Category Archives: Database

London Technology Week 2015

I cant believe it has already been a year since the last London Tech Week! As usual, my company generously gave me a day off work to explore and broaden my mind to the latest tech trends. Lets see what the London Technology Week 2015 flagship event had to offer.

Out with the old, In with the new

One of the really noticeable difference this year was the lack of 3D printing – I failed to find a single stand! Last year I couldn’t turn a corner without bumping into 3D printing machines, they seemed to have gone out of fashion completely this year. Instead every other stand was talking about IoT or more specifically how their product is online and can be controlled via smartphones or talk to other equally smart devices to work together seamlessly.

Now I wouldn’t say IoT is a particularly new concept, but I would say that 2015 is the year where it entered the mainstream language. Previously, you would probably buy a new washing machine based on its ability to clean clothes whilst using minimal amount of energy and power. Today, if the washing machine cant be remotely turned on from your smart phone, you will probably look past everything else it can do – this is of course a little tongue in cheek, but not far off the truth!

While there were many excellent IoT themed products on show, two catch my attention, not because they were the most advanced but because they attempt to solve two issues close to my heart.

1. Keeping track of small children

The first is Lineable – a smart band your children can wear to help alert parents if their child wanders off too far from them or track them down when lost. What is really exciting about this project is that it was crowdfunded (here) – and now it is using the power of the crowd again (see below) to help track down lost children!

Lineable wristband

The concept is very simple, the band will connect to the parent’s smartphone via bluetooth – if the child moves out of range, your phone will alert you. They claim this is approximately 20-30 meters, but will decrease depending on environment.


If your child is lost, the band will also send out a unique identifier to all phones with the lineable app installed. This information will be uploaded to the central servers and can be accessed by the parent of the child. The success of this will then of course rely on the number of people who download this app and have bluetooth turned on all the time – hence power of the crowd. While this is a massive short coming, you have to applaud the makers for coming up with such a wonderful initial prototype. The band itself is priced at only $5 so should be no brainer for any parent. The app itself will be free so even if you are not a parent you could still download it and help in the movement.

2. Keeping heating costs down!

The second IoT product I wanted to call out was openTRV. Some of you will know I am  passionate about keeping my energy costs down having recently invested in a NEST. However, my current NEST setup is only able to sense the temperature of one room in my property. This means I still have to manually adjust the valve control of each radiator individually as some rooms are hotter/cooler than others. openTRV attempts to solve this problem by having a smart radiator valve attached to each individual radiator. Each Valve will attempt to learn the usage pattern of that particular room and adjust temperature accordingly. Each Smart Valve uses a couple of standard AA batteries and will automatically open or shut the flow of hot water to its respective radiator.

One of the founders of openTRV with their prototypes

Currently their prototype feels early stage and they are finalising the end product.  I think the perfect end goal for them will be when they are plugged into the open APIs from all the major players (NEST, Hive etc) and push additional information to the central boiler control – i.e. these devices essentially act as additional input sensors for NEST. Their target price will be £10 for each valve, which would cost me £70 for all valves to be replaced at home. This would seem like a no brainer if their claimed savings of upto £300 a year is achieved, although I would think this is much less if you already have a NEST/Hive solution, but still cheap enough to be worth a punt.

NEST security camera

It will be great if we get to a stage where every device in the house is able to communicate to each other via a set of open standards. Just today, NEST announced their various products will be able to talk to each other – i.e. the fire alarm will inform the thermostat to shut down the boiler if it detects carbon monoxide. In addition, they will also be bringing out a security camera which will hopefully leverage the thermostats motion detection! To that extend, I hope major players like Google/NEST can help drive open standards for IoT in the home and allow other devices to work with their in house devices. Lets all hope they are able to do this before Apple gets here and try to lock us into their ecosystem even more! (Apple fanbois incoming 3, 2, 1 ..)

Big Data needs Big Databases

There were many many talks about Big Data and Big Data Scientists at various London Technology Week 2015 events. But I wont go into these here, mainly because I am bored to death talking about about how cool Big Data is and how it will enlighten you to all of life’s mysteries. Instead I have recently grown a curious interest in large distributed databases. Ofcourse you cant talk about this subject without bringing up Hadoop. My limited interactions so far have been through Hive which is the de-facto tool for Hadoop.

At the event I came across two other Database technologies, Tajo which is an Hive alternative and Brytlyt which offloads compute-intensive portions of the application to the GPU (bitcoin mining anyone?).

1. Apache Tajo and Hadoop

Tajo is an alternative to Hive – in fact, until today it hadn’t even occurred to me why I would need an alternative to Hive. It feels like I have opened Pandora’s box and now I feel obliged to rummage through all the Hadoop tools.

Tajo Logo

I found a good blog on the various SQL-on-Hadoop frameworks, it is definitely worth a read if you have time. In summary, whilst Hive is easy to setup and use, it is widely accepted to be slow due to the overheads of using MapReduce. Various alternative exists with varied degrees of enterprise support. Tajo is one of these and claims to be upto 10 times faster than Hive. Although as with all metrics, probably needs to be taken with a pinch of salt as different use cases will result in varied performance outputs.

Although Tajo is open source, it is mainly developed and supported by Gruter (who were at the show to introduce me to Tajo today). One of the major disadvantages mentioned in the above blog about Tajo is the lack of community and support outside of its native South Korea. In all fairness, this was very much aligned with my own experience today when I tried to speak with the representatives from Gruter. They hardly spoke a word of English and after 10mins of trying to get them to explain Tajo to me, the only thing I had learnt was that it was meant to be a faster alternative to Hive (I guess they got the main message across).

2. Brytlyt – GPU

Brytlyt produces a product called MakuluXtrm, its USP being its ability to offload compute-intensive portions of the application to the GPU, thus massively increasing performance of queries. Unfortunately they didn’t have any live demos available at the show, but the concept itself has a lot of potential. Given the amount of attention GPUs have been getting in recently years I can easily see how this could work if applied well.

Brytlyt architecture
Brytlyt architecture

Brytlyt’s representatives claim some impressive stats, which I wont quote here as I couldn’t find these claims printed on their website. Understandably they seem to be guarding their technology and giving out minimal information online, however this could also be their ultimate down fall as they are competing in a vast field where everyone is going open source. As a new entry into this field, I think they could do just as well if not better were they to open source their technology and offered their services to support enterprises make use of it instead.

Techlash (aka Technology backlash)

Finally I want to mention a very interesting talk given by Lord Wei of Shoreditch about a recent trend we are seeing of back lash against new technologies. The issue is that while Technology is improving society as a whole, there are vast amounts of people who are directly affected and often left behind. In the U.K., 35 percent of jobs could be replaced by automation in the next 20 years, a Deloitte study showed. That figure rises to 47 percent in the U.S., according to researchers at Oxford University.

A recent example of Techlash is Ububerer. Over an extremely small period of time it was able to disrupt the livelihood of cab drivers around the world. Hold on I hear you ask, didn’t Uber also create thousands of jobs? Indeed they did, in fact not just Uber, but generally it has been estimated that for every new Technology position filled, 7 other jobs are created on the back of that. While that is indeed true, you still have a large set of people left behind when a new technology comes along. We need to help those people transition or else resentment will set in and society as a whole will suffer as people find themselves out of work and relying on the state.

A real example of this happening close to home is the modernisation programme on the underground (UK), where the government is met with fierce opposition from the unions about driver-less trains and also closing of thtubee ticket offices. In this case, the transition plan offered is to retrain a lot of tube staff to be visible on the platforms helping passengers instead of behind desks. Inevitably new technology often means you need less people to do the same job. However Lord Wei argues that this doesn’t mean there will be less and less jobs, while technology means you may require less people in certain roles, it will also help create completely new labour markets which didn’t exist before.

Lastly, Lord Wei mentioned of some of the conversations he had with Google employees recently in California where they admitted in certain parts of the city they would not reveal where they work for fear of being beaten up. In some parts of those cities, the way the trend is going, technologist will be more hated than bankers! Its just as well we are not technologists working in banking – oh wait.

That’s a wrap, thanks!

Finally, I leave you with a random selection of photos from the event, see you again next year.


London Technology Week 2014

This week I was lucky enough to be able to take some time off work and attend the London Technology Week. This is a brief write up of the talks I attended and photos of some cool exhibits. Lets start with the talks

When smart-phones sense how you feel: The era of intelligent mobile devices

Talk description: “Argus Labs uses deep learning algorithms to sense, understand and predict human behaviour and emotions, based on the sensors in a smart-phone and general usage of a smart-phone. The presentation will demonstrate how smart-phones will start to behave as intelligent entities that know how a user feels and improve our lives.”

  • Argus Labs is working on APIs and SDK which will allow programmers to build emotion aware mobile apps
  • A variety of tools is used to predict mood, based on learning over time – i.e. every Monday morning you feel depressed and you tell the app this is how your feeling and over time it will form that pattern or maybe after everySaturday night after you come out of a club you feel happy. They also make use of the gyroscope to determine how happy / angry you are – for example, you are happy after clubbing and reading a book, but you might be jumping around after clubbing whereas your quietly happy after reading
  • The Presenter mentioned how personality is now really easy to predict – i.e. based on what feeds we subscribe to and what circle of friends we have. However mood is still extremely hard
  • An interesting topic was whether you should act on the user’s current mood or their predicted mood – i.e. if you know someone will be unhappy in an hour’s time, maybe play some happy music now to start counteracting that
  • A good example of this application is if the apps sees the user stuck in traffic and knows the user is typically angry – time to play some gentle music to help calm the user
Will cars be able to detect your mood in the future to make you a better driver?
Will cars be able to detect your mood in the future to make you a better driver?
  • The apps you build will learn about an individual’s mood and behavior over time, but is locked to that person so you cant pass your mobile to another user and expect it to work
  • Future of mobile is not about mobility, its been done – it will be about AI

This was a really interesting talk, made appreciate a lot more about how much companies are able to learn about me from my general online profile and keeping track of the places I visit etc. I start to think that one day my phone will know me better as a person than I do! Also the general idea of data mining and building a reasoning engine behind the massive amounts of data you collect is a theme that runs through many growth areas.

The Database of You – The opportunities and issues that arise from wearable technology everywhere

Talk description: “The last 5 years have seen an explosion in the growth of wearable technology, health and fitness data and mobiles apps and devices. Yet we are still only at the very start of our journey towards the database of me; – a personal cloud containing details of every aspect of our lives. How will the availability of huge quantities of health and performance data on everyone affect how companies do business, how people lead their lives and the general wellness of the population?”

  • Talk was delivered by CIO of Virgin Active Health Clubs and he started by mentioning that attrition rate in health club industry is a massive 60% – i.e. 60% of their members leave every year
  • The way to reduce attrition is to better facilitate users with their goals and allow users to see that they are making progress
  • The presenter mentioned the term “Cambrian explosion”, this is the period when there was a massive explosion of different life forms on earth and before settling on the best design – he sees wearable technology currently going through this period and everything we have not yet seen anything like the final design yet. He compared this to initial stages of the mobile where they came in all different shapes and sizes before everyone making phones that look like an iPhone
Graph showing popularity of a new technology as it comes out and then over time as it matures
Graph showing popularity of a new technology as it comes out and then over time as it matures
  • In terms of collecting data about individuals, he was nervous about how much personal data we are now able collect, but at the same time he felt everyone was starting to feel at ease with their data being in the cloud
  • This is an area which Apple, Samsung and Google are now all heavily investing with different platforms
  • Finally he compared current wearable tech to the early stages of car GPS systems which would plot on the map exactly where you are and the route you took – which was amazing when it first came out – to current GPS systems which also tell you about traffic and weather and best route home. So to make the tech actually useful you need a lot more context and a collective database about the person – The database of you.

Personally I really love the design of the new Samsung Gear – although I do worry slightly about the massive database of myself being held online, does that mean my medical insurance will go up with every McDonalds meal I eat? Maybe it should – but maybe this is just the perfect reason I needed to keep myself healthy as well.

Make sense of your (big) data and analyze in real-time like you have never done before!
Talk description: “With companies like Bloomberg, Facebook, Goldman Sachs and The Guardian using Elasticsearch, discover how the power of search can power your data analysis. We’ll enlighten you with some impressive use-cases for on-the-fly real-time data analysis and share dashboards that will delight you (and your stakeholders) beyond belief.”
Free Elasticsearch USB stick - yup I am easily brought
Free Elasticsearch USB stick – yup I am easily brought
  • Elasticsearch actually mentioned Goldman a few times as one of their star clients
  • Major advantage is its server plug and play capability – by adding new servers, the load will be automatically distributed for you
  • One of the most useful way to use it is not to look for the items with the biggest volumn / ticks / treads, but to look for exceptions in your data – i.e. comparing this week’s and last week’s dataset and seeing which currency is doing more or less relative to last week
  • Everything is open source and available for commercial use for free – Elasticsearch themselves make money by providing training and some APIs and tools
  • Do not use Elasticsearch db as your golden source if the data is important for you, always keep a separate copy and elasticsearch can lose data in certain edge case scenarios
Other cool London Technology Week exhibits