Xiaxian Zhang

Production Engineer Lead responsibilities for DRUID (a real-time analytics database) & Targeting Insights platforms.

AWS migration for the Apache Druid cluster.

Tools used:

Technologies: Druid, Java, superset, EMR, RDS, EKS, presto, Lambda, athenz, CICD, S3, KMS, IAM, Route53, EBS, ALB, NLB. Secret manager, Cloudformation, Git, Screwdriver

AWS migration. Began working on AWS in 2014, until migration was complete in 2019. Be the first in the team to create VPC, route tables, subnets, internet gateway, NAT gateway, VPC endpoint, service endpoint, RDS, EBS etc. Migrated billing applications, include both online and batch applications. Online apps in ECS fargate, ECR, RDS, Batch billing jobs and Billing Revenue report jobs on EC2 with EBS as persistent volume .

BRR project leader: Migrated all BRR jobs from on-prem abinito to AWS EC2, EMR. Dry run for make sure there wasn’t discrepancy between on-prem jobs and AWS jobs.

As lead for components/projects, I always carefully review what developers delivered and provide feedback for design, security, maintainability, etc. Example, at the early stage of AWS migration, we were all new to AWS technology, developer send plain text from AOL data center to s3 through directly to www.amazon.com. When I noticed that it’s public endpoint which mean the sensitive data would leave AOL data center to public and enter AWS s3 through public endpoint amazon.com, I worked with developer to implemented proxy for data to go through VPN and VPC endpoint, so process won’t leave AOL data center or amazon network. Avoid possible data breach

For billing, any outage may lead to potential revenue loss or bad customer experience, so it’s critical to keep the system running all the time. Our billing database is the central component that many applications connect to it. Password change used to be error prone and troublesome process which leads to outages. Implemented blue/green databases password change process to achieve zero outage during password change

PCI compliance, BRR, SOX compliance

Technologies: Terraform, Cloudformation, Git, Screwdriver, Java, SQL, Python, PERL, Shell, TCL, Penatho, Ab Initio, Oracle, MySql, Memcache, Sybase, Apache, Tomcat, Hadoop, Druid, oozie, Jobscheduler, Autosys, Weblogic, New Relic, Jenkis, OpsGenie

Various commercial billing systems such as Infrante (C), Velista (J2EE weblogic), Informatics (weblogic) and in-house developed billing system OBI (Open Business Infrastructure)(java)

OBI/BRR (Billing Revenue Reporting)

Informatics billing system

Velista billing system: project lead

Infrante billing system

Database failover, setup automate failover process, from one and half hour error prone to 5 minutes error free

Improve intranet trouble shooting process by separate online and batch processes

Infrante (C), Velista (weblogic) team lead, design meeting, product meeting, provide feedback/requirement for operation persbect, Informatics (weblogic), OBI (Open Business Infrastructure)(java)

For billing, any outage may lead to potential revenue loss or bad customer experience, so it’s critical to keep the system running,

Billing Database is a central piece that many components connect to it. In the past, database failover, password change took long time and cause outage and long tine trouble shooting due to

Blue/green for database, setup blue/green databases password change process to achieve zero outage during password change

Database failover, setup automate failover process, from one and half hour error prone to 5 minutes error free

Improve intranet trouble shooting process by separate online and batch processes.

For billing, the data contain customer sensitive information, credit card information, so it’s very important to keep secure.

For billing, data integrate is very important. Migration focus on data integrity,

I joined the project in the very challenge moment. With my excellent knowledge of Portal/lnfranet and trouble shooting skills, plus my understanding of RMI.NET business requirements, I solved a lot problems occurred in customer registration, price plan, billing, reporting and integration with Siebel Systems.

• USA.NET is an internet messaging company and mainly provides email services to companies who have regional offices around world like airlines, hotels, and communications. The Portal/lnfranet system serves as both billing and provisioning systems. I completed data testing and customer conversion in the second month after I joined this project, which had not made any progress in the past six months. I worked with marketing and customer services to fix data problems in the legacy system. Since we loaded high quality data, the system runs very stable. I also implemented usage-loading module, which load customer usage automatically on daily bases. I performed data auditing between lnfranet and Siebel Systems and ensure the data integrity across multiple systems.

• Platform: Window NT, Sun Solaris. Language: Java, JSP, Java severlet, JDBC, DBI, CIC++, SQL, XML, and Perl. Database: Oracle. System: Portal/lnfranet.

One of the key developers in the HSA Portal/lnfranet development team.

I participated in designing and developing many different object-oriented GIS applications for telecommunication and utility companies such as Level 3 Communications and Duke Energy.