Mike Heffner Email & Phone Number

San Francisco, California, Us

San Francisco, California, Us

San Francisco, California, Us

Austin, Texas, Us

Led data teams across the multiple Solarwinds acquired startups Librato, Tracelytics (now AppOptics) and Papertrail, building a unified cloud monitoring solution. Data teams were responsible for building distributed systems capable of consuming and processing high-volumes of metric and event data in real-time. Managed multiple direct reports and coordinated common technology adoption across multiple product platforms.Primarily responsible for AWS cost analysis and management, with over 10 years experience shipping on AWS. Built tools to track and optimize AWS spend that saved the company upwards of $100k/month in COGS.Led development for a custom in-memory TSDB built with C++ and RocksDB that implemented the Facebook Gorilla time series compression algorithms. Shipped an initial version to production that could process 100k metrics/second/core and could support single-digit millisecond latency read queries. Eventually this system replaced a large Apache Cassandra DB deployment, reducing infrastructure cost 15x.

Austin, Texas, Us

Librato acquired by SolarWinds, Jan 2015.

San Francisco, Ca, Us

Responsible for hiring and growing the data engineering team that supported the metrics platform used by thousands of customers. Delivered distributed systems built in Java and C++ that handled millions of metrics/second. Assisted team in delivering streaming, real-time solutions built around Kafka.Built the initial backend data platform for the Librato Metrics SaaS solution and took on many different roles/responsibilities as we grew the Librato Metrics product with a small team. Delivered the first version of our time series DB running on top of the Apache Cassandra DB using a custom schema built for multi-tenant time series data. Led the DevOps team and helped scale the platform running on AWS.

San Francisco, Ca, Us

Cofounded Librato, an observability platform delivered as a SaaS.Worked on user-level operating system virtualization and workload management solution, implemented with a pre-loaded system library transparent to applications. The solution pre-dated Linux cgroups/Docker containers and provided CPU, memory, net & file I/O resource-isolated containers for Linux and Windows. It allowed customers to run low-priority batch workloads alongside front-line web servers for increased density without impacting latency sensitive workloads. Primary responsibilities included implementing support for signal, file I/O, network I/O virtualization. Also developed CPU time-slicing algorithms to ensure correct CPU isolation for multi-threaded applications.Professional services work for a major financial institution implementing HFT messaging solutions. Primarily responsible for developing a BSD sockets-like API for Infiniband, supporting low-latency (~10 μs) and high-bandwidth (~10 Gbit/s) messaging. Developed a framework for building streaming HFT applications with hard real-time and deadline scheduled processing tiers.

Us

Senior member of team building a user-level checkpoint-restart library for large scale HPC applications. Allowed long-running distributed computations across many-node systems to be restored on single-node failures, thereby preventing loss of work completed.The CP/R library was responsible for checkpointing and restoring: processor context (x86 and x86-64), memory space (thread stacks, heap, mmap), mutex lock ownership, file modifications and replaying/restoring network communication state; all while transparent to the user application.Led efforts to quiesce MPI over Infiniband (MVAPICH) network state for CP/R.

Redmond, Washington, Us

Software Developer in Test, Internship

Ms, Computer Science

Bs, Computer Science