Lightbend’s Akka Platform Facilitates Serverless Architecture
Today, we are excited and honoured to welcome Jeremy Pollock to The Prime View. Jeremy is VP of Product Management&Developer Platform at Lightbend. From speaking with Jeremy, we can say he sits in the top tier echelon of truly special technologists with a combination of intelligence, passion and focus to build a game changing technology. Prior to joining Lightbend, Jeremy led product development at PubNub and before that at Mashery, a company acquired by TIBCO.
Jeremy, can you tell us a bit about the steps in your career and how your career has led you to this point?
Looking back on it now, I see a common thread: building platform products that help people build things. My Product Management career has been rooted in trying to understand what people, companies, and individuals are trying to build, what problems they are trying to solve via tooling, apps and infrastructure, and then be a part of building the solution for them.
Further reflection, for me, points to a significant inflection point: my time at API management pioneer, Mashery. Before then, the platforms that I managed were, while used by developers, had very targeted business purposes: e-commerce, content management, personalization. Starting with Mashery, and for the 10+ years since, the problem space, that I play in, is very much rooted in more broad, horizontal developer problems: enabling them to be more productive, faster, cheaper, better, through general-purpose building blocks: from securing and scaling APIs to integrating high-speed data messaging into their applications to developing and deploying serverless APIs and microservices.
Throughout your career, you’ve experienced a number of acquisitions. I’m wondering how the acquisition process (especially, when the buyer is a large company like Intel and TIBCO) affects the work of the Product Manager Leader? What were your lessons learned?
First, get vision and strategic alignment ASAP. Hindsight is 20/20 but a lack of alignment around where are we going, what problems are we trying to solve, how do we approach solving them, at minimum hampered our efforts at Mashery, which was acquired by Intel back in 2013 and later sold to its current owner, TIBCO.
At the time back in 2013, we had the right position in the API management market – now dominated by the SaleForce’s Mulesoft, Google’s Apigee and the OSS product, Kong – and really good momentum with customers (Starbucks, ESPN, Netflix, among many others). And this isn’t to place blame; but back then, before entering a multi-year period of confusion, ever-changing priorities and unmotivated teams, I wish that we had sat down with our acquirers and hammered down this alignment around the big picture and how Mashery was supposed to fit into it. Clear goals. Clear purpose. If you don’t have that, how do you define your own product strategy? Motivate teams? Win in the market?
Second, continuing the theme of strategic alignment: get clarity on planned new investment and be sure to push back aggressively. Are we losing people? Free to hire? Expand to additional locations? I know it is easy to say; I just saw too much nebulous direction in this area. And we broke up teams, scattered teams on to different projects, brought them back, retained people with large packages only to have them drift. Vision and strategy are important of course. But the execution is equally critical, if not more so, post-acquisition. Great planning is required!
The third lesson, and even if you’ve got the great planning in place, with the right investment, communication is absolutely required for keeping people and teams aligned and to help get them over the motivation humps that will always emerge to slow or block progress. I consider myself a good communicator and I firmly appreciate the value of communication. But I look back on my acquisition time and recognize that I could have done better.
What type of individuals are you looking to bring to your company to establish the innovation culture?
This goes back to that career progression and focusing on developers and their pains and opportunities for the last decade. The spirit of wanting to tackle problems in new and interesting ways. Ceaseless desire to solve really hard problems. How could we ever build products for developers if we truly can’t get ourselves into the same mindset? And for me “build” is defined as “develop creative and valuable solutions to hard problems”. Yes, the basics of building – getting stuff done, ticking items off my backlog – are important. But it is the desire to find and develop super valuable solutions, perhaps for problems people and companies didn’t even foresee as being required or possible, is at the heart of the folks that I like to bring on to my teams.
Could you give our listeners a quick background on your current company, Lightbend, and the product you offer, Akka Serverless?
Most people have not heard of Lightbend. But most of us have been touched by technologies powered in part by its Open Source project, Akka. Buying a coffee at Starbucks? Akka is playing a big role there. Tuning into Disney Plus? Ditto. If you’re a cricket fan and following the big events, like IPL, over in India, there are parts of the technological experiences that are powered by Akka. The company and this open-source project – both have been in existence for 10 years – have had as its central purpose the enabling of developers to build massively scalable, extremely fast applications. Another good example – and I bring this up because I just recently went through this process myself – is the auto lending process for CapitalOne. What once took days, of gathering personal financial data, running that through sophistical Machine learning models and ultimately spitting out a result – new loan or denied!- now takes minutes. But what has taught us over the past decade of working with such great companies like Starbucks, Disney, Dream11, CapitalOne among many others, is that building these applications is tough, even with the help of great tech like Akka. That’s where Akka Serverless comes into play.
Akka Serverless gets rid of the inherent complexity in building business-critical apps that can scale to great heights and deliver near-instantaneous performance and keeps the focus of building on only what the developer needs to worry about: codifying business requirements and shipping this into production.
Sure you can do great things with other incredible serverless products like AWS’s Lambda, but we’re not focused on that area. We believe that the modern enterprise dev team is looking to ditch their servers – from databases to message brokers to caching infra – and simply build their APIs and microservices. That’s where Akka Serverless can help really accelerate roadmaps.
Many companies are pushing serverless architecture, including Cloud giants, Microsoft and Amazon. Does it mean serverless architecture is mainstream now?
I hope so, although I’ve seen recent articles touting the decline of serverless. But those aren’t – I think – expansive enough in the definition of what serverless is now. The world has grown up with serverless as defined by the Functions as a Service capability of AWS Lambda. But now you have more products – like Serverless databases from Yugabyte, Planetscale, FaunaDB and serverless messaging tech like Confluent’s Kafka serverless offering – entering into the market, enabling development teams to focus simply on the business need that they’re trying to meet, as opposed to all of the ancillary but important and often complicated infrastructure.
I believe, in the emergence of serverless into the mainstream, is in the inclusion of functions-as-a-service-like capabilities into products like Twilio, Cloudflare, GatsbyJS and Vercel, among many others. It is hard to debate: why would you want to worry about underlying infrastructure? If you’ve ever built something and then struggled to deploy and manage it, then you intimately know the value of serverless. And if you work with the cloud giants, and companies like Planetscale, Twilio and Cloudflare, and their serverless offerings, you can easily see where the reality is quickly meeting the promise. I’d say that’s also why I’m really excited about Akka Serverless: it is another serverless tool to tackle specific developer pains and expands the scope of applications that can be built in the serverless world.
I looked at the cases studies on your website and see that some large companies are choosing the Akka Platform due to its data streaming capabilities. Can you share with our listeners some interesting cases of such kind?
Data streaming is just one of the uses of the Akka Platform. And to be clear for those following the space, the Akka Platform is not Kafka. It is not Pulsar. Both of which are great streaming data platforms as well. But we do exist and deliver great value! Very large companies and business-critical apps have Akka in place because of its ability to enable near-instantaneous decision making and action when it is required. People often talk about data at rest and data in motion. We like to think of data in action.
That CapitalOne use case of making very quick decisions on loan approval, even when executing dozens of machine learning models in parallel. Starbucks, Norwegian Cruise Lines, and several banking customers leverage the same tech to make very quick decisions on personalization. Unique offers based on prior behaviours and what’s happening right now, for the individual and across various cohorts. Approaching these problems with more traditional streaming data platforms doesn’t yield the critical millisecond latencies that make or break the user experience. Your kids are going crazy during the holiday break and you’re looking for Disney+ to calm them down? What’s next, what could they watch, what are the shows, movies that are options for placating the masses? No one’s going to die but getting them into a show fast is supercritical for a parent! For this and so many similar use cases, the Akka Platform’s streaming data capabilities are all about making sure that the right data is where it needs to be in order to make a decision, take an action in the fastest possible way. That’s what we’ve optimized for: keeping the streaming data and application logic tied to the hip, if you will, co-located, in another way to describe, so that it can happen immediately, for that insurance app to be approved or denied, for a personalized offer to be presented on your next cruise with NCL, for the personalized viewing experience of Disney+.