Distributed Scrum In Large Projects

[article]
Summary:
This article discusses the highlights of a distributed Scrum project run by SirsiDynix (Provo, UT) and StarSoft Development Labs (Cambridge, MA and St. Petersburg, Russia). The project focused on the new implementation of platform and system architecture for a complex Integrated Library System, which is best compared to a vertical market ERP system with a public portal interface used by more than 200 million people.

This article was written by Peter Vaihansky and Jeff Sutherland

One of the common objections to using agile approaches is, "What if our project is too big?" Indeed, agile almost invariably suggsts small teams. However, our experience of the past several years shows that agile methods such as XP and Scrum can be successfully applied to large-scale projects. In fact, Scrum is the first process with well-documented linear scalability. When you double team size in a well-implemented Scrum, you can double software output, even when the teams are distributed and outsourced.[1][2] In contrast with historical experience large projects can be just as productive as small projects for the first time.[3]

This article discusses some of the highlights of a distributed Scrum project by SirsiDynix (Provo, UT) and StarSoft Development Labs (Cambridge, MA and St. Petersburg, Russia). The project focused on the new implementation of platform and system architecture for a complex Integrated Library System (ILS) called Horizon 8.0, which is best compared to a vertical market ERP system with a public portal interface used by more than 200 million people.[1]

Best current Scrum practice is for local Scrum teams at all sites to synchronize once a day via a Scrum of Scrums meeting. In this case, {sidebar id=1} we describe something rarely seen on large, distributed teams. At SirsiDynix, all Scrum teams consist of developers distributed across different sites. Any team member from any site can work on any team task. While some Agile companies operate in this geographically transparent manner on a small scale, SirsiDynix and StarSoft have successfully implemented fully integrated Scrum teams with 56 developers in the U.S., Canada, and Russia.

New best practices for distributed Scrum seen on this project include:

    1. Daily Scrum meetings of all developers from multiple sites.
    2. Daily meetings of Product Owner team.
    3. Hourly automated builds from one central repository.
    4. No distinction between developers at different sites on the same team.
    5. Seamless integration of XP practices like pair programming and aggressive refactoring with Scrum.

While similar practices have been implemented on small distributed Scrum teams, this is the first documented project that demonstrates Scrum hyperproductivity for large distributed/outsourced teams building complex enterprise systems.

Let's look at the project in more detail and see how some of the obvious challenges were addressed.

Business Problem
SirsiDynix is a global leader in technology solutions for libraries. SirsiDynix has approximately 4,000 library and consortia clients, serving more than 200 million people through more than 20,000 library outlets around the world.

SirsiDynix was confronted with the requirement to completely re-implement its flagship product, a legacy library system with over 12,500 installed sites across the globe. The large number of developers required over many years in the midst of a changing business environment threatened to obsolete many feature requirements in the middle of the project. To complicate matters further, the library software industry was in a consolidating phase. Dynix started the project in 2002 and merged with Sirsi in 2005 to form SirsiDynix.

Fortunately, Dynix started the project with a scalable Agile process that could adapt to changing requirements throughout the project. Building a large product and facing time-to-market pressure, Dynix needed to quickly double or triple its productivity to exploit a business opportunity. The local talent pool was not sufficient to expand team size, and salary costs exceeded the constrained Ramp;D budget. The project could only succeed by augmenting Dynix' resources with outsourced teams.

On the other hand, outsourcing is only a solution if Agile practices are enhanced by capabilities of the outsourced teams. The primary driver was enhanced technical capability resulting in dramatically improved throughput of new

Pages

Tags: 

About the author

Peter Vaihansky's picture Peter Vaihansky

Peter Vaihansky is the General Manager for First Line’s North American operations. In this capacity, he oversees strategic client relationships, plays a key role in the sales and pre-sales process, and directs the company’s global marketing activities. Peter has been working with distributed Agile development teams since 2003. 

AgileConnection is one of the growing communities of the TechWell network.

Featuring fresh, insightful stories, TechWell.com is the place to go for what is happening in software development and delivery.  Join the conversation now!

Upcoming Events

Sep 22
Sep 24
Oct 12
Nov 09