Reliability and scalability critical to mGage SMS gateway service
mGage provides a gateway service for SMS messages and transactions, routing from customers to operators and back again. The global mobile messaging and engagement provider has more than 15 years of mobile CRM experience, with offices around the world including in the UK, USA, and India, and employs around 500 staff worldwide.
mGage has been using Riak KV since 2013 to ensure the reliability and scalability that its message service requires. We sat down with Tom Rees, European Head of Operations at mGage, to see how the company is using the solution, and how its benefits have reached beyond the business.
What are the data processing challenges facing the SMS transactional industry and mGage, in-particular?
Well, the SMS transactional industry is a heavily regulated one, and with good reason. As a mobile messaging company we are held responsible for the messages we coordinate and for ensuring that transactions are executed successfully – if this isn’t handled correctly we can be hit by significant fines by regulatory bodies. With this in mind, it’s absolutely key for us to be able to maintain our performance no matter the volume of data we’re processing – if we experience a surge in transactions, we have to be able to cope with this and ensure all transactions are successful. For example, if someone uses our service to enter an SMS competition and their vote isn’t counted, we’d face a severe backlash from regulators. Often, events that are voted upon, whether it’s a TV-talent show-type vote or another form of competition, happen in real time, meaning that if a vote isn’t processed successfully and the window of entry passes, there is no way to remedy the mistake. This would lead to angry and dissatisfied customers and operators.
A prime example of our work in this space is our management of Vodafone UK’s JustTextGiving, a service where customers text a phrase or word to a certain number in order to donate money to charity. One of these campaigns was last year’s ALS ice bucket challenge, the viral phenomenon that was embraced by millions of people, including a large number of celebrities. Due to the popularity of the ice bucket challenge we saw data volumes go from a few thousand messages in a month to millions in days. We needed a solution that was robust and reactive enough to deal with these spikes in data, while maintaining reliability and performance and ensuring that all those who donated were able to do so unimpeded and that the charity received these donations without delay.
What were you looking for to address these issues, and how did Riak fit the bill?
We realised that in order to cope with peaks in data traffic, caused by campaigns like the ALS ice bucket challenge, without compromising on performance or reliability, we needed a platform that could cope with sporadic but large volumes of data without error. Our service doesn’t constantly generate massive levels of data volume, so we needed a platform that was scalable enough to handle both the peaks and dips in demand without delay. Again, this was an imperative as any delay could impact both our reputation, through poor customer service, and our bottom line, through fines from the relevant regulatory body.
Therefore we needed a distributed database that was capable of capturing high volumes of data. While we were evaluating the market it became clear that Riak KV was best known for its exceptional scalability and speed, and wouldn’t be impacted when data volumes peaked, so we identified it as a potential solution. Of course, we went away and evaluated other solutions but were let down by other database vendors who were unable to meet these needs. With this in mind, we wanted to try something new and turned to Riak KV.
Our faith was well-placed, as Riak KV proved ideal for supporting our mission critical transaction volumes, and flawlessly managed the peaks in data traffic.
How did you deploy Riak KV?
For our deployment of Riak KV we maintained nine ‘live’ nodes in production and one node reserved for testing. This ensured that we had a quality assurance environment where our team could analyse and develop software changes, which would then be deployed after we had run them through our vigorous assessment process. This environment could be used to test data throughput and transactional volumes before deploying the tested methods in a live, mission critical environment.
We also rely upon Riak’s support offerings, which were crucial during our data center migration – an event where downtime is not an option if we are to continue serving our customers. We needed to migrate our live gateway and there could be no errors in database replication. Here, Riak KV was deployed, with the team using a function known as replication error.
We hadn’t used this feature before so there was some trepidation, but Riak’s incredibly helpful technical support staff guided us through the entire process. Riak KV replicated the entire database to our new data center, successfully shut down the applications in the old center, and there was no downtime in terms of processing the traffic.
This was an excellent experience for the team and we were particularly impressed with Riak’s assistance in training and support. We have a preferred contact at Riak who the team speaks with on a monthly basis and this kind of support offers invaluable guidance when implementing software and using features for the first time.
How has mGage benefited since the implementation?
Now that we no longer have to worry about the peaks in our data traffic we can spend more time focusing on innovating and improving our own offerings, instead of dealing with maintenance headaches. Being able to rely on a solution that doesn’t require lengthy maintenance or ‘babysitting’ boosts productivity internally – I wouldn’t be able to quantify to the last second all the time we’ve saved now that we’re not constantly looking for new solutions to deploy, but I imagine it’s a lot.
With Riak KV we can confidently say that we are able to deliver a flawless and reliable experience to our customers while also satisfying the strict criteria of the market’s regulatory bodies.
What is the future for Riak KV and mGage?
We are very happy with our deployment and have realized that by staying up to date on the software version, maintaining good communications with Riak and flagging any issues as soon as they appear there’s very little chance that anything will go wrong, so it’s unlikely that we will change. Instead we intend to extend its use. We’d like to roll out Riak KV globally to help us to streamline our processes and reduce operational and maintenance costs, so this is the next step in our implementation.
