Years before founding Nventify, Jack Levin, was employee number 21 at Google. He spent much of his career scaling huge infrastructure systems, and learned some valuable lessons along the way. Today, Jack uses those lessons and and shares them with the team at his newest company, Nventify. He’s helping companies scale their systems and scale their images as efficiently as possible. In his keynote at the 2016 Annual LDV Vision Summit, Jack exemplified how to scale large systems and not burn yourself, or your infrastructure, out.
Jack’s first day at Google looked like this -- 300 servers, no routers, a single data center and a single internet provider. He actually walked into Google’s data center and wired the whole thing himself. A week later, the company had already outgrown this setup. They brought on one of their first big clients and needed to launch quickly. So, within a week, Google added an additional 2,000 servers. It was a massive undertaking. As a young person starting off his career, Jack shares that he “had no idea what he was supposed to be doing.” One of Google’s founders, Larry Page, handed him a bunch of cables, he plugged them in, and carried on. That story continued in the same fashion for the next couple of years.
Jump forward a few years and Jack was no longer going to the data center on a daily basis. Instead, a team of 25 people managed it. At that point, there were rows upon rows of hardware in place, countless kilowatts being consumed, abundant heat being generated, and droves of queries being served.
But let’s get to the real reason you’re reading. Let’s talk disaster. Jack claims responsibility for “killing Google with my bare hands” more than once during his time there. In full transparency, he didn’t always know what he was doing, pushed the wrong button more than once, and Google would go down. Luckily, at that time, the fix was as easy as Jack jumping on his scooter, running to the data center, unplugging the power supply and plugging it back in to reset his configurations. It took a few of these panicked trips to the data center before Jack found a better way and implemented a dial up to the data center. He could then quickly dial in and undo his work. By putting this system in place, and properly scaling alongside the company, Jack’s work remained manageable and productive. But those sorts of ideas only come with experience, and the stakes are high.
Google’s early growing pains are not at all uncommon. In fact, one of the biggest problems that quickly scaling start-ups experience is a lack of IT and operations knowledge. Many founders are not experts in these fields. They can hire people to run their data center, but most of these people are just like Jack in his first years at Google, they don’t really know what they are doing, or how their decisions are going to impact the business. For a long time at Google, they didn’t even know what would kill the search engine. They were constantly monitoring, but like any early market business, they were unsure of how to interpret results. In the early 2000s, any end user could kill Google simply by sending a query of words that had no meaning when combined together like “theological silhouette.” If you sent five of those queries per second from your laptop, Google’s search engine crashed. Another lesson learned... turns out monitoring queries and spam detection is important for Google!
Jack said it best himself, “sometimes it takes luck, and persistence, and people working more than 9-5 just to get things up and running the right way. But even then, 80% of the time you won’t get it right.” To help you avoid some common operations and IT mistakes, Jack has 3 tips to ensure that scaling remains manageable for your team.
- Think post-mortem AND premortem: You’ll never move forward if you only think about disasters after they happen, you must pre-emptively attack problems. In the early days of Google, there wasn’t a backout plan. But today, Jack makes sure Nventify engineers always have a backout plan to revert their changes. Even if you think you know what you are doing, always have a safety net.
- Accept that you don’t know everything: You are a better engineer if you know how to search for answers and learn from other experts. Most of the building blocks you need are already there. Spend your time focusing on the core business, not re-inventing the wheel. Or, in the case of image scaling use a service, like Nventify’s Imagizer, to compress and resize your images for you so that you can spend more time on the things you do best.
- Don’t be afraid to outsource: When companies scale up, they are often hesitant to buy any tools. They want to create everything in-house. But that is often more expensive, slows progress, and the ability to deliver a product to the end user.
If you are ready to start scaling your business appropriately, but don’t know where to start, image resizing is a logical choice. Grainy images and hard-to-see photos shine a poor light on your great brand. Don’t halt progress by manually resizing images across various devices, mobile apps, and webpages. Imagizer is the perfect solution for startups to easily and dynamically manipulate images in real time, cutting hours of production work. It’ll also scale with your company as it grows. Check out our free demo.