quality is an important characteristic of a software product. be quality not happen spontaneously, but must actively constructed
XenServer technology does this by our "Quality Engineering" process that goes something .:
- [1945008definieren] and agree quality targets for XenServer release
- define and plan resource objectives (the "quality plan") to make
- tracking and reporting progress of the quality plan under corrective action where necessary
And because quality is too important to the quality we leave fairies a "quality manager" appointed for all this to be responsible.
The Quality Manager ensures quality has an equal place at the table when release decisions are made, we fall from stopping in the event of too often silent in favor of quality scope or time trade. I have previously written about this in detail. In this blog we take a look into the quality plan for Dundee, a XenServer release currently in development.
Dundee quality objectives
First and most importantly-we have a number of quality targets directly to Dundee release regarding the quality. The core target is in terms of customer Raised Unique defects (CRUDs) after 180 days drafted after publication. As shown in Figure 1, we have to track this metric for a number of messages over a number of years.
The breakthrough of this measure in quality with the XenServer 6.2 release came where paid in quality increased investment with a very significant reduction in the number of CRUDs. 6.5 XenServer maintain this trend of improvement. For Dundee our goal, at least to the quality of XenServer is match 6.5, as measured by CRUDs. This is a more aggressive goal than the exciting and complex functions, where developed sounds for XenServer Dundee
1 -. XenServer CRUDs since 2010 
When CRUDs are a lagging indicator (we do not know whether we have met the target, up to 180 days after publication) we can use the CRUD goal of "in project" objectives complement, which we believe to be positively correlated with CRUDs. The most important we use is a function of in Project impurity levels. We have rich historical data about how the number of unresolved defects at different stages of XenServer projects (especially at the time of publication), a map to CRUDs.
Furthermore, we have to evaluate a methodology for defects and to focus our efforts on those who are most likely to result in CRUDs, or generally in customer dissatisfaction. This also ensures we minimize effort spent fixing defects that are unlikely to affect our customers. We also break our total CRUD goal into subgoals specific product areas targeting, functions or abilities to focus attention where it is most needed and where we get the most bang per dollar.
We identify these areas from sources such CRUDs from previous versions and Intel to analyze from our technical support teams. In Dundee we have a general goal does not regress our performance and scale KPIs as well as some specific improvement goals.
We also have a number of "indirect" quality objectives. These are objectives to engineering efficiency and effectiveness continue to optimize the use of our project budget and thus ensure that we devote a greater proportion of our energy to building great features with great quality. In other words, these objectives have the side effect of making our publications more predictable and easier to achieve our 'direct' quality objectives.
For Dundee, we have targets to improve target "in the project" test run rates and quality reporting. We also have goals for a metric we call "automation signal-to-noise ratio," the proportion of test case failures ascribed importance product defect rather than test infrastructure problems.
Dundee Quality Plan
Great, we got some goals. But how to achieve it? The Dundee quality plan includes a number of measures to ensure we meet our quality objectives and a series of detailed quality criteria are met at different internal milestones throughout the project duration.
Before the project implementation under way, we have the quality plan are required reviewed and agreed by the engineering team as a whole, with commitment to provide all necessary resources available. Example actions in Dundee Quality Plan are:
- using static analysis enhance error to find early
- by systematizing new techniques System test planning to improve on complex functions piloted on XenServer 6.5
- extend the reach of our automated interop testing (ie testing new XenServer builds with other Citrix products) to ensure a full program of automated system testing, including functional, stress and performance testing
- is every 2 weeks performed
- staged program of alpha releases and a tech preview
Keeping everyone honest
Even with agreed quality goals and a signed-off quality plan is vigilance required quality at the top remains the guarantee of everyone's mind! For this purpose, published the Quality Manager a bi-weekly quality report. This includes data on the test coverage and test pass rates for the period, lack inflow and outflow statistics and a general report on "quality risks and problems."
This report is a major highlight of the project-review meetings. If the quality manager's view that the team is not the goal, the quality goals then need to take corrective action to be taken is - it is not an option to ignore it, or quietly quality act adhere out release dates. This results in a robust method for ensuring the quality of XenServer on a number of publications proved
Figure 2 -. Fragments two weeks Dundee quality reports, on the report of a left very early a report from one point of the Dev approaches complete milestone in the project and on the right side. Note. The increased amount of test coverage and improved success rates
What is the next version?
The next major release XenServer to Dundee codenamed Ely.
It is not yet time Elys identify quality targets, let alone the quality plan to meet them. But what we can say that we are pursuing our strategy of gradual improvements in both product quality and efficiency in engineering will continue and efficacy.
We will continue the former based on up-to-date feedback from the field and from our tech support experts. We are the latter on our own analysis of engineering KPIs.
An area support, the considerable attention in Dundee but had to work in Ely continues will require that the feedback time between code checkins and error detection to reduce. A fertile area is here that our powerful automated system-level testing to complement with greater unity and component level testing.
success in this area will have a direct positive impact on the efficiency and effectiveness, and thereby have an indirect positive effect on the product quality by the time and effort savings that can be reinvested in improving product quality!
0 Komentar