BEGIN:VCALENDAR VERSION:2.0 PRODID:-//Talks.cam//talks.cam.ac.uk// X-WR-CALNAME:Talks.cam BEGIN:VEVENT SUMMARY:Composition and Inheritance in Declarative Configuration Languages - Paul Anderson\, University of Edinburgh DTSTART:20170518T090000Z DTEND:20170518T100000Z UID:TALK72542@talks.cam.ac.uk CONTACT:Microsoft Research Cambridge Talks Admins DESCRIPTION:Configuring large computing installations is a difficult probl em -- there are many different subsystems involved\, all with their own la nguage\, and many different people with an interest in overlapping aspects of the overall configuration. Deploying and maintaining a configuration w hich reliably meets everyone's "higher-level" requirements is hard\, and configuration errors are responsible for a significant proportion of syste m failures.\n\nAlmost all large installations use some type of "configurat ion tool" (*) in an attempt to automate and unify the process. These often involve some kind of custom specification language\, which is often claim ed to be "declarative". But all practical configuration languages have dev eloped in very informal ways\, with a complex semantics which are error-pr one and difficult to use. Declarative configuration languages in particula r\, are quite different from programming languages\, and have not been wid ely studied in any formal way\, despite their critical role in most large infrastructures.\n\nFor a long time\, I have been interested in understand ing the overall configuration process\, and particularly in designing more usable languages\, which provide specific features to support this in a c lear in and correct way. Although there is little directly related work\, this has led to interesting connections with a number of other areas\, tw o of which have been the subject of previous Microsoft Research Phd Schola rships:\n\n- The use of constraint-based specifications[1] (with Andy Gord on) particularly the novel use of these of these to support conflict-free composition of partial specifications from different stake-holders\n\n- Th e use of provenance techniques (from the database field) to track the depe ndencies between final configuration parameters and the source specificati ons which affect them[2] (with Dimitrios Vytiniotis). This has implication s for fault diagnosis\, and security (which is not addressed well\, if at all\, in current languages)\, and it suggests new ways of resolving potent ial conflicts.\n\nI have also been involved in the deployment of the resul ting configurations\, particularly distributed agent-based approaches usin g automated planning.\n\nRecently\, we have been trying to formalise the s emantics of practical declarative configuration languages[3\,4] and the di fficulty of this has highlighted some of the reasons why this type of lang uage has acquired an (unnecessary) reputation for complexity. This has led to an interest in human-factors work on the usability of configuration la nguages[5]\, and it seems that these considerations should be a fundamenta l part of the design process for any new language - I have a (currently un funded) proposal for a Phd in this area[6].\n\nI plan to use some recent w ork on composition of configurations as the focus for a talk which will dr aw on various aspects of this previous work. The short note [7] is an acc essible description of this for a non-specialist audience. However\, I wou ld like to share my experiences with the configuration problem in general to see if there was any further possible synergy\, and I would be happy to focus more on any aspects which may be of interest\, and to discuss anyth ing else which may potentially be of common interest.\n\nPaul Anderson \nhttp://homepages.inf.ed.ac.uk/dcspaul\n\n(*) SOME CONFIG URATION TOOLS\n\n- "Puppet" (https://puppetlabs.com/) is probably the most common "declarative" tool.\n\n- "Cfengine" (http://cfengine.com/) was one of the first-generation tools which claimed to be "declarative".\n\n- "Sm artFrog" (http://www.smartfrog.org/) probably has the cleanest semantics\, but has never been widely used outside of HP.\n\n- "LCFG" (http://www.lcf g.org/) was developed here in Informatics\, and is probably the oldest "de clarative" tool. But the "language" is very primitive\, and the tool is no t widely used outside of the University.\n\n- Chef (https://www.chef.io/) is derived from Puppet\, but takes a more imperative approach\, probably d ue to the apparent difficulty of using the Puppet language.\n\n- "Ansible" (https://www.ansible.com/) is a more imperative approach focussing on "ta sks" which are executed in order" - this is claimed to be more simple than a declarative approach.\n\n- Microsoft Powershell (https://msdn.microsoft .com/en-us/powershell/mt173057.aspx) is an an imperative language and fram ework for scripting configuration changes\, however DSC ("desired state co nfiguration") introduces some declarative features.\n\nREFERENCES\n\nSumma ry: The short note [7] is an informal introduction to the composition prob lem which avoids the distracting details of ``real'' configurations\, and the technicalities of ``real'' configuration languages. [4] is a good intr oduction to the difficulties and complexities of ``real'' languages\, and the more formal work on the semantics. [2] is a short position paper which illustrates some usability problems and motivates the current work proven ance work.\n\n[1] Constraint-based autonomic configuration\n2013 Self-adap tive and self-organizing systems conference (SASO)\, September 2013\nJohn A Hewson & Paul Anderson & Andrew D Gordon\nhttp://homepages.inf.ed.ac.uk/ dcspaul/homepage/live/pdf/saso-2013.pdf\n\n[2] Toward provenance-based sec urity for configuration languages\nThe 4th Usenix workshop on the theory a nd practice of provenance\, June 2012\nPaul Anderson & James Cheney\nhttp: //homepages.inf.ed.ac.uk/dcspaul/homepage/live/pdf/tapp12-final15.pdf\n\n[ 3] A formal semantics for the SmartFrog configuration language\nJournal of Network and Systems Management\nPaul Anderson & Herry Herry\nhttp://homep ages.inf.ed.ac.uk/dcspaul/homepage/live/pdf/sfsemantics.pdf\n\n[4] μPuppe t: A Declarative Subset of the Puppet Configuration Language\nTechnical re port\, December 2016\nPaul Anderson\, James Cheney\, Weili Fu & Roly Perer a\nhttp://homepages.inf.ed.ac.uk/dcspaul/homepage/live/pdf/arxiv-puppet.pd f\n\n[5] Usability of System Configuration Languages: Confusion caused by Ordering\nAdele Mikoliunaite\nhttp://homepages.inf.ed.ac.uk/dcspaul/homepa ge/live/pdf/msc_20162088.pdf\n\n[6] Usability of System Configuration Lang uages - IGS Studentship Proposal\nhttp://homepages.inf.ed.ac.uk/dcspaul/ho mepage/live/pdf/usability-phd.pdf\n\n[7] Composition in the L3 Configurati on Language\nhttp://homepages.inf.ed.ac.uk/dcspaul/homepage/live/pdf/l3com position.pdf\n\n LOCATION:Small Lecture Room \, Microsoft Research Ltd\, 21 Station Road\, Cambridge\, CB1 2FB END:VEVENT END:VCALENDAR