MICHAEL J. ABBOTT
Unix Server Software Architect & Performance Optimist
To design and develop large-scale high-performance OS X/BSD/Linux/Unix server
software from the Minneapolis/St. Paul area.
- Technical architect experienced with multi-petabyte
globally-distributed 24/7 production storage systems, high-performance
multi-threaded real-time client-server 64-bit object-oriented server
software development, and system-wide performance tuning.
- 29 years experience with C and Unix, especially networking and file
- 22 years experience with multi-threaded programming and
- 10 years experience with C++, 13 years with object-oriented
- 8 years experience with Perl, 3 years with internet mail protocols, Java,
- Open Source Projects: State Threads Library
and Accelerating Apache
- US Patents: 5,751,280,
regarding streaming media.
- Samples: Code and
documentation, available on-line.
California March 2008 - present
Senior Server Software Engineer
Contributed to the quality of two major releases of Mac OS X Server, Snow Leopard 10.6 and Lion 10.7.
- Enhanced and maintained open source components of the mail service
including Dovecot and Postfix. More than quintupled
Dovecot's scalability. Improved responsiveness for mobile users by
adding support for key Lemonade features including forward-without-download, compression, and
server-side indexing and searching of mail including non-text
attachments (e.g. PDFs, word processor documents). Contributed features
to the open source projects.
- Represented Apple with SPEC's
mail server subcommittee until the latter disbanded. Led development of
the SPECmail2009 benchmark.
Provided anonymized data
on real mail characteristics from Apple's corporate mail system.
Received a SPECtacular award for my contributions.
- Measured and tuned the performance of Apple mail servers. Published
a briefly-industry-leading benchmark
- Developed mechanisms for automatically reconfiguring services after
a server's network configuration changes.
- Assisted other teams by fixing bugs in Xsan, Address
Book, and several services.
California April 2003 - December 2007
Managed, enhanced, and streamlined the distributed storage system
supporting Yahoo! Photos, Briefcase, Groups, Shopping, Personals, and 40 other
properties as it grew from tens of terabytes in one farm to several
petabytes in dozens of farms around the world, serving a billion files a
day to hundreds of millions of users.
- Designed, implemented, and documented hundreds of improvements to
increase scalability, reliability, usefulness, and performance while
reducing expenditures on capital and operations. Charted a long-term
worldwide growth strategy. Identified and redesigned underperforming
system components. Evaluated new hardware and strategized with hardware
vendors. Refined the system into a mature managed storage service,
easy for properties to join.
- Led a team of up to six local and international engineers and
operators developing and maintaining the storage service. Introduced,
assigned, and prioritized tasks while balancing the complex, competing
business requirements of our client properties. Liaised with scores of
other teams. Trained new team members. Instituted policies of sharing
information and of documenting history, procedures, and best
- Held ultimate authority and responsibility for 24/7 production
operations of this storage service, safeguarding millions of dollars of
revenue. Served routinely as the primary pageable on-call contact.
Regularly audited configuration, security, backups, and performance.
Devised and refined dozens of continuous historical data plots and
extrapolations using open source software. Simplified, automated,
documented, and outsourced operations whenever possible. Investigated
complaints escalated from customer care. Mined content and access
patterns for business intelligence.
- Managed the relentless, exponential growth of the amount of content
stored. Collected growth data, projected trends, and predicted bursts.
Planned capacity expansion and ordered tens of millions of dollars of
hardware for just-in-time deployment. Presented and defended hardware
orders and system architecture to executive review boards. Shepherded
install tickets through the many stages of installation by disparate
- Significantly reduced the incremental cost of storage while
considering capacity, reliability, density, power consumption, speed,
and the ease and cost of operations. Reduced the number of servers per
unit of storage by over 90%. Created and implemented a framework for
transparently compressing stored content to reduce storage and save
millions of dollars more. Deployed multiple tiers of storage farms
based on different hardware platforms to meet the various business needs
of the properties and to minimize cost.
- Conceived, implemented, and deployed a complete system of automatic
fault detection, isolation, mitigation, and failover across
geographically disperse data centers. Met the highest corporate
standards for business continuity planning and high availability for
- Served as the primary security contact for nine properties.
Conducted biannual compliance reviews of those properties. Improved,
proselytized, and enforced security policies. Improved the prevention,
detection, and remediation of abuse.
See 21 more years of experience history on my full
Cornell University, College of Engineering, Ithaca, New York
Bachelor of Science Degree with Distinction, May 1989
Major: Computer Science; Concentration in Electrical Engineering
SKILLS AND INTERESTS
OS X, BSD, Linux, Irix, Unix
Objective-C and Cocoa, C++ and STL, C, Java, Perl,
shell/grep/awk/sed, PHP, assembly
Blocks and GCD
Servers and Services
Akamai and AkaDNS
Tracking and Ticketing
Storage and File Systems
Network Appliance filers,
pervasive social computing,
rev Jul 2011