eurekapd Chapter 295 Late-Night Phone Call
Chapter 295 Late-Night Phone Call
March 30, 1998, 11:40 PM.
Lingyun walked out of the development and testing area, and the glass door closed gently behind him.
My phone vibrated in my pocket.
He took it out, the screen lit up, and the caller ID showed a number—area code 408, a Silicon Valley number range, but he didn't recognize it.
connected.
"President Ling." The voice on the other end of the phone was a little hoarse and clearly tired. "This is Jiang Feng."
Lingyun stopped.
"Jiang Feng? Do you have any good news to tell me?"
"Hmm." There was a pause on the other end. "The database is working."
The corridor was quiet.
Where are you now?
"The lab," Jiang Feng said, "just finished the final round of stress testing. It ran continuously for 72 hours without crashing, without data loss, and read/write latency was within the acceptable range."
Ling Yun didn't speak.
"Let me start from the beginning." The sound of pages turning came from the other end. "At the end of January, we completed the final version of the kernel refactoring. We spent the entire month of February doing stress testing—single-machine, cluster, and distributed. At the beginning of March, we started running 72-hour continuous tests, which just ended today."
"What were the test results?"
"A single machine can achieve a QPS of 8,000, and the cluster scales linearly, reaching 23,000 with three machines. The average read/write latency is 12 milliseconds, with a peak of no more than 30 milliseconds. All data consistency tests have passed, and the average fault recovery time is 47 seconds."
Lingyun pushed open the door to the stairwell and walked to the second-floor corridor.
"How much did it cost?" he asked.
"So far, it's been twelve million," Jiang Feng said. "Mainly servers and people. We bought eighty-seven servers to build a test cluster, and the number of engineers has expanded from the initial five to ninety-one. Salaries, equipment, and server room rent add up to this amount."
"If we can make it, all the investment will be worthwhile."
There was a two-second silence on the other end of the phone.
"What's next? What should I do next?" Jiang Feng asked.
Lingyun walked to the window.
"Fly back to China next week, I'll be waiting for you in Jinan," he said. "Bring the seven core members of the main team back with you. The rest of you will stay in Silicon Valley to continue optimizing the business."
"Go back to China?"
"Yes. We'll establish a new R&D center in Jinan," Ling Yun said. "Ni Guangnan has existing office space; you and Wang Jianguo's operating system team can share a floor."
"And what about Silicon Valley..."
"Leave one technical lead to lead the rest of the team in continuing the iteration. This includes version updates, customer support, and liaison with Xingyu," Lingyun said. "You and the people you brought back will form a new team."
"What direction should we take?"
Ling Yun did not answer.
"President Ling?" Jiang Feng's voice came through the receiver.
"You'll find out when you get back," Lingyun said. "Let me know once you've booked your plane ticket, and I'll pick you up at the airport."
"Okay," Jiang Feng said.
"Don't tell anyone about the database yet," Ling Yun said, "not even the team members. We'll talk about it when you get back."
"clear."
"Make sure the handover is clear in Silicon Valley before you leave. Don't rush, and don't overlook anything."
"clear."
"Also," Ling Yun paused, "you've had a tough year and a half."
There was silence for a few seconds on the other end of the phone.
"All the hard work was worth it," Jiang Feng said.
After hanging up the phone, Lingyun stood still.
He recalled Jiang Feng's first visit to Los Angeles a year and a half ago.
At that time, he was a young man in his early twenties, who had recently graduated from the University of Science and Technology of China. He was writing an internet cafe management system for Xinghuo Internet Cafe and had built a rudimentary member database. Lingyun picked him up from the airport and took him directly to the hotel, not letting him show up at the company.
"I'm giving you a task," Ling Yun said at the time. "Develop a lightweight database based on the Star System. You'll be the CEO of the new company, and you can recruit your own team."
Jiang Feng didn't ask anything, he just nodded.
It took a year and a half and cost $12 million. The team consisted of 91 people and 87 servers.
It ran continuously for 72 hours without crashing.
Now, that database is working.
Lingyun recalled 1998, the 1998 of her previous life.
Back then, nobody knew how important databases were. Internet companies were using Oracle, IBM DB2, and Microsoft SQL Server. Open-source MySQL was just starting out, and PostgreSQL was still only circulating in academic circles.
Back then, no one would spend twelve million dollars and a year and a half to write a database from scratch.
But in this life, he did it.
It's not for now. It's for the future.
My phone vibrated again. It was a text message from Jiang Feng, containing only one line:
I will arrive in Jinan next Wednesday afternoon.
May 10, 1998, 9:00 AM.
Redmond, Microsoft Campus.
Ballmer stood at the door of the ICQ provisional project operations room, but didn't go in. The door was ajar, and a cacophony of sounds came from inside—the clatter of keyboards, the ringing of a telephone, and a rapid-fire argument.
Jim Olson walked down the hallway, carrying a folder. He stopped next to Ballmer and peeked through the crack in the door.
"What time will you be online?" Ballmer asked.
"Ten o'clock." Olsen glanced at his watch. "Fifty minutes left."
"What about the test?"
"We conducted four rounds of testing overnight. Of the thirty-seven bugs we found last night, thirty-one have been fixed, and there are still six known issues—two interface misalignments, three performance delays, and one occasional data synchronization loss."
Ballmer did not speak.
"The probability of it being lost is very low," Olson added. "It's about three in a thousand, and it only happens during peak hours. We'll monitor it once it goes live and fix it immediately."
Will users notice?
Olsen paused for two seconds. "It's possible to find it, but the chances are slim; the probability of this bug is extremely low."
Ballmer pushed open the door and walked into the office.
The operations room was nearly full, with all forty workstations occupied. Some were catching up on sleep at their desks, while others stood in front of the whiteboard, which was covered with last-minute to-do lists. The air was thick with the mingled scents of coffee grounds, printer toner, and human sweat.
Jonathan Meyer stood in front of the control panel, holding a stack of printing papers, confirming something with three engineers. Seeing Ballmer enter, he put down the papers and walked over.
"Mr. Ballmer."
"Ready?"
Meyer nodded. "We'll be online at 10:00 sharp. The servers have been expanded to three times their usual capacity, the entire technical support team is on standby, and the press release from the public relations department has been sent to the media."
Is it fully functional?
"Complete," Meyer paused, "but some parts still need optimization."
"Users don't know they need optimization," Ballmer said. "They just know it exists."
Meyer didn't respond.
Ballmer walked to the window. Outside was the central lawn of the Microsoft campus, the morning light turning the grass a vibrant green. Several employees in casual clothes were crossing the lawn toward the office building, coffee in hand.
He recalled the day Netscape Navigator released its new version in 1995. He was then in sales, watching his competitors' products dominate every headline. Later, they won, with bundling, with price, with every tactic.
Today, the same war is being fought on a different battlefield.
He turned to face Meyer.
"I'll be online at 10 a.m.," he said. "I'll report any problems immediately."