{"id":433,"date":"2017-07-19T12:15:01","date_gmt":"2017-07-19T12:15:01","guid":{"rendered":"http:\/\/imalogic.com\/blog\/?p=433"},"modified":"2023-08-19T14:32:04","modified_gmt":"2023-08-19T14:32:04","slug":"iot-le-passage-dun-proto-a-un-produit-industriel","status":"publish","type":"post","link":"https:\/\/imalogic.com\/blog\/2017\/07\/19\/iot-le-passage-dun-proto-a-un-produit-industriel\/","title":{"rendered":"IoT – Le passage d’un proto \u00e0 un produit industriel \/ Transition from Prototype to Industrial Product in IoT"},"content":{"rendered":"

Beaucoup de personne sont persuad\u00e9e que le fait d\u2019avoir un prototype qui commence \u00e0 donner des r\u00e9sultats \u00e9quivaut \u00e0 avoir un produit fini \u00e0 80%\u2026<\/span><\/p>\n

Je vous le dis tout de suite : Il n\u2019en est rien !<\/b><\/p>\n

Les \u00abNon Informaticiens\/Vendeurs\/Manager\/Casse couille \u00bb [ Biffer la(les) mention(s) inutile(s) ] ont tendances \u00e0 croire que le boulot est termin\u00e9 par ce qu\u2019ils ont entre leurs mains un prototype qui r\u00e9ponds aux besoins principaux et ce, de fa\u00e7on toute relative. (Oui, il y a encore quelques bugs,\u2026 Mais 90% du boulot est fait\u2026)<\/span><\/p>\n

Pour la plupart, ils ne comprennent pas (ou tendent \u00e0 l\u2019oublier) la d\u00e9marche initiale d\u2019un POC : Celui-ci vise \u00e0 r\u00e9duire les risques, \u00e0 valider certains choix technologiques ou fonctionnelles plut\u00f4t que d\u2019autre\u2026 Bon nombre de vendeur prenne cela pour argent comptant et en profite pour vendre votre \u00ab bidouillage c\u00e9r\u00e9brale \u00bb comme \u00e9tant LA nouvelle 7\u00e8m merveille du monde\u2026 (Oui, il y a encore quelques bugs,\u2026 Mais 95% du boulot est d\u00e9j\u00e0 fait\u2026)<\/span><\/p>\n

Quand on en vient \u00e0 annoncer qu\u2019il faut passer du temps pour transformer ce prototype en v\u00e9ritable produit, il ne r\u00e9alise pas l\u2019effort qu\u2019il reste \u00e0 fournir pour arriver \u00e0 ce but. Cette phase est n\u00e9cessaire pour r\u00e9duire les risques li\u00e9s \u00e0 la phase d\u2019industrialisation. En effet, produire quelques exemplaires d\u2019un syst\u00e8me embarqu\u00e9, ou en produire des centaines de milliers d\u2019exemplaires, ce n\u2019est pas tout \u00e0 fait la m\u00eame chose : une s\u00e9rie de nouveaux facteurs li\u00e9es \u00e0 la production de masse doivent entrer en ligne de compte. ( Oui,il a encore quelques bugs,\u2026 Mais 95% (1.500.000 pi\u00e8ces ) ont d\u00e9j\u00e0 \u00e9t\u00e9 livr\u00e9 \u00e0 travers le monde 8-\/ )<\/span><\/p>\n

Bien d\u00e9finir les sp\u00e9cifications et r\u00e9aliser un mode d\u2019emploi permet dans un premier temps de mettre en \u00e9vidence certaines failles dans le produit : on peut, par exemple,\u00a0le comparer \u00e0 ce qui existe d\u00e9j\u00e0 sur le march\u00e9 et prendre des dispositions si sur le papier il est d\u00e9j\u00e0 obsol\u00e8te par rapport \u00e0 la concurrence. Il ne faut pas oublier qu\u2019a ce stade, ce produit doit encore \u00eatre d\u00e9velopp\u00e9 et\u00a0produit, ce qui peut prendre plusieurs mois.\u00a0 On peut aussi\u00a0confronter \u201cla r\u00e9alit\u00e9 techniques\u201d \u00e0\u00a0la demande initiale voulue par le client. Bon nombre de projet ont d\u00fb \u00eatre arr\u00eat\u00e9 avant la fin du d\u00e9veloppement pour cause de probl\u00e8me li\u00e9 \u00e0 des \u00e9l\u00e9ments techniques qui n\u2019ont pas \u00e9t\u00e9 pris en compte d\u00e8s le d\u00e9part.<\/span><\/p>\n

L\u2019exemple qui me vient en t\u00eate \u00e9tant la cr\u00e9ation d\u2019un syst\u00e8me audio embarqu\u00e9 qui \u00e0 n\u00e9cessit\u00e9 le travaille d\u2019une \u00e9quipe d\u2019une dizaine d\u2019ing\u00e9nieur durant 8 mois, qui a d\u00fb \u00eatre arr\u00eat\u00e9 pour un probl\u00e8me d\u2019autonomie. Sur le papier, le syst\u00e8me devrait permettre d\u2019enregistrer de la voix durant 8 heures. Au final, le syst\u00e8me r\u00e9alis\u00e9 tenait \u00e0 peine 1 heure\u2026 (Oui, il a encore quelques bugs,\u2026 )<\/span><\/p>\n

La phase d\u2019industrialisation n\u00e9cessite de prendre pas mal de pr\u00e9caution, et de pr\u00e9voir tous les cas d\u2019utilisation du produit final. Les outils de test int\u00e9gr\u00e9 ou pas au firmware ne doivent pas \u00eatre sous estim\u00e9, ainsi que les outils de configurations, les proc\u00e9dures de test,\u2026 En effet, programmer et configurer 10 unit\u00e9s ou protos, ne demande pas la m\u00eame infrastructure que de g\u00e9rer un flux de plusieurs centaines d\u2019unit\u00e9 par jour. Des outils sp\u00e9cialis\u00e9s doivent \u00eatre d\u00e9velopp\u00e9s afin d\u2019automatiser cette phase. La gestion des retours, le support sont autant de point qui doivent \u00eatre \u00e9tudi\u00e9 avec soin. Et n\u00e9cessite le cas \u00e9ch\u00e9ant des ressources non n\u00e9gligeables (R\u00e9daction de doc, Support t\u00e9l\u00e9phonique, Formation de personnelle (ouvrier\/ing\u00e9nieurs), Gestion de forum, Mise \u00e0 jours correctif de firmware,\u2026 )<\/span><\/p>\n

Quelques autres points int\u00e9ressants \u00e0 aborder lors d\u2019un prochain articles :\u00a0<\/strong><\/p>\n

Quid de la s\u00e9curit\u00e9 au niveau de nos cher petite machine ? As-t-on s\u00e9curiser l\u2019acc\u00e8s \u00e0 l\u2019upgrade du firmware, les donn\u00e9es export\u00e9es sont elles critiques ? N\u00e9cessite-t-elle d\u2019\u00eatre crypt\u00e9e ?<\/span><\/p>\n

Quid de l\u2019abonnement pour des syst\u00e8me n\u00e9cessitant la connexion \u00e0 un r\u00e9seau GSM ? La bande passante est assez large\/ bien la calculer avant\u2026<\/span><\/p>\n

\u00a0<\/p>\n\n\n

Transition from Prototype to Industrial Product in IoT<\/strong><\/h2>\n\n\n\n
Many people are convinced that having a prototype that starts yielding results is equivalent to having a finished product at 80%\u2026<\/p>\n\n\n\n
Let me tell you right away: That\u2019s not the case!<\/p>\n\n\n\n
\u201cNon-IT Professionals\/Salespeople\/Managers\/Pain in the neck\u201d [Cross out irrelevant mention(s)] tend to believe that the job is done because they have a prototype in their hands that meets the main requirements, albeit to some extent. (Yes, there are still a few bugs,\u2026 But 90% of the work is done\u2026)<\/p>\n\n\n\n
For the most part, they don\u2019t understand (or tend to forget) the initial purpose of a Proof of Concept (POC): It aims to reduce risks, validate certain technological or functional choices over others\u2026 Many salespeople take this at face value and seize the opportunity to sell your \u201cbrainchild hack\u201d as the new 7th wonder of the world\u2026 (Yes, there are still a few bugs,\u2026 But 95% of the work is already done\u2026)<\/p>\n\n\n\n
When it comes to announcing that time is needed to transform this prototype into a real product, they fail to grasp the effort still required to achieve this goal. This phase is necessary to mitigate risks associated with the industrialization phase. Indeed, producing a few units of an embedded system versus producing hundreds of thousands of units is not quite the same: a series of new factors related to mass production must be taken into account. (Yes, there are still a few bugs,\u2026 But 95% (1,500,000 units) have already been delivered worldwide 8-\/)<\/p>\n\n\n\n
Clearly defining specifications and creating an instruction manual initially highlights some flaws in the product: for instance, it can be compared to what already exists in the market, and adjustments can be made if it\u2019s already outdated on paper compared to the competition. It\u2019s important to remember that at this stage, the product still needs to be developed and manufactured, which can take several months. The \u201ctechnical reality\u201d can also be compared to the initial client requirements. Many projects have had to be stopped before development was completed due to technical elements not considered from the outset.<\/p>\n\n\n\n
An example that comes to mind is the development of an embedded audio system that required the work of a team of about ten engineers for 8 months, only to be halted due to battery life issues. On paper, the system was supposed to record voice for 8 hours. In the end, the implemented system barely lasted 1 hour\u2026 (Yes, there are still a few bugs,\u2026)<\/p>\n\n\n\n
The industrialization phase requires taking many precautions and anticipating all use cases of the final product. Integrated or standalone firmware testing tools should not be underestimated, nor should configuration tools, testing procedures, etc. After all, programming and configuring 10 units or prototypes doesn\u2019t require the same infrastructure as managing a flow of several hundred units per day. Specialized tools need to be developed to automate this phase. Handling returns, providing support, and addressing these points with care are essential. It might require significant resources if necessary (Documentation writing, Telephone support, Personnel training [workers\/engineers], Forum management, Firmware patch updates,\u2026)<\/p>\n\n\n\n
A few other interesting points to address in a future article:<\/p>\n\n\n\n
What about security at the level of our dear little devices? Have we secured firmware upgrade access? Are the exported data critical? Do they need encryption?<\/p>\n\n\n\n
What about subscriptions for systems requiring GSM network connectivity? Is the bandwidth sufficient? Calculate it carefully beforehand\u2026<\/p>\n<\/body>","protected":false},"excerpt":{"rendered":"
Beaucoup de personne sont persuad\u00e9e que le fait d\u2019avoir un prototype qui commence \u00e0 donner des r\u00e9sultats \u00e9quivaut \u00e0 avoir<\/p>\n","protected":false},"author":1,"featured_media":213,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[7,8,9],"tags":[],"class_list":["post-433","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-coding","category-embedded","category-iot"],"jetpack_publicize_connections":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/imalogic.com\/blog\/wp-content\/uploads\/2016\/11\/ibm_iot_trial-avatar.png?fit=180%2C176&ssl=1","jetpack_sharing_enabled":true,"jetpack_shortlink":"https:\/\/wp.me\/p8J21V-6Z","jetpack-related-posts":[],"_links":{"self":[{"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/posts\/433","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/comments?post=433"}],"version-history":[{"count":2,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/posts\/433\/revisions"}],"predecessor-version":[{"id":1117,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/posts\/433\/revisions\/1117"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/media\/213"}],"wp:attachment":[{"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/media?parent=433"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/categories?post=433"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/imalogic.com\/blog\/wp-json\/wp\/v2\/tags?post=433"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}