Thomas is a guy who I always enjoy speaking to – not just because he is a great personality but also because he always talks sense about the data architecture world. We had some questions for him about the relationships between big data, traditional architecture and the looming prospect of GDPR
Collaboration and bringing together the decentralised teams is the key to GDPR success within Analytics domain.
1) You describe yourself as a Solutions Architect who loves to talk technology. Your linkedin references compliment your commercial or enterprise knowledge. Is it possible that the traditional architecture job titles are becoming somewhat blurred?
Yeah, I understand why this needs to be described. Coming from a commercial division before joining IT, I have a somewhat closer relation to business in general, which was actually why I was transferred to IT. Due to my background I have done some very techy stuff although being in Commercial, thus forming an informal bridge between two normally competing divisions. My architecture role today has become a success for this reason, I am respected in both divisions due to my history and deep knowledge about legacy, as-is and the roadmap ahead of us. I challenge both business and IT, and due to my flair for new technology adoption, I am given the tech responsibility of introducing new platforms as well as securing the coherency to existing ones. I am thus not a traditional IT architect, and I am also missing quite a few high-level architecture experience points, but this is compensated by my own experience. So you can say that I am domain expert in both it and business, and coupling this knowledge to a more low level solution architect role with the ability to talk straight ahead to CXOs in both areas.
2) The Harvard Business School has today claimed that ‘companies love big data but lack the strategy to use it’. How do architects ensure that the business ‘get’ the point of big data?
Well this is a really good question. I hear a lot of similar statements from contacts within the domain. Coming from Business side, i have a strong gut feel that this is the situation in Telenor as well. Telenor is a typical telco, with a strong focus on selling simcards, retaining customers and survive the rise of online communication tools invading the mobile domain as fancy user-friendly apps that even elderly adopt quickly. As an architect, i constantly challenge business to face the fact, that we do not promote the company towards greater times like 15 years ago if we do not focus on also increasing the topline. Big Data is key to doing so, as telcos are amongst the most data-savvy companies out there, and the huge customer base is representing a lot more value than just being something you charge for close to flat rate / no frill services. I mean, we have the option to utilize the insane amounts of data we have, to strengthen our network, collaborate with government or tourist driven infrastructure companies etc. I am here avoiding focusing on the tempting subjects of the popular contextual marketing domain, as a will never fancy the big brotherish way of tailing individual behaviour, consent or not. As an architect I believe that striving for generating new revenue by respecting individual right not to be tracked is very satisfying, and to work with huge data loads provides exactly those options. But telco business is very conservative, and it takes more than a solution architect to convince business professionals to focus on completely new lines of business. Let me here mention that solution architects do not handle the full architecture function of a company. According to TOGAF, solutions are only to be provided once architectural work is required, establishing or changing capabilities, triggered by Business attention and requirements. So i believe that Big Data is fully exploited only if business architects can promote the tempting capabilities for Business.
3) You were involved in the first commercial big data project at Telenor. Often greenfield big data projects depend as much on engagement from business users as it does the quality of the technology. What was key to making sure that Telenor made the best out of your installation?
Ah, actually, the very first commercial Big Data was not implemented by me. It was deducted from the IT transformation program I was following, only to be rapidly deployed directly by techy new business entity, completely isolated from IT participation. Being all vendor driven on a low cost basis, with less than minimum setup on old hardware generation, the scene was set for a small development / explorative platform which was used to implement somewhat attractive products facing public markets. It has been going on for quite some time now, with only little success. It has however got big attention from customers, and from internal sponsor. What is currently missing for going to market with a full blown solution is a stable production platform, which I am currently implementing. It has been on its way for more than a year now, being challenged by the lack of investment will from management. But recently, with the big re-org, my new department has got the full management support to start the enterprise focused Big Data project that will ultimately deliver a production environment for the existing prototyping environment, but also serve as a common data lake for all data consumers to go to one place for getting both data and tools for working with data.
4) Job Specs with the titles ‘Big Data Architect’ are becoming increasingly common as the market matures and develops. Do you think that it is possible for a solid architect to successfully run a big data project or is the domain big data knowledge a necessity?
Good question, I dont believe that big data projects can succeed without big data competencies, but it does not necessarily require Big Data architect experience. However, it requires interest, and will to enter a new world. It is indeed different, and at least you should be well known to the overall analytics domain where issues often occur when traditional technology gives up on huge data sets.
5) There are a lot of things claiming to be the next big thing in the big data world – though GDPR is rapidly coming onto the horizon. How do you see that Big Data can coexist with the GDPR
Actually, I see Big Data solutions as being complimentary to GDPR, especially when used as Data Lakes comprising all the data that are normally found scattered around in big enterprises numerous BI environments where Deep dive knowledge into smaller business areas often reside as either satellite solutions or even shadow IT. Historically centralising reporting/BI teams into a coherent and enterprise covering single unit has been popular efforts but results are in my opinion small and often combined with huge lagtime on the Deep dive efforts once so popular and effective. Instead, focusing on centalising the platform, rather than the teams, enables enterprises to keep the domain knowledge level while benefitting massively from the presence of all data at one place, omitting the need for copying relevant data sets into numerous Places for inclusion in domain specific algorithms. Multiple copies of data is one of the key painpoints within the analytical side of GDPR. Knowing WHO is using what data and where this data resides is exponentially harder when the same data is used differently in different environments. Therefore, using Big Data solutions as Data Lakes WITH the tools to analyse the data sufficiently is GDPR supporting, as data copies are avoided to a large extent, given a sufficient governance setup. Big Data does not do this itself, it requires a lot of effort towards securing the Whole team around the solution, from data engineers ingesting data over data scientists to end users.
6, Imagine you were an architect in a country soon to leave the EU … such as the UK … and subject to GDPR for a limited time only. Would this impact your design work in relation to GDPR?
Short answer is NO, not to the full extent at least. I personally think that GDPR is a good thing, offering a lot of potential trust from customers. I find it very useful to actually consider what you are using data for, why the sensitive personal information is actually needed when lots of alternatives to actually do the anonymization is generally available. Also, the Whole idea of documenting and seeking tight solutions that does not fiddle around with vast storage consumption and endless retention just because it is not required are generally good considerations to keep healthy environments with high levels of content quality for the purpose it is serving.