the technical departments.
Source: Nokia
A data lake is a bit like an industrial scale version of the ‘data dump’ that’s performed when a business associate throws random abstracts at you and expects you to make sense of them. In both cases they operate on a ‘store now, analyse later’ basis.
The next phase of ‘on premise’ or ‘in warehouse’ data analysis was traditionally known as ETL (extract, transform, load) in which data was pulled and prepared for analysis.
However, the machinations of a modern communications service provider run on sub-millisecond response times. Let’s examine how some systems integrators are migrating a telco to a public cloud.
There is a mismatch in form involved in a telecom data warehouse migration to public cloud.
If, say, Spain’s flagship mobile operator Telefonica wants to offer subscribers an upgrade on their data package the second they exceed their limit, as they celebrate a goal by Real Madrid, they need to do it instantly, before the excitement dies down.
So, the analytics tools for the telco cloud are quite different.
Moving the old Data Architecture of warehousing to the cloud may well fall short of your business requirements. The phrase telecom data warehouse migration to public cloud evokes images of castles in the air!
There are four qualities you must retain once your data has left the lake, the warehouse or the silo and heads into the cloud. You want to embrace an open architecture without being trapped in a database lock-in.
How telcos are moving analytics to the public cloud
If you want as many brains – human and robotic – applying their analysis to a situation then by necessity you must grant ubiquitous access to the intelligence (aa the data) available. You will want to constantly refine it, so that it is constantly operational and maintains accurate historical integrity.
So, there are huge risks involved in moving data to the cloud. There seem to be three types of risk. The data types, the analytics and most ominously of all, the choice of cloud technology in the telecoms industry.
Any object store such as Amazon S3 will be a good fit for the cloud. Amazon Web Services (AWS) also brings constant gratification. However, once it has wrapped itself all over your business processes and your data, it might be very difficult to extract the business if another service provider offers better options.
Less cost, more revenue, unlimited growth – this all sounds very easy. Obviously, the decisions about cloud migrations for any telco are anxiety-inducing strategic choices. Witness the industry machinations as telcos look for partners to help them with all the difficult decisions involved.
In Britain, for example, telco VMO2 is a hotch-potch of incongruous systems that once belonged to local cable TV companies, networks and a mobile operator. There are better integrated fly-tips.
The task of integrating them is so vast that they’ve never achieved it. They just bully any unhappy customers with legal letters, which is cheaper! So the data side is incredibly hard.
Meanwhile, Telenor (in Norway) seems to have handed its data into the custody of Google Cloud, whereas Vodafone is being more hands on with its data infrastructure conversion and working with VMware’s Telco Cloud.
The perfect compromise is to work with a hyperscaler without being hyperscalped. It can be avoided. Once more, there is a layer of context, that lies between The Data of of a company’s written ambition and the knowledge that comes from completing the project. That context comes from an experienced provider of telecommunications software development services – Intellias has a good track record in turning ideas into execution for telco clients.
Pros and cons of moving a telco data to a public cloud
The archetypal public cloud supplier is obviously Amazon Web Services (AWS), but Azure and Google will also perform a telecom data warehouse migration to public cloud.
AWS migration services for telecom will do everything quicker because it’s done the job before. Many, many times more than you. It learned a lot about mobile and broadband data transmission services when dealing with 23 million subscribers.
Will AWS scale you without scalping you? Will it optimise costs and then unoptimise them all over again?
There is no doubting the hyperscaler’s ability to sweep data up from on-premises servers and spirit it into the cloud. By migrating its existing data warehouse to AWS, one Intellias client had designs on streamlining its data processing and business intelligence (BI) systems. So Intellias conducted the proof of concept (PoC) to verify the efficiency of such a transition.
The knowledge and experience gained from this successful project applies to any telecoms company.
Intellias customised the design by assessing resources needed, measuring the performance of the developed systems and estimating the expected infrastructure costs. This allowed the client to validate their ideas and estimate running costs to help them decide and plan the AWS cloud migration strategy.
There is no easy data warehouse migration strategy for telcos to public cloud.
Telcos like their database to be on the premises so they can use their business intelligence tools on it and watch the lights flash as the data is processing. In this case though its data had mushroomed and grown out of the existing data warehouse. Worse still, the costs of maintenance, vendor licenses and hardware had gone through the roof. So the telco saw migration to AWS as a trip to the promised land of low cost and bountiful resources.
As with any migration there was an acclimatisation needed. Upgrading all the underlying big data and improving the process of analysis would prove very valuable.
That involved three big tasks: accounting for customer consumption and billing them, predicting their behaviour (through modelling) and, finally, watching it, through stream data indexing. Obviously, for the purposes of brevity I have oversimplified, but there are many labours involved in detailing a monolith into microservices.
The bottom line is that by working with an integrator a telco can sign a pact with AWS without promising a pound of flesh.
The telecom data warehouse migration to public cloud will open new vistas for the telco, with lower costs and wider perspectives. By using its intelligence to personalise services it can keep its customers happy.
A telco doesn’t have to be locked in to their premises or locked in to any vendor.
That sounds like the perfect salad.
Contact if you are interested in fruit salads, moving telcos to a public cloud or just plain data warehouse migration strategy for telcos to the public cloud, we have a team of experts who can make any complex question seem as clear as an Azure. We do this every day so we know how telcos are moving analytics to the public cloud and are involved at every level from the architecture to the operator to the utility.
