I recently wrote a detailed article about Large Agile Teams that was a detailed walkthrough of how to structure agile teams of various sizes. I suspect that this is the most comprehensive online discussion of this topic. The article addressed the following topics:
Organizing Agile Teams. The article starts with a summary of the results of some industry research that I've done regarding the size of agile teams, showing that agile techniques are in fact being successfully applied on a variety of team sizes. It then goes into detail describing the organization structure of agile teams at various sizes. The article starts with a discussion of small agile teams, covering the common rhetoric of how to organize such a team and then making observations about what actually happens in practice. It then walks through two approaches to organizing medium sized teams of 15 to 50 people - a structure for a single team and a structure for a team of teams. Finally, it walks through how to organize a large agile program of 50+ people, focusing a fair bit on the need for a leadership team to coordinate the overall activities within the program. This advice is similar to what is seen in the SAFe framework although proves to be a bit more flexible and pragmatic in practice.
Supporting Large Agile Teams. The leadership structure to support a large agile team is reasonably straightforward once you understand the issues that such a team faces. In this section the article overviews the need for three important sub teams within your overall leadership team: The Product Delivery Team, The Product Management Team, and The Architecture Team. It also describes the need for an optional Independent Testing/Integration Team, something misleadingly labeled an integration team in SAFe, that reflects some of the known agile testing and quality practices that I've been writing about for several years.
Organizing subteams. The article includes a detailed discussion for how to organize the work addressed by agile sub teams within a large agile program. These strategies include feature teams, component teams, and internal open source teams. As you would expect with the Disciplined Agile Delivery (DAD) framework, the article clearly summarizes the advantages and disadvantages of each approach on provides guidance for when (not) to apply each one. I suspect you'll find this portion of the article to be one of the most coherent discussions of the Feature vs. Component team debate.
Tailoring agile practices. The article provides a detailed overview of how the various DAD process goals are tailored to address the issues faced by large teams. This advice includes: Do a bit more up-front requirements exploration; Do a bit more up-front architectural modelling; Do a bit more initial planning; Adopt more sophisticated coordination activities; Adopt more sophisticated testing strategies; and Integrate regularly. My hope is that you find this part of the article very illuminating regarding how the DAD framework provides flexible and lightweight advice for tailoring your approach to address the context of the situation that you face.
Other Resources. The article ends with a collection of links to other resources on this topic.
I welcome any feedback that you may have about Large Agile Teams.
On Nov 16 2011, Kim Werner, Agile Coach from ATSC and Liz Parnell, Solution Design Manager from Blue Cross Blue Shield of North Carolina (BCBSNC), gave a webcast
sharing their experiences regarding how BCBSNC adopted a few Agile techniques, with the help of some good coaching, and adopted some IBM Rational Jazz tools
(Rational Team Concert and Rational Requirements Composer) to reduce time to market and lower development costs. BCBCNC works in the highly-regulated healthcare industry, so faced a few interesting constraints when adopting agile.
activities are evolutionary (iterative and incremental) and highly collaborative in nature. Initially requirements are explored at a high level via requirements envisioning
at the beginning of the project and the details are explored on a just-in-time (JIT) basis via iteration modeling
and model storming
activities. The way that you perform these agile practices, and the extent to which you do so, depends on the situation in which a project team finds itself. The Agile Scaling Model (ASM)
is a contextual framework for effective adoption and tailoring of agile practices to meet the unique challenges faced by a system delivery team of any size. To see how this works, let's apply the concepts of the ASM to see how we would scale our agile approach to requirements.
First, let's consider how a small, co-located team would work. The first two categories of the ASM are core agile development and disciplined agile delivery
, the focus of both are small co-located teams in a fairly straightforward situation. In these situations simple techniques such as user stories
written on index cards and sketches on whiteboards
work very well, so the best advice that I can give is to stick with them. Some teams will take a test-driven development
(TDD) approach where they capture their requirements and design in the form of executable specifications
, although this sort of strategy isn't as common as it should be (yet!), likely because of the greater skill and discipline that it requires. Traditionalists often balk at this approach, believing that they need to document the requirements in some manner. But, for a small co-located team working in a collaborative manner, requirements documentation proves to be little more than busy work, often doing nothing more than justifying the existence of a business analyst who hasn't made the jump to agile yet. Don't get me wrong, there are good reasons to write some requirements documentation, and we'll see this in a minute, but you should always question any request for written specifications and try to find more effective ways to address the actual goal(s) motivating the request. Never forget that written documentation
is the least effective communication
option available to you.
Although inclusive tools
such as whiteboards and paper work well for requirements, for development activities you will need electronic tools. You will either put together an environment from point-specific tools or adopt something more sophisticated such as IBM Rational Team Concert (RTC)
which is already fully integrated and instrumented. RTC is a commercial tool, but luckily you can download a 10-license environment free of charge, which is just perfect for a small team. Larger teams, of course, will need to purchase licenses. One of the things that a disciplined agile delivery approach adds to core agile development is it addresses the full delivery life cycle, which is important because it explicitly includes pre-construction activities such as requirements envisioning. The first step in scaling agile techniques is to adopt a full delivery life cycle which covers the full range of activities required to initiate a project, produce the solution, and then release to solution to your end users.
More interesting is the third category of the ASM, Agility@Scale, and how its eight agile scaling factors
affect the way that you tailor your process and tooling strategy. Let's explore how each one could potentially affect your agile requirements strategy:
- Geographical distribution. The majority of agile teams are distributed in some manner -- some people are working in cubicles or private offices, on different floors, in different buildings, or even in different countries -- and when this happens your communication and coordination risks goes up. To counter this risk you will need to perform a bit more requirements envisioning up front to help ensure that everyone is working to the same vision, although this doesn't imply that you need to write detailed requirements speculations which would dramatically increase the risk to your project. Remember, agilists do just barely enough modeling and are prepared to iteratively elicit the details when they need to do so. The more distributed the team is the more likely they will need to adopt software-based requirements modeling tools such as IBM Rational Requirements Composer (RRC) which supports streamlined, agile requirements elicitation throughout the delivery life cycle. Index cards and whiteboards are great, but they're difficult to see if you're outside the room where they're posted. I've written a fair bit about distributed agile development in this blog.
- Team size. Some organizations, including IBM, are successfully applying agile techniques with teams of hundreds of people. A team of one hundred people will naturally work much differently than a team of ten people, or of one thousand people. Large teams are organized into collections of smaller teams, and the requirements for the overall project must be divvied up somehow between those teams. The implications are that as the team size grows you will need to invest a bit more time in initial requirements envisioning, and in initial architecture envisioning for that matter; you will need to use more sophisticated tools; and may need to use more sophisticated modeling techniques such as use cases and functional user interface prototypes. See large agile teams for more advice.
- Compliance requirement. When regulatory issues – such as Sarbanes Oxley, ISO 9000, or FDA CFR 21 – are applicable you are likely going to be required to capture requirements specifications in some manner and to enact traceability between those requirements. However, I highly recommend that you read the actual regulations yourself and don't let bureaucrats interpret them for you (doesn't it always seem that their interpretation always results in an onerous, documentation heavy solution?) because I have yet to run into a regulation which required you to work in an ineffective manner. Managing your requirements as work items in RTC can often more than meet your regulatory requirements for documentation and traceability, although you may want to consider a tool such as IBM Rational RequisitePro for complex regulatory situations.
- Domain complexity. The manner in which you elicit requirements for a data entry application or an informational web site will likely be much simpler than for a bio-chemical process monitoring or air traffic control system. More complex domains will require greater emphasis on exploration and experimentation, including but not limited to prototyping, modeling, and simulation. Although user stories may be effective as a primary requirements artifact in simple domains, in more complex domains you are likely to find that you need to drive your requirements effort with more sophisticated modeling techniques.
- Organization distribution. Sometimes a project team includes members from different divisions, different partner companies, or from external services firms. In these cases, particularly where the work is strictly organized between the various organizations (perhaps for security concerns), you may need a more sophisticated approach to managing the requirements. RTC enables you to organize the requirements between teams, and then to automatically track progress in real time via the RTC project dashboard.
- Technical complexity. The technical complexity of a solution can vary widely, from a single platform silo application to a multi-platform application working with legacy systems and data to a full-blown systems engineering effort. Complex technical domains, just like complex business domains, require more complex strategies for requirements elicitation and management. The requirements for your legacy systems are likely to have been captured using tools and techniques appropriate for that platform, for example the requirements for your COBOL application may have been captured using data flow diagrams and data models, whereas the requirements for your Java legacy application where captured using UML diagrams. The subteam working on the COBOL system might be using IBM Rational Application Developer (RAD) and RTC for Z whereas the Java subteam may use Eclipse with RTC. Because systems engineering projects can stretch on for years, particularly when the hardware is being developed in parallel to the software, sophisticated tooling such as IBM Rational DOORS is often used in these situations. For more information about systems engineering, see the IBM Rational Harmony process.
- Organizational complexity. Your approach to requirements elicitation and management will be affected by a host of organizational complexities, including your corporate culture. When the culture is flexible and collaborative you can be very agile in your approach to requirements, but as it becomes more rigid you become more constrained in what is considered acceptable and thus take on greater project risk. For example, many organizations still struggle with their approach to funding projects, often demanding that the project team provides an "accurate" estimate up front to which they will be held to. This in turn motivates risky behavior on the part of the development, including a "big requirements up front (BRUF)" approach where a detailed requirements speculation is developed early in the project. This is just one example of how questionable corporate culture can impact the way in which an agile team works.
- Enterprise discipline. Some organizations have enterprise-level disciplines, such as enterprise architecture, enterprise business modeling, strategic reuse, and portfolio management in place. These disciplines can easily be agile and from what I can tell the more successful efforts appear to lean more towards the agile end of the spectrum rather than the traditional end. Having an enterprise business modeling effort underway will affect your project-level requirements strategy -- you'll be able to leverage existing models, have access to people who understand the domain at an enterprise level, and will likely need to map your project efforts back to your enterprise models. The enterprise modelers will likely be using tools such as IBM Rational System Architect or IBM Websphere Business Modeler.
It is important to note that the way that you tailor the agile practices that you follow, and the tools that you use, will reflect the situation that you find yourself in. In other words, you need to right size your process and the Agile Scaling Model (ASM) provides the context to help you do so. As you saw above, in simpler situations you will use the simpler tools and techniques which are commonly promoted within the core agile development community. But, when things become a bit more complex and one or more of the scaling factors applies you need to modify your approach -- just don't forget that you should strive to be as agile as you can be given the situation that you find yourself in.
The explicit phases of the Unified Process -- Inception, Elaboration, Construction, and Transition -- and their milestones are important strategies for scaling agile software development to meet the real-world needs of modern organizations. Yes, I realize that this is heresy for hard-core agilists who can expound upon the evils of serial development, yet these very same people also take a phased approach to development although are loathe to admit it. The issue is that the UP phases are like seasons of a project: although you'll do the same types of activities all throughout a project, the extent to which you do them and the way in which you do them change depending on your goals. For example, at the beginning of a development project if you want to be effective you need to do basic things like identify the scope of the project, identify a viable architecture strategy, start putting together your team, and obtain support for the project. Towards the end of a project your focus is on the activities surrounding the deployment of your system into production, including end-of-lifecycle testing efforts, training, cleaning up of documentation, piloting the system with a subset of users, and so on. In between you focus on building the system, including analysis, design, testing, and coding of it. Your project clearly progresses through different phases, or call them seasons if the term phase doesn't suit you, whether your team is agile or not.
The UP defines four phases, each of which address a different kind of risk:1. Inception. This phase focuses on addressing business risk by having you drive to scope concurrence amongst your stakeholders. Most projects have a wide range of stakeholdres, and if they don't agree to the scope of the project and recognize that others have conflicting or higher priority needs you project risks getting mired in political infighting. In the Eclipse Way this is called the "Warm Up" iteration and in other agile processes "Iteration 0".2. Elaboration. The goal of this phase is to address technical risk by proving the architecture through code. You do this by building and end-to-end skeleton of your system which implements the highest-risk requirements. Some people will say that this approach isn't agile, that your stakeholders should by the only ones to prioritize requirements. Yes, I agree with that, but I also recognize that there are a wide range of stakeholders, including operations people and enterprise architects who are interested in the technical viability of your approach. I've also noticed that the high-risk requirements are often the high-business-value ones anyway, so you usually need to do very little reorganization of your requirements stack.3. Construction. This phase focuses on implementation risk, addressing it through the creation of working software each iteration. This phase is where you put the flesh onto the skeleton.4. Transition. The goal of this phase is to address deployment risk. There is usually a lot more to deploying software than simply copying a few files onto a server, as I indicated above. Deployment is often a complex and difficult task, one which you often need good guidance to succeed at.
Each phase ends with a milestone review, which could be as simple as a short meeting, where you meet with prime stakeholders who will make a "go/no-go" decision regarding your project. They should consider whether the project still makes sense, perhaps the situation has changed, and that you're addressing the project risks appropriately. This is important for "agile in the small" but also for "agile in the large" because at scale your risks are often much greater. Your prime stakeholders should also verify that you have in fact met the criteria for exiting the phase. For example, if you don't have an end-to-end working skeleton of your system then you're not ready to enter the Construction phase. Holding these sorts of milestone reviews improves your IT governance efforts by giving senior management valuable visibility at the level that they actually need: when you have dozens or hundreds of projects underway, you can't attend all of the daily stand up meetings of each team, nor do you even want to read summary status reports.
These milestone reviews enable you to lower project risk. Last Autumn I ran a survey via Dr. Dobb's Journal (www.ddj.com) which explore how people actually define success for IT projects and how successful we really were. We found that when people define success in their own terms that Agile has a 71% success rate compared with 63% for traditional approaches. Although it's nice to that Agile appears to be lower risk than traditional approaches, a 71% success rate still implies a 29% failure rate. The point is that it behooves us to actively monitor development projects to determine if they're on track, and if not either help them to get back on track or cancel them as soon as we possibly can. Hence the importance of occasional milestone reviews where you make go/no-go decisions. If you're interested in the details behind the project, they can be found at http://www.ambysoft.com/surveys/success2007.html .
Done right, phases are critical to your project success, particularly at scale. Yes, the traditional community seems to have gone overboard with phase-based approaches, but that doesn't mean that we need to make the same mistakes. Let's keep the benefit without the cost of needless bureaucracy.[Read More
Modified by ScottAmbler
In my previous blog posting, http://www.ibm.com/developerworks/blogs/page/ambler?entry=strategies_for_distributed_agile_teams , I overviewed several strategies for improving your effectiveness at geographically distributed development (GDD). Those strategies were fairly generic and directly applicable to both traditional and agile development teams. In this posting I focus on strategies which are more agile in nature, although they could also be applied to more traditional approaches as well. These strategies are:
1. Get the whole team together at the beginning of the project. Your goals are to build rapport amongst the team, to get to know the people that you’re working with to facilitate communication later on, and to better understand the situation on the ground. The implication is that you will need to fly some people around, increasing your initial expenses, an investment that many organizations balk at. The reality is that you will eventually end up paying for travel anyway, either because you actually flew people around or because your communication costs are higher throughout the project. In short, don’t be penny wise and pound foolish.
2. Organize your team around the system architecture. The most effective way to organize a distributed team is around the architecture of the system that you are building, not around the job functions of the people involved. In other words, if your team is in Toronto, Rome, and Bangalore then each subteam should be responsible for one or more subsystems. It would be a mistake to organize the teams around job function, for example to have the architects and analysts in Toronto, the developers in Rome, and the testers in Bangalore because this structure would require significantly more documentation and other forms of communication to coordinate the teams, increasing both cost and risk. As I mentioned in my previous blog posting you will need to invest in some initial architecture envisioning at the beginning of a project to identify the subsystems and their public interfaces, and that to do that you’ll also need to do some initial requirements envisioning to drive this architecture effort. I suggest that you take an Agile Model Driven Development (AMDD) approach to this to enable you to gain the value from modeling without the costs and risks associated with up-front comprehensive modeling and documentation that get many traditional project teams in trouble.
3. Have “daily stand-up meetings”. A common practice on co-located agile teams is to have daily stand-up meetings where people share the status of what they did yesterday, what they intend to do today, and whether they’re running into any problems. These short meetings enable team coordination. Distributed teams can do this as well, the people in a given geographical location can hold local stand-up meetings and then representatives from each location can hold a shared meeting to coordinate the subteams. Whereas local stand-up meetings are held first thing in the morning, distributed daily stand-up meetings may need to be held at unusual times so as to include people at distant locations.
4. Have Ambassadors. Ambassadors are people who travel between sites, often technically senior people or senior business experts, to share information between the subteams. Getting the team together at the beginning of the project sets the foundation for communication, but without continual investment in maintaining effective collaboration between teams you run the risk of your subteams deviating from the overall strategy. These are typically short engagements, a week or two in length, because of the pressures it puts on the people doing the actual traveling. The implication is that you’ll have several people flying between sites at any given time on a reasonable rotation schedule. Because you’ll have some people flying around, your local team rooms should accommodate visitors by having one or more desks available for them to use when they’re visiting.
5. Have Boundary Spanners. A boundary spanner is someone who is located on site who focuses on enabling communication between subteams as well as within their subteam. On large distributed teams you’ll find that you have three flavors of boundary spanners – team leaders who take on project management responsibilities on the subteam, product owners who are responsible for representing the business within the subteam, and architecture owners responsible for technical direction on the team. These boundary spanners will work closely with their peers, having regular coordination meetings across all subteams as well as impromptu one-on-one meetings to deal with specific issues between individual subteams.
6. Ensure that the global team gets the credit it deserves. In both offshoring and nearshoring environments it’s common to see small teams in North America or Europe driving the efforts of significantly larger teams in another country. Yet, at the end of the project it always seems as if the smaller team, often because they work for the direct customer, gets the lion’s share of the credit – unless of course the project failed, then the subcontracting team often seems to get virtually all of the “credit”. This clearly isn’t fair, and it clearly doesn’t promote effective teamwork between the subteams in the future.
7. Take a lean approach to development governance. As I’ve written in the past, effective governance is based on enablement and collaboration instead of the traditional approaches of management and control. Good governance measures progress through regular delivery of working software, not through status reports or delivery of detailed specifications. Good governance is based on the idea of having a living process which changes to reflect lessons learned as your project progresses. In a previous blog posting at http://www.ibm.com/developerworks/blogs/page/ambler?entry=lean_development_governance I’ve discussed lean development governance in greater detail.
The strategies that I’ve described are clearly nothing more than common sense, something that can be said of all agile strategies. Sadly, as Mark Twain lamented, common sense isn’t very common in practice.[Read More
A common question that I keep running into with customers is whether you can take an agile approach to service oriented architecture (SOA). The quick answer is yes, because Agile is orthogonal to the implementation technologies used. You can take an agile approach developing COBOL applications running on mainframes, fat-client Java applications, multi-tier J2EE applications, and yes, even services. Granted, it's easier to do with some technologies than others, either because of the nature of the technology or because of the supporting tools.
The long answer is "yes, but". You don't adopt an SOA approach for the sheer joy of doing so, instead you very likely want to improve the level of reuse within your organization. To succeed at SOA-driven reuse you need an enterprise focus, something that doesn't appear to be very common on many agile teams. Therein lies the challenge. Several strategies for improving your chances with Agile SOA, and SOA in general, follows:1. Invest in some initial enterprise architecture modeling. You don't need to identify all of the details up front, that would take too long and actually put the effort at risk, but you do need to set a starting point to guide development teams. Identifying the technical architecture is critical, and identifying a few basic services which would provide immediate business value to one or more teams is critical. Involve people from several application project teams to ensure that you get a wide range of input. See http://www.agiledata.org/essays/enterpriseArchitecture.html for a streamlined approach to enterprise architecture modeling. Creating big, detailed models often proves to be a waste of time because development teams are rarely motivated to read mounds of documentation.2. Build out the initial infrastructure on a real application development project. This proves that your SOA strategy actually works and puts the technical foundation in place for future teams. During this period you'll be tempted to try to support several development teams, which is feasible but dramatically increases your risk. It's also tempting to focus simply on getting the infrastructure in place without delivering any business functionality, but this risks producing an ivory-tower architecture that nobody is interested in.3. Spread the service architects out onto application development teams. The people that formulated and then proved your SOA should be actively involved on the development teams that are working with it to ensure that the teams use it appropriately and to ensure that the architects get concrete feedback which they can use to evolve the architecture. When working on agile teams, these people will need to work in a collaborative and evolutionary approach just like other team members.4. Fund reuse separately. I've lost track of the number of organizations that I've run into that fail at reuse because their development teams never have the resources to develop reusable assets. That's simply the nature of the beast -- project teams will always be more interested in addressing their own specific requirements than they are in investing the time and effort to make something reusable. The real problem here is that you expect them to act differently. A better strategy is to have a separate reuse engineering team that has the resources to monitor existing projects to look for potentially reusable assets. When they find said assets this team does the work to harvest the asset, to reengineer it to make it reusable, and then to integrate back into the original source project. The goal is to make it as painless as possible to produce reusable assets such as services. If you expect project teams to do this work out of the goodness of their hearts then you're effectively punishing them when they do the right thing. That's not a very good governance strategy, IMHO.5. The reuse team now owns the asset. Any reusable asset, including services, will need to be maintained, evolved over time, and supported. This isn't free nor is it viable for project teams to do so.
If you're interested, I provide agile strategies for both enterprise architecture and strategic reuse in the book "Enterprise Unified Process". Although written under the assumption that you're taking a RUP-based approach to development, the reality is that the EUP can extend any evolutionary/agile software development process so that it addresses the larger-scale needs of modern IT organizations.
- Scott[Read More
When you are first adopting agile techniques in your organization a common strategy is to run one or more pilot projects. When organizing these projects you typically do as much as you can to make them successful, such as finding:
- Projects where the stakeholders are willing to actively work with you.
- IT people who are flexible, willing to try new things, and willing to collaborate with one another.
- IT people who are generalizing specialists, or at least willing to become so.
- Finding a project which is of medium complexity (therefore it's "real" in the sense that it's significant to your organization) but not one where it can make or break your organization (therefore it's safe to experiment with).
In North America we refer to this as "cherry picking" because you're picking the cherry/best situation that you can find.
- Being agile may not have been the primary determinant of success. You set up an environment where you have a good relationship with your stakeholders, where you have good people who want to work together, and the project is challenging but not impossible. Oh, and by the way you adopted a few agile techniques as well. Sounds to me that situation you could have adopted a few not-so-agile techniques instead and still succeed. Although my various project success surveys, see my IT surveys page for details, have shown time and again that agile project teams are more successful than traditional project teams I haven't been able to tease out (yet) whether this success is attributable to agile or just attributable to improved project initiation efforts.
- When adopting agile/lean widely across your organization, you can't cherry pick any more. For the past few years I've been working with IT organizations that are in the process of adopting agile/lean strategies across their entire organization, not just across a few pilot projects. What these organizations are finding is that they need to find ways to adopt agile where the business isn't as willing to work with IT, where some of the people aren't so flexible or collaborative, where some of the people are narrowly specialized and not as willing to expand their skills, or where the project exhibits scaling factors which motivates you to tailor your agile approach. It's harder to succeed with agile in these situations because they're not as "cherry" as what you've experienced previously. Luckily, if you've been successful previously then you now have some agile experienced people, you have successes to reference, and you've likely overcome some problems even in the cherry situations which you have learned from. So, your cherry successes will hopefully improve your ability to succeed even in "non cherry" situations.
- You need to work smarter, not harder. If the source of your success was actually from improved project initiation practices and not from agile, then recognize that and act accordingly. Realistically part of your success was from that and part was from agile, and the organizations that adopt a measured improvement approach potentially have the data to determine which practices lead to success and which didn't. Without the metrics you're effectively flying blind when it comes to deciding how to improve. There is clearly a mandate for smarter work practices within IT, within your organization as a whole for that matter.
If you want to gain more insight into some of the issues that you'll face when adopting agile across your organization, I suspect that you'll find my recent paper Scaling Agile: An Executive Guide
to be interesting. I've got a more detailed paper in the works, so stay tuned to this blog.
One of the scaling factors
of the Agile Scaling Model (ASM)
is technical complexity.
The fundamental observation is that the underlying technology of solutions varies and as a result your approach to developing a solution will also need to vary.
It’s fairly straightforward to achieve high-levels of quality if you’re building a new system from scratch on a known technology platform, but not so easy when there are several technologies, the technologies are not well known, or legacy assets are involved.
There are several potential technical complexities which a Disciplined Agile Delivery (DAD) team may face:
- New technology platforms. Your team may choose to work with a technology platform which is either new to the team or sometimes even new to the industry. In the past few years new technology platforms include the Android operating system, Apple’s iPad platform, and various cloud computing (http://www.ibm.com/ibm/cloud/) platforms. Working with these platforms may require you to adopt new development tools and techniques, not to mention the need to train and mentor your staff in their usage. Furthermore, your team may need to allocate time for architectural spikes to explore how to use the new technology and to prove the overall architecture with working code early in the project lifecycle (this is a DAD milestone).
- Multiple technology platforms. IT solutions often run on multiple platforms. For example, a system’s user interface (UI) could run in a browser, access business logic implemented using J2EE on Websphere which in turn invokes web services implemented in COBOL running on a Z-series mainframe, and stores data in an Oracle database, a DB2 database, and in several XML files. Implementing new business functionality, or updating existing functionality, could require changes made on several of these platforms in parallel. The implication is that you’ll need to adopt tools and strategies which enable your team to develop, test, and deploy functionality on all of these platforms. Testing and debugging in particular will become more difficult as the number of technology platforms increases, potentially requiring you to adopt the practice of parallel independent testing. The Agility at Scale survey found that 34% of respondents indicated that their agile teams were working with multiple technology platforms.
- Legacy data. IT solutions should leverage existing, legacy data wherever possible to reduce the number of data sources and thereby increase data quality within your organization. Also, using existing data sources can potentially speed up development, assuming your team has a good relationship with the owners of the legacy data sources (sadly, this often isn’t the case as the Data Management Survey found). Working with legacy data sources may require improved database regression testing, practices, database refactoring practices, and agile approaches to data administration. The Agility at Scale survey found that 42% of respondents indicated that their agile teams were working with legacy data sources (personally, I’m shocked that this figure is so low, and fear that many agile teams are contributing to data quality problems within their organization as a result).
- Legacy systems. There are several potential challenges with legacy systems. First, the code quality may not be the best either because it was never really that good to begin with or because it’s degraded over the years as multiple people worked with it. You know you’ve got a quality problem if you’re either afraid to update the code or if when you do so you have to spend a lot of time debugging and then fixing problems revealed when doing the update. If the legacy system is a true asset for your organization you will want to pay off some of this technical debt by refactoring the code to make it of higher quality. Second, you may not have a full regression test suite in place, making it difficult to find problems when you do update the code let alone when you refactor it. Third, your development tools for your legacy code may be a bit behind the times. For example, I often run across mainframe COBOL developers still working with basic code editors instead of modern IDEs such as Rational Developer for System Z. Some of the strategies to deal effectively with legacy systems are to adopt a modern development toolset if you haven’t already done so (better yet, if possible adopt a common IDE across platforms and thereby reduce overall licensing and support costs) and to adopt agile practices such as static code analysis, dynamic software analysis, and continuous integration (CI). The Agile Project Initiation Survey found that 57% of respondents were integrating their new code with legacy systems and 51% were evolving legacy systems.
- Commercial off-the-shelf (COTS) solutions. COTS solutions, also called package applications, can add in a few complexities for agile teams. The packages rarely come with regression test suites, they often have rules about what you can modify and what you shouldn’t (rules that are ignored at your peril), and they’re often architected with the assumption that they’re the center of the architectural universe (which is a valid assumption if they’re the only major system within your organization). As I describe in my article Agile Package Implementations it is possible to take an agile approach to COTS implementations, although it may require a significant paradigm shift for the people involved. The Agility at Scale survey found that 15% of respondents indicated that their agile teams were working with COTS solutions.
- System/embedded solutions. For the sake of simplicity, if your team is developing a solution with both hardware and software aspects to it then you’re a systems project. Embedded systems are a specialization where the system has a few dedicated functions often with real-time constraints. Bottom line is that systems/embedded projects are typically more challenging than software-only projects – it gets really interesting when laws of physics starts to kick in, such as when you’re building satellites or space probes. I highly suggest Bruce Douglass’s book Real-Time Agility if you are interested in taking an agile approach to systems/embedded solution delivery.
The technical complexity faced by a project team is contextual – Working with four technology platforms is straightforward for someone used to dealing with seven, but difficult for someone used to dealing with just one. Recommended Reading:
Modified by ScottAmbler
I just wanted to round out my discussion about agile approaches to geographically distributed development (GDD) with a few important words of advice:1. Get some experience. Worry less about enterprise adoption and instead get started with a small project, or better yet a series of increasingly more complex projects. There will be learning experiences as you build a relationship with the offshore service provider. This advice is applicable whether you’re working with your own offshore division or with an independent service provider.2. Have a long-term staffing strategy. It may be great in the short term to have work done in a lower cost country, but how are you going to transfer the necessary skills to the maintenance and support team. Outsourcing that work is also an option, but it can be a risky one as you would need to build up expertise in “your” systems if you ever decide to insource that work again.3. Be concerned about intellectual property (IP). The rules are different around the world, and you may inadvertently be financing the creation of a new international competitor if you don’t have a clear division of ownership. And yes, this may mean that some components of your systems are still built internally by your own organization.4. Show off locally before you go global. GDD makes things harder to manage, so if you’re struggling to manage local teams you’re really going to struggle managing teams at a distance. Make sure you have local success first and are good at agile development in general. Furthermore, if your agile GDD projects run into trouble, don’t end your local agile adoption just because of difficulties with distributed projects.5. Let your offshore partners lead. The offshore partner likely has more experience than you at successful distributed development, and this is particularly true when you’re dealing with an established service provider.6. Do some reading. There’s a great IBM Redbook entitled “Global Development and Delivery in Practice: Experiences of the IBM Rational India Lab” which can be downloaded free of charge from http://www.redbooks.ibm.com/abstracts/sg247424.html7. Do some viewing. We recorded a Rational Chat a few months ago entitled "Being Agile in a Global Development Environment" which is posted at https://www14.software.ibm.com/webapp/iwm/web/reg/acceptSignup.do?lang=en_US&source=dw-c-wcsdpr&S_PKG=120607&S_TACT=105AGX23&S_CMP=TALKS&cp=UTF-8 . I also gave a keynote on Agile approaches to GDD at Software Development Practices 2007 held in Boston in the Autumn of 2007. The video can be downloaded free of charge from http://www.life20.net/video/scottambler.mov .[Read More
The popular Agile literature can often seam naive when it comes to how Agilists work with project stakeholders:- Extreme Programming (XP) has a practice called On-Site Customer where one or more people work closely with your team to provide information and to make decisions in a timely manner.- Scrum has the role of Product Owner who is the one single person that the development team goes to for decisions about requirements. - Agile Modeling (AM) has the practice of Active Stakeholder Participation which extends On-Site Customer to get the stakeholder(s) actively involved with the modeling effort through the use of inclusive tools and techniques.
These are great strategies for small, co-located teams doing straightforward development, but they quickly fall apart at scale. This occurs for several reasons:1. Stakeholders are a diverse group. Your stakeholders include end users, business management, project funders, enterprise architects, operations staff, support staff, other system development teams, and many others. Different people have different, and often contradictory, requirements and they certainly have different priorities. It's questionable whether a single person, or a handful of persons, can adequately represent this diverse group.2. One person becomes a bottleneck. Even with a small co-located team this is a problem, let alone one that is geographically distributed or one that is very large. There's no way that a single person can be available 24/7 in a responsive manner to support distributed teams.3. It's a difficult role. The Product Owner/Customer (POC) is responsible for representing the business to the development team. They're making important decisions on a regular basis, decisions which they'll be held accountable for.4. One person becomes a serious project risk. Not only is it questionable whether a single person can fairly represent all stakeholders, even if they could what happens if you lose that person? They effectively become a single point of failure for your team.
To scale this role, consider the following strategies:1. Recognize the true scope of the POC role. Not only are they stakeholder proxies they also are a development team representative to the stakeholder community as a whole. As stakeholder proxies they'll make decisions and prioritize the work, they'll run requirements elicitation sessions, they'll negotiate priorities, and they'll put the development team in contact with stakeholders who have expertise in specific aspects of the domain. As team representatives they'll often demo the current version of the system to other stakeholders, communicate the status of the project to people, and respond to various requests for information from the stakeholders.2. Have multiple people in it. A single POC works well for small, co-located teams developing simple software. At scale you'll soon discover that you need multiple people in this role so that they don't become a bottleneck. For distributed teams it's common to see each subteam have one or more POCs who are managed by a primary/chief POC. The primary POC typically works on the coordinating team with the chief architect (I'll talk about this role in a future blog posting) and the program manager (also a topic for a future blog posting).3. Train them in business analysis skills. The person(s) in the POC role need good business analysis skills. If fact, it's common for people who were formerly BAs for traditional teams to step into the POC role, particularly with BAs who originally come from the business side of your organization. This strategy has its advantages and disadvantages. As a BA they've likely got solid business knowledge but their instincts may motivate them to take a documentation-driven approach to providing information to the development team instead of a collaboration-based approach. Be careful.4. Consider the full system development lifecycle. There's far more to the POC role than supporting the development team during Construction iterations. During "Iteration 0", the Inception phase for an Agile RUP project or the warm-up phase for an Eclipse Way project, the POC(s) will often lead the initial requirements envisioning efforts. The product backlog, or better yet your work item list, needs to come from somewhere after all. During the release iteration(s), the Transition phase for RUP or the End-Game phase for Eclipse Way, the POC(s) will focus on communicating the upcoming release to the stakeholder community, will be actively involved with any final user acceptance testing (UAT), and may even be involved with training end users.
In my January 2008 column in Dr Dobb's Journal, posted at http://www.ddj.com/architect/204801134 , I provide detailed advice about how to scale the way that you work with stakeholders on Agile projects by applying the practices of Agile Model Driven Development (AMDD). There's no magic solution, you just need to choose to organize yourself effectively. The good news is that you can easily work with stakeholders at scale.[Read More
I recently wrote an "e-book" for Internet Evolution overviewing agile software development at scale. The goal of the Agility at Scale: Become as Agile as You Can Be
ebook is to get people thinking outside of the box a bit when it comes to agile development strategies and see that they really are ready for primetime.
Test-driven development (TDD) is a common agile programming technique which has both specification and validation aspects. With TDD, you specify your software in detail on a just-in-time (JIT) basis via executable tests that are run in a regression manner to confirm that the system works to your current understanding of what your stakeholders require.
TDD is the combination of test-first development (TFD) and refactoring. With TFD, you write a single test (at either the requirements level with customer/acceptance tests or the design level with developer tests) and then you write just enough software to fulfill that test. Refactoring is a technique where you make a small change to your existing code to improve its design without changing its semantics.
TDD offers several benefits:1. It enables you to take small, safe steps during development, increasing programmer productivity.2. It increases quality. Agile developers are doing more testing, and doing it more often, than ever before. We're also fixing the problems that we find right on the spot.3. It helps to push validation activities early in the lifecycle, decreasing the average cost to fix defects (which rises exponentially the longer it takes you to detect them).4. Through single sourcing information, by treating tests as both specifications and as tests, we reduce the work required, increasing productivity.5. We leave behind valuable, up-to-date, detailed specifications for the people who come after us. Have you ever met a maintenance programmer who wouldn't want a full regression test suite for the code that they're working with?
But TDD isn't perfect. Although TDD is great at specifying code at a fine-grain level, tests simply don't scale to address higher level business process and architectural issues. Agile Model Driven Development (AMDD) enables you to scale TDD through initial envisioning of the requirements and architecture as well as just-in-time (JIT) modeling at the beginning and during construction iterations. To scale requirements-level TDD, you must recognize that customer tests are very good at specifying the details, but not so good at providing overall context. High-level business process models, conceptual domain models, and use cases are good at doing so, and these work products are often created as part of your initial requirements envisioning and iteration modeling activities. Similarly, to scale design-level TDD you must recognize that developer tests are very finely grained but once again do not provide overall context. High-level architecture sketches created during envisioning activities help set your initial technical direction. During each construction iteration, you'll do more detailed design modeling to think through critical issues before you implement them via TDD.
You also need to scale the validation aspects of TDD. TDD is in effect an approach to confirmatory testing where you validate the system to the level of your understanding of the requirements. The fundamental challenge with confirmatory testing, and hence TDD, is that it assumes that stakeholders actually know and can describe their requirements. Therefore you need to add investigative testing practices which explore issues that your stakeholders may not have thought of, such as usability issues, system integration issues, production performance issues, security issues, and a multitude of others.
For further reading, I suggest:1. My article "Introduction to TFD/TDD" at http://www.agiledata.org/essays/tdd.html which overviews TDD.2. My February 2008 column in Dr. Dobb's Journal entitled "Scaling TDD" at http://www.ddj.com/architect/205207998 which explores this issue in detail. 3. Andrew Glover's article "In pursuit of code quality: Adventures in behavior-driven development" at http://www.ibm.com/developerworks/java/library/j-cq09187/ which describes a new-and-improved take on TDD called BDD.[Read More
It's customary to start a blog by describing the vision for it. Although this vision will undoubtedly evolve over time, it's always good to put a stake in the ground to get things started. Agile software development is clearly taking off and in my opinion is becoming the dominant development paradigm. Furthermore it appears that Agile approaches enjoy a higher success rate, providing better value for your IT investment, than do traditional approaches. Although organizations are succeeding at simpler projects with agile, many are struggling when applying Agile in more complex situations. They're finding that the "Agile rhetoric" doesn't always live up to its promises once you move into these complex situations. My goal with this blog is to share strategies for applying Agile techniques at scale.
When applying Agile strategies at scale you are likely to run into one or more of the following complexity factors:1. Geographical distribution. Is your team, including stakeholders, in different locations? Even being in different cubicles within the same building can erect barriers to communication, let alone being in different cities or even on different continents.2. Regulatory compliance. Regulations, including the Sarbanes-Oxley act, BASEL-II, and FDA statutes, to name a few, can increase the documentation and process burden on your projects. Complying to these regulations while still remaining as agile as possible can be a challenge.3. Entrenched policies, people, and processes. Most agile teams need to work within the scope of a larger organization, and that larger organization isn't always perfectly agile. Hopefully that will change in time, but we still need to get the job done right now. Your existing culture and organization can really hinder your ability to scale agile approaches, then a few "simple" changes can really help your efforts.4. Legacy systems. Although the politically correct term would be "proven assets" the reality is that it can be very difficult to leverage existing code and data sources due to quality problems. The code may not be well written, documented, or even have tests in place, yet that doesn't mean that your agile team should rewrite everything from scratch. Some legacy data sources are questionable at best, or the owners of those data sources difficult to work with, yet that doesn't given an agile team license to create yet another database.5. Organizational distribution. When your teams are made up of people working for different divisions, or if you have people from different companies (such as contractors, partners, or consultants), then your management complexity rises.6. Degree of governance. If you have one or more IT projects then you have an IT governance process in place. How formal it is, how explicit it is, and how effective it is will be up to you. IBM has been doing a lot of work in this topic over the past few years, and just recently Per Kroll and I have done some work around Lean Governance strategies. 7. Team size. Large teams will be organized differently than small teams, and they'll work differently too.8. System complexity. The more complex the system the greater the need for a viable architectural strategy. An interesting feature of the Rational Unified Process (RUP) is that it's Elaboration phase's primary goal is to prove the architecture via the creation of an end-to-end, working skeleton of the system. This risk-reduction technique is clearly a concept which Extreme Programming (XP) and Scrum teams can clearly benefit from.
It is definitely possible to scale Agile software development to meet the real-world complexities faced by modern organizations. Based on my experiences, I believe that over the next few years we'll discover that Agile scales better than traditional approaches. Many people have already discovered this, but as an industry I believe that there isn't yet sufficient evidence to state this as more than opinion. My goal with this blog is to provide advice for scaling Agile so as to increase your chances of success.
So, it looks like I have my work cut out for me. My strategy will be to address common questions which I get when working with customers and with internal IBM development teams. I have the privilege to work with a variety of software development teams worldwide, helping them to become more agile. They're all struggling with the same basic issues although don't recognize it because they're too focused on their own situation. So hopefully I'll be able to spread the word about what's actually working in practice.
I hope that you stay tuned.
- Scott[Read More
Yesterday I was involved with a workshop around agile development at scale. At one point in the conversation we started talking about the relationship between cost and quality. Some of the people in the workshop were relatively new to agile and still believed the traditional theory that to build in high quality it costs more, sometimes substantially more. This does appear to be true on traditional waterfall projects, but some people were making the mistake that this was an "natural law of IT" which also must apply to agile project teams. I naturally jumped on that idea and described how agile developers have found that writing high quality code leads to lower development costs and shorter time to value, in direct contradiction to traditional theory. A few people struggled with the idea for a bit, and one was pretty adamant that in some cases the need for very high quality does in fact lead to greater cost and time. He talked about his experiences on large-scale Rational Unified Process(RUP)
projects and in particular how some URPS (usability, reliability, performance, and supportability) requirements can increase your cost. At this point Per Kroll, co-author of Agility and Discipline Made Easy: Practices from OpenUP and RUP
, jumped into the conversation and pointed out although higher quality does lead to lower cost in most cases, using Toyota's lean approach to manufacturing as an example, that the agile community didn't completely have the relationship between quality and cost completely correct. My spidey sense told me that a learning opportunity was coming my way.
Per and I had an offline discussion about this to explore what he'd been observing in practice. In most situation it appears to be the case that higher quality does in fact lead to lower costs and shorter time for delivery, something that Per and I had observed numerous times. This happens because high quality code is much easier to understand and evolve than low quality code -- the agile community has found that it is very inexpensive to write high quality code by following practices such as continuous integration
, developer regression testing [or better yet test-driven development(TDD)
], static code analysis
, following common development conventions, and agile modeling strategies
. When you "bake in" quality from the start through applying these techniques, instead of apply traditional techniques such as reviews
and end-of-lifecycle testing (which is still valid for agile projects, but should not be your primary approach to testing) which have long feedback cycles
and therefore prove costly in practice. But, as we've learned time and again, when you find yourself in more complex situations of Agility@Scale sometimes the mainstream agile strategies fall down. For example, in situations where the regulatory compliance scaling factor is applicable, particularly regulations around protecting human life (i.e. the FDA's CFR 21 Part 11), you find that some of the URPS requirements require a greater investment in quality which can increase overall development cost and time. This is particularly true when you need to start meeting 4-nines requirements (i.e. the system needs to be available 99.99% of the time) let alone 5-nines requirements or more. The cost of thorough testing and inspection can rise substantially in these sorts of situations.
In conclusion, it does seem to be true in the majority of situations, which is what the level 1 rhetoric focuses on, that higher quality leads to lower development costs. But at scale this doesn't always seem to hold true.
PS -- Sorry for the corny title, but a couple of days ago at the Rational Software Conference I had the pleasure of interviewing Jamie Hyneman and Adam Savage from the Discovery Channel's Mythbuster's show as part of the conference keynote. They're great guys, BTW, who have had a really positive impact on motivating children to be interested in science (apparently kids like to see stuff get blown up, go figure).[Read More
Recently I visited a customer who had adopted Scrum. They were a few sprints, what Scrum calls iterations, into the project and were running into some difficulties. Although I was primarily brought in to educate senior management on disciplined agile software development, I was also asked to sit in on the team’s daily stand-up meeting so that I could hopefully provide some suggestions as to how to address the problems they were running into.
Their work area was fairly typical. They had some whiteboards which they were using for project planning and tracking, with sticky notes to indicate what work had been taken on by each team member. The current status of the task (not yet started, in progress, and completed) was indicated by putting each sticky note in a corresponding column for the status and corresponding column for the team member. This allowed everyone on the team to easily share their status and to see the status of everyone else. On the sides were sketches of the architecture as well as some business oriented models. In addition to Scrum the team had adopted several practices from Agile Modeling, in this case they had done some initial requirements envisioning
and architecture envisioning
, as well as practices from Extreme Programming (XP) for construction. In short, they had followed a fairly common strategy of combining practices from various agile methods.
This would have worked perfectly fine if they had tailored the practices to reflect the situation that they were in, but instead they adopted them "straight out of the book". First, the team was distributed, with most of the team in the location that I was visiting but some people located in two other distant cities. Therein was the source of most of their problems. The people at the other two locations weren’t getting much value out of the daily stand-up meetings, even though they would dial in, because they couldn’t see the project status information. Although people at this location were trying their best to represent these distant people in the daily stand-ups it wasn’t working well – their status information wasn’t being kept up to date and for some people it was a bit of mystery as to what they were actually working on at all.
This team also had 30 people in it, which isn’t a big deal although it can stretch the limits of the simple modeling and planning tools (in this case paper and whiteboards) that they were using. Because the team was larger they were investing a fair bit of time creating burn down charts at both the iteration/sprint and project levels. One of the unfortunate implications of using manual tools for project management is that any associated metric/status reporting in turn becomes manual as well. Considering how the agile community is so concerned with working efficiently, I find it comical that we have a tendency to overlook our own potentially unnecessary bureaucracy such as this.
The problem was that the team was applying strategies, in this case using sticky notes and whiteboards to capture the detailed iteration plan, applying similar strategies to capture key models, and were verbally relaying of status information between sub-teams. There are perfectly fine strategies for smaller co-located teams, but not so good for large or distributed teams. The solution was to recognize that they were in an Agility@Scale situation and needed to tailor their approach to reflect this fact. In this case they needed to forgo some of the manual tools and instead use electronic tooling such as Rational Team Concert (RTC) to share information across disparate locations, in particular the work assignment and corresponding status information. RTC also creates common agile reports such as burn-down charts based on the activities of the developers, providing accurate (nearly) real-time information while removing the burden of status reporting. The RTC project dashboard does more than just this, to see an actual example of one visit www.jazz.net
to see the dashboard for the RTC development team itself. You can also see their actual work item list too, a more advanced version of Scrum’s product and sprint backlogs.[Read More
My new paper Scaling Agile: An Executive Guide
is now available. As the title suggests the paper overviews how to scale agile strategies to meet your organization's unique needs. The executive summary:
Agile software development is a highly collaborative, quality-focused approach to software and systems delivery, which emphasizes potentially shippable working solutions produced at regular intervals for review and course correction. Built upon the shoulders of iterative development techniques, and standing in stark contrast to traditional serial or sequential software engineering methods, agile software delivery techniques hold such promise that IBM has begun to adopt agile processes throughout its Software Group, an organization with over 25,000 developers. But how can practices originally designed for small teams (10-12) be “scaled up” for significantly larger operations? The answer is what IBM calls “agility@scale.”
There are two primary aspects of scaling agile techniques that you need to consider. First is scaling agile techniques at the project level to address the unique challenges individual project teams face. This is the focus of the Agile Scaling Model (ASM).
Second is scaling your agile strategy across your entire IT department, as appropriate. It is fairly straightforward to apply agile on a handful of projects, but it can be very difficult to evolve your organizational culture and structure to fully adopt the agile way of working.
The Agile Scaling Model (ASM) defines a roadmap for effective adoption and tailoring of agile strategies to meet the unique challenges faced by a software and systems delivery team. Teams must first adopt a disciplined delivery lifecycle
that scales mainstream agile construction techniques to address the full delivery process, from project initiation to deployment into production. Then teams must determine which agile scaling factors
– team size, geographical distribution, regulatory compliance, domain complexity, organizational distribution, technical complexity, organizational complexity, or enterprise discipline, if any — are applicable to a project team and then tailor their adopted strategies accordingly to address their specific range of complexities.
When scaling agile strategies across your entire IT organization you must effectively address five strategic categories — the Five Ps of IT
: People, principles, practices, process, and products (i.e., technology and tooling). Depending on your organizational environment the level of focus on each area will vary. What we are finding within many organizations, including IBM, is that the primary gating factor for scaling agile across your entire organization is your organization’s ability to absorb change.
Timo Tenhunen has recently published his master's thesis, Challenges in Scaling Agile Software Development
, and has been kind enough to make it available online. I suspect you'll find it to be an interesting read.