When I invited Tim Wagner to our recent “State of Serverless in 2020” webinar, I was very surprised and excited to hear he was working on something new, but at the time it was still in stealth mode. Considering the vision he created for serverless, I immediately thought that this will be the next big thing for the cloud enthusiasts. Shruthi Rao, former Head of Blockchain at AWS, and Tim Wagner, widely regarded as the “Father of the Serverless movement”, inventor, and the first GM of Serverless at AWS, united their expertise and experience to establish Vendia. Like anyone else, I was very curious about what Vendia is about to deliver and started pinging Tim and Shruthi with my endless questions.
We realized this could be an opportunity to inform anyone who’s as curious as we were to ask questions of Vendia’s founders and ran a short survey to gather additional questions from the community. In this post, Shruthi and Tim answer the questions we gathered and share more about their vision and plans for Vendia.
1. Very cool name - how did you decide on “Vendia”, and what does it represent?
Vendia is short for, “Venn Diagram”, reflecting our focus on sharing and on bringing together the best of what’s come before. Companies, data, and people are strongest when they work together effectively and harmoniously, while also retaining their individuality. Sameness and the “least common denominator” approaches don’t work for people or companies!
The idea of Venn Diagrams is also reflective of our approach. When we first started looking at the dispersed data problem, we realized that incumbent solutions - both APIs and legacy blockchains - weren’t helping customers succeed. We needed something that took the tenets of Blockchains such as immutability, lineage and cryptographic security and combined them with easy to deploy, scalable and flexible cloud-based architectures - the strong point of Serverless. Serverless + Blockchains represents the best of both worlds - our “peanut-butter and jelly” combo that helps customers solve their dispersed data challenges.
2. Serverless has grown from its roots in small, event-based tasks into a serious enterprise tool. But it’s still missing critical support for stateful computing. How will Vendia address the “missing piece” of Serverless?
Serverless is a powerful paradigm - before Lambda, doing anything in the cloud meant having a team of distributed systems experts and “around-the-clock” operational coverage...even for relatively simple tasks. Even today, it’s not uncommon for developers to spend more time getting servers and containers to scale effectively and efficiently than they do on the actual business logic.
But “out-of-the-box” Serverless offerings from cloud vendors are still very incomplete - stateful, cross-region, cross-account, cross-company, and cross-cloud applications are the norm for enterprises. They need a platform that reaches all of their data and all of their code...in whatever region, account, cloud, and company it happens to live. Vendia helps companies retain the best of Serverless applications, including their scalability and high availability, but adds a built-in high-speed, fully replicated data store that enables stateful computing and cross-cloud access. Vendia’s serverless native platform makes that possible by delivering instant scalability and secure data sharing across multiple clouds…at a fraction of the cost and complexity of traditional solutions.
3. I can imagine sharing data through Vendia’s built-in distributed ledger, but what do you mean by sharing code between cloud vendors?
Today, customers have to make unpalatable choices. Want the most scalable and feature-rich serverless function? Hope you wanted to run on AWS. Want unfettered portability? Hope you love complexity and the least-common-denominator approach of Kubernetes. Want guaranteed equivalency of code execution? Hope you’re content sharing the few transactions per second Ethereum can manage with everyone else in the world.
Vendia takes a more pragmatic approach to customer needs. Our tenets are that (1) Customers are inherently multi-cloud, (2) Workloads span not just clouds but companies, orgs, regions, and tech stacks, and (3) Customers want it all - the best-of-breed services on every cloud and the ability to use their own code, wherever it happens to live today.
We help customers connect those dots, getting their data to their code (often in the form of cloud services they want to use), offering to run cloud-independent code on their data model without caring where it runs, and leveraging their historical investments in APIs, including on-prem services.
4. When I’m talking to a person from Kubernetes world, they always criticize serverless because it creates vendor lock-in by its very nature. Will Vendia solve the vendor-lock problem? If so, how?
From our experience running Serverless at AWS, vendor lock-in is probably the single biggest concern customers express. Even though AWS has a history of lowering prices, “bad behavior” by other vendors has scarred an entire generation of IT professionals.
To avoid ‘vendor lock-in’ Kubernetes demands that you take on vast amounts of unnecessary, undifferentiated work. What’s worse, workloads of the future - such as smart mobility, autonomous things, predictive replenishment, etc. - create even more complexity for server- and container-based approaches, as they span clouds, edge, and on-premise data centers. Least common denominator services that ignore the best and brightest cloud offerings, unfettered server-based complexity, and endless porting challenges are a high price to pay to avoid vendor lock-in with Kubernetes.
We realized that it’s important to address the real challenge here: Customers don’t want to port their code, they want to know that they’re not locked into a single platform. They want solutions that are multi-cloud, not “theoretically portable among clouds”. Vendia embraces this ideal, by creating cross-cloud code and data solutions that enable customers to work with cloud services on Azure, AWS, Google, IBM, and other clouds as well as on-prem data centers...without porting and without dealing with server- and container-based complexity. Want to ingest IoT data on Azure, process it with a Lambda function on AWS, and store it on Google cloud to perform advanced analytics? Vendia is there for you. With a strongly consistent data model, SaaS-based management, and access to serverless capabilities on every cloud, Vendia can offer the best, rather than the worst, of cloud computing across vendors.
5. What metrics are you aiming to fix in application teams? Developer productivity, time-to-market, MTTR, security concerns? All of the above? Can you elaborate more on the value that Vendia provides?
We think of this in three waves: The Day 1 problems, the Day 100 problems, and the Day 1000 problems.
The Day 1 problems are all about productivity and time-to-market. We’ve seen customers struggle with deployments of Hyperledger Fabric, DIY solutions, and old-school ERP solutions that can take months, even years, and sometimes collapse under the weight of their complexity and cost. With Vendia, you can turn a data model into a multi-account, multi-cloud, multi-region, multi-party solution in about 5 minutes...as fast and as easily as you can create a regular database table. Our serverless architecture helps customers demonstrate value quickly because we can deploy rapidly and without operational burden. In the time it takes to configure a Kubernetes cluster or set up networking on a VPN, you can be delivering value with Vendia.
The Day 100 problems are all about scale and cost. Can you process the business load that real enterprises actually face? Can you avoid charging customers when they’re not doing useful work? Again, our serverless heritage helps, as we can scale instantly to zero and then back up to address fluctuating workloads, all while “shrink wrapping” costs on a per-transaction basis. Existing blockchain technologies fail dramatically here, given their “single box” deployment architectures.
Finally, the Day 1000 problems are all about evolution. Can you add clouds, orgs, and regions seamlessly? Can you evolve your data model without having to recode your existing applications or replumb your infrastructure? Having thought through these challenges before customers encounter them is part of our strength. Here is where DIY solutions become increasingly costly and complicated, versus a schema-driven approach that can incorporate the learnings of many customer scenarios.
6. Do you see any competition? Or are you just creating a new market? Do you expect new competitors?
Multi-party data aggregation and transparency is not a new problem...the Romans were maintaining distributed ledgers of grain production and army deployments...computers just made it possible to do it better. What’s changed just in the last few years is the volume: IoT and mobile mean we’re producing more data than ever before, while AI/ML and analytics have enterprises consuming and processing ever more data. Companies need easy, flexible, and scalable mechanisms to aggregate these disparate sources together in real time so their ML and analytics solutions can produce usable business intelligence.
Historically, large enterprises looked to expensive ERP systems to help them with these challenges. More modern solutions, including most of the “born in the cloud” approaches, are DIY implementations based on point-to-point APIs. APIs are relatively easy to set up initially, but accrete complexity over time - as “dumb pipes”, they often fail to deliver on consistency and timeless outcomes, and the need to repeatedly poll can waste data transfer capacity and drive up costs. DIY APIs are the most used alternative to Vendia today, but companies often struggle to successfully connect multiple parties in multiple clouds at scale with the homegrown approach.
On the other end of the spectrum are blockchains. With a promise to seamlessly connect and share data and code, they seemed like an ideal solution for enterprises tired of struggling with data silos. Unfortunately the “single machine” deployment architecture of blockchains makes them ill suited to the enterprise - their high latency, low throughput, deployment complexity, specialized languages, and difficult enterprise integration caused many early adopters to simply give up.
We had the opportunity to build and run both Serverless and Blockchain businesses at AWS and Coinbase over the last few years. We got to see, firsthand, the root cause of these problems and the challenges that customers face with both DIY solutions and attempts to adopt blockchains. With over 1,000 customer meetings and thousands of hours of operational exposure, we’ve developed a deep understanding of what enterprises need and used our insight to create a simpler, more effective solution.
7. Congrats on the new seed round, now you have fuel to grow! What are your short and medium-term plans? Are you hiring? What are you looking for in your first few hires?
Thank you, we are very excited and thankful to have such incredible investors who share our convictions about this space! The next stage of our journey is to staff our seed round positions and deliver our initial product release. Interest in Vendia has been high and we’ve filled all our open roles for the moment, but we’re always happy to chat with interested folks and are already looking forward to our Series A when we can grow the team again. You can always check out jobs.vendia.net to see any roles we have open.
All our early hires share some common characteristics - including a deep expertise and interest in Serverless, distributed architectures, and cloud computing. Both developers and Solution Architects at Vendia have to also be passionate about customer success and the goal of making it easy for companies to develop smart mobility and distributed applications. We also have a strong focus on growing a healthy culture: We started Vendia not only because of the exciting problem space, market opportunity, and the innovation it affords but also because we love working with each other. We work hard, hustle, and build every day, with no room for toxicity or negativity. To maintain this sacred space for building, we have a strict “Kind Humans Only Policy” that we require every hire to sign and adhere to. We’re both former Amazonians, so we also embrace Amazon Leadership Principles such as Ownership, Customer Obsession, and Bias for Action, but we realize those “business” goals have to exist in the context of supporting each other. Our Kind Human policy is the missing glue that makes Amazon-style leadership principles humane and healthy...because we can all use a little empathy, love and belonging in our workplaces, where we spend about a third of our lives.
8. Who are your first customers? How would you describe their use cases and the benefits they receive by using Vendia? What is the “pitch” to companies who might not realize they have a need for Vendia’s offerings?
We are currently working with 8 companies in either a design partner or early customer capacity. They range in size and market cap - from large auto manufacturers and airline companies to startups. But their use case and problem is the same - they all need to aggregate data from multiple partners on multiple clouds or on-prem data centers in real time. They and their partners inevitably all require fine-grained control over the data they share, along with the ability to update and delete it. And finally they and their business partners need assurance - the ability to prove that the right data has been shared with the right people at the right point in time.
Let’s take a look at some specifics: One of our startup customers aggregates data from approximately 64,000 data sources including traffic, weather, signal status of traffic lights, and public camera feeds. They augment these data streams with connected car data, enabling them to run advanced predictive algorithms that establish propensity of fault in vehicular accidents. These analytics are then shared with auto insurance company adjusters. There are multiple layers of challenges here. First, the data sources are disparate - some on cloud, some on-prem, some with APIs but others without. The data contains many duplicates, has varying degrees of freshness, and lives in every conceivable place - it’s not “just on AWS” or “just on Azure”. Some sources, such as weather data, are monetized, so the owners need assurance that all users have appropriately compensated them for access, an especially important concern in any multi-party data sharing scenario. With Vendia Virtual Datalake, this startup is now able to seamlessly connect to all data sources, while only retrieving data from a source when it’s relevant to their business, such as restricting geographically-correlated data to be near the scene of an accident. Vendia also provides the lineage and control needed to track monetized access, delete and update data that changes in real time, and enable data source evolution as information is added or removed over time. This startup formerly spent 60% of their total efforts on data aggregation, a challenge they were able to offload to Vendia. Now they can process all of their data in a single, continuously monitored S3 bucket, with Vendia performing the “heavy lifting” of tracking lineage and maintaining data integrity over time.
For those customers we haven’t met yet, our pitch is a simple one: Your data is everywhere you are. You want the best of the cloud, not its lowest common denominator. We can help your business span partners, clouds, regions, and accounts.
9. What is the typical developer experience using Vendia to create their first Serverless, distributed apps or “Unis”? Is there a particular language or skill required?
We’ve made it super simple to get started. All you need is a JSON schema that describes your data model - it can be as simple as a single string if that’s all you need to share! You also decide how many participants (nodes) you want, and what region, account, and owner they have. That’s it - with a couple of simple JSON files and about 5 minutes you can create a scalable, secure, production-grade multi-node network with full GraphQL support for web and mobile apps. You can create applications in your choice of languages, and no special skills (such as learning Solidity) are required.
10. When Lambda was first launched, we saw that developers used it for cron jobs and then for batch processing. From there it grew to encompass use cases that no one would have predicted early on, such as high-speed video transcoding. What are some of the things developers can Unis for out of the gate? How do you see the innovation flywheel starting?
Unis are a super simple way to create applications that need to run in “more than one place”. Ingesting data in different regions or accounts and then consolidating or processing it usually requires a complicated ETL process or data streaming architecture. With Unis, you can set that up in a few minutes if your data already lives in an AWS service. Any “multi-primary” system is a good candidate for consideration, and the fact that it comes with a built-in AppSync makes it easy to wire up web and mobile applications quickly.
11. What attributes of “classic” blockchains does Vendia plan to retain? Is Vendia still a good fit where there is no need for distributed ledger?
Many of our customers neither know nor care that there’s a distributed ledger running “under the hood” - they just produce and consume data (or files or messages) and leave the implementation details to us. But under that hood are all the things you’d recognize from a formal blockchain: a distributed (fully decentralized) ledger with fully replicated, totally ordered, ACID-semantics transactions, tamperproofing through hashes, and a shared “world state” database. Consensus is carried out through a combination of highly scalable and parallelized serverless components, providing the same “mutually trustless” semantics of a conventional blockchain but at a much, much higher rate of throughput and with substantially lower latency.
But distributed ledgers are only one of the elements we provide. Customers also benefit from our “batteries included” approach of creating a customized AppSync interface. Vendia offers message passing, ledgers, replicated databases, cross-region blob and streaming data support, a simple way to create typed GraphQL solutions … in short, we make modern apps easy to build and even easier to operate!
12. Are you considering an open source offering? Would you consider a CNCF-style relationship or other means for opening the interfaces while allowing differentiation through implementation? What partner ecosystem do you foresee in a Vendia world?
We believe in developers coming together...it’s part of the inspiration for our name! A strong, vibrant ecosystem is critical for any platform to both grow and thrive. We’re committed to open solutions that make it easy for anyone - commercial, nonprofit, or individual - to bring additional data producers and consumers, optimized solutions such as specialized indexers, and higher-level application constructs and tools into Vendia networks. It’s still early days but we look forward to working within a framework such as the CNCF when the time is right to help create an open and transparent process for that. We’re also excited about open source at Vendia, and our first major release will be our serverless CDK deployer, which will be out by the end of 2020.
13. What are your predictions for the future? Will Vendia affect the adoption of Serverless architectures and/or distributed ledgers?
Our goal is to help customers deliver solutions while ignoring boundaries - geographic boundaries across regions, the “walled garden” of cloud service providers, the gap between different companies and even departments within the same company, and so forth. When customers can write applications that “run everywhere” without the need for porting or managing servers, we’ll know we’ve done our job!
At the same time, we believe it’s critical that customers get business benefits, not that they buy into specific technologies. Many of our customers aren’t Serverless experts and some of them don’t have any experience with distributed ledgers. That’s ok - to them, and to us, those are implementation methodologies. They provide important benefits, but it’s like a page lock in an Oracle database - critical to correct operation, but not necessarily something a customer needs to understand in detail in order to benefit from.
14. The community wants to play with Unis! When should we expect Vendia to be open for preview access???
Vendia Share developer preview will be open on Sept 15th, when you will be able to develop your first Unis with initial support for AWS. Visit https://www.vendia.net/ to sign up for free preview access!