The future looks different in the data economy

Our Technology Evangelist, Neil King spends most of his time helping our customers align with industry best practice. Neil took some time away from his desk to discuss what the future data economy looks like, and how Bechtle ensure our customers are taking advantage of their most valuable asset; their data!

My thought is that the world is changing, right? I mean, the technology world has gone through a period of change, our customers have now got IoT (Internet of Things), distributed systems, edge systems, all of which integrate seamlessly into IT, whether that be a data lake on premise or in the cloud.

If you think about it, smart IoT is everywhere, you have sensors in your fridge at home, or your car, sensors are everywhere. What are they doing? They’re creating data, and a lot of it! For example, a popular electric vehicle manufacturer may upload close to 1 Gb of data every day – per car; and that data is landing somewhere, right?

There is a business need to find value in that data to improve services and deliver next level customer engagement.

Also, not only do you have massive numbers of sensors in our every day lives, you’ve also got distributed edge IT out in locations such as oil rigs, mobile phone towers, manufacturing plants and retail stores, to name just a few.

The Retail industry is full of edge devices because they have Remote Office/Branch Office (ROBO) locations. Consider a supermarket chain, you might have stores spread throughout the country, each one with an edge device connected to the Corporate network. Those edge devices are getting fed data by a multitude of local devices and local sensors.

When a customer goes into a supermarket -say- they are doing things like self-scanning, and that is collecting huge amounts of data points. It’s not just for ease of shopping; the retailer is recording everything you’re doing. That data is being fed into the edge device in store, then uploaded to the companies’ data lake, from that they can work out shopping habits and build up a profile of not only what you shop for but how you shop, how long you spend in store, the route you take in store, where you pause, what items you scan and put back, how much you spend and when.

The retailer will then monetise that by sending you vouchers and offers that they know will appeal to someone with your spending habits. Generally, a member of the public will then return to that retailer with their money saving voucher instead of choosing to go to a competitor’s store.

That is just one -relatable- example of how everything's changing in the data economy, because data is the new gold!

Because companies don’t need to guess anymore!

Data can easily identify trends, daily, weekly, yearly, what ever the company needs to extrapolate from the data they are harvesting.

That’s how the strategy of our customers businesses is being transformed by this revolution.

So, it’s basically, the fact that hardware has got bigger, better, faster and we get a lot more performance out of the hardware we have available to us today.

We now have GPUs (Graphics Processing Unit) where we used to have CPUs (Central Processing Unit). Just digging into that for a second. GPUs are not technically as powerful as CPUs, but they can run a lot of processes in parallel. This makes them significantly more powerful than the equivalent CPU with more memory. It’s a volume thing. If you can multitask better, you can get more done. This is why GPU’s have been at the forefront of Generative AI (Artificial Intelligence) technology.

From a server perspective, it’s the amount of processing power that we have now as a result in this technology shift. Next generation servers are now designed for AI and Machine Learning (ML), and because of the amount of data out there, these servers need to be able to process and manage a huge amount of data in very, very short periods of time.

To give you a real-world example, I was in an AI meeting at the beginning of the week. This was a Bechtle led call, and we were going through how we were working with one of our partners on a genome sequencing project. Working with our customer (a large healthcare provider), we have been able to cut down a genome sequencing process from 20 days down to 2 hours.

Using, obviously, enormous amounts of customer data and our partners server and storage stack, Bechtle were able to deliver huge time savings to our customer by enabling them to manipulate the data they have collected very quickly.

From a storage perspective, traditionally customers tend to rely on their existing, ageing, slow storage footprint. As a recent example, one of our customers asked for high-performance servers as part of their infrastructure estate refresh. However, they still wanted to utilize their existing storage array that is 7 years old.

What customers sometimes don't realize is that the performance of their infrastructure is only as good as its weakest link. This applies to storage as well as compute; if either is slow, it won't be able to meet demand.

Ok, lets stay with storage for a second. So, with data lakes right, it's exactly what it says. It's a massive lake of data where customers can apply machine learning (ML) to large language models (LLMs) and extract valuable insights from their data.

Now, a data lake relies on underlying storage hardware platforms for its foundation. Modern technological advancements and efficiency improvements, such as compression, compaction, and deduplication, are making data lakes more efficient. Because of these efficiencies, customers can reduce the size of their data lake and factor this into their infrastructure strategy.

Cloud connected is also very important. Sometimes you might need to burst out into the cloud. Let’s say a retail customer has got an on-premise solution but then ‘Black Friday’ is approaching. They might need to temporarily expand their on-premise infrastructure into the cloud to enable them to burst (and keep up with demand) for that period. Solutions like this are easy to turn on and off and manage from a single management framework.

We now live in an ‘adaptive cloud’ world. This is a slightly different approach to cloud computing. It’s the ability to manage multiple clouds, private clouds, edge devices and IOT with familiar cloud tools.

In the solutions we design for our customers, a single pane of glass manages all those pieces of the customers infrastructure in one place.

This solves the problem when you want to manage everything from the cloud, but you have geographically dispersed technology stacks, in other words, you're making the cloud ‘adapt’ to your needs and work for you (unified operations, centralised management, true scalability and flexibility), and across all public cloud hyperscalers.

One recent example is one of our customers who were heavy users of Azure and MS365. However, this wasn’t suitable for a data explosion they were experiencing. By enabling them to deploy an adaptive cloud model, they were able to use Google Cloud to host and manage a data lake, continue to use Azure for their office needs and still have centralised management of both clouds.

It’s simplicity and ease of use, but what you must do, the key, is really prepare and plan.

IT leaders should have a very good understanding of what the business strategy and the objectives of the business over the next two to three years. They should know what the objectives are and determine what they need to do to realise those objectives.

It’s no longer a case of “we have this budget what can we buy?”, it must be the other way around.

This allows our customers to “buy small” at the outset and then grow. Scale up and scale out - it’s a new way of thinking about IT. Gone are the days of siloed infrastructure.