BEGIN:VCALENDAR VERSION:2.0 PRODID:-//HasGeek//NONSGML Funnel//EN DESCRIPTION:On SRE\, systems engineering and distributed systems X-WR-CALDESC:On SRE\, systems engineering and distributed systems NAME:Rootconf Hyderabad edition X-WR-CALNAME:Rootconf Hyderabad edition REFRESH-INTERVAL;VALUE=DURATION:PT12H SUMMARY:Rootconf Hyderabad edition TIMEZONE-ID:Asia/Kolkata X-PUBLISHED-TTL:PT12H X-WR-TIMEZONE:Asia/Kolkata BEGIN:VEVENT SUMMARY:Check-in and onsite registrations DTSTART:20191116T034500Z DTEND:20191116T040000Z DTSTAMP:20260421T020940Z UID:session/RXU7XPY3tZFkrMXyUmupXM@hasgeek.com SEQUENCE:0 CREATED:20190628T095305Z DESCRIPTION:\n LAST-MODIFIED:20191106T044624Z LOCATION:Hyderabad ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Check-in and onsite registrations in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Introduction to Rootconf\; how to navigate the schedule DTSTART:20191116T040000Z DTEND:20191116T041000Z DTSTAMP:20260421T020940Z UID:session/3SZgKaofHWkpo7tTsCVLcs@hasgeek.com SEQUENCE:0 CREATED:20190628T095339Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20191111T043359Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Introduction to Rootconf\; how to navigate the schedule in Mai n auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Observability and control theory DTSTART:20191116T041000Z DTEND:20191116T045000Z DTSTAMP:20260421T020940Z UID:session/BgJgBnLfi9nB8gPkCJ7oXc@hasgeek.com SEQUENCE:3 CATEGORIES:SRE,Full talk (40 mins) CREATED:20190819T042750Z DESCRIPTION:- Sample problem that go unnoticed with traditional Blackbox m onitoring\n- Sample problem that goes unnoticed with advanced whitebox mon itoring.\n- Timeline of Monitoring tools.\n- Evolution of Monitoring needs .\n- What is monitoring vs Observability?\n- Servers vs.Services.\n- Real questions that Distributed Systems have to answer\n- Domains of Observabil ity\n- Need of Observability\n- Debuggability in advanced systems.\n- Math s of Stability\n- What is Control theory? \n- Practicale example of Contro l Theory\n- Types of Control-Systems\n- How Control theory uses Observabil ity\n- How do we use Control theory\n- How TCP Implements all of this toge ther\n- FLow Control vs Congestion Control\n- Sample implementation in a w orking application\n- Thank you?\n\n### Speaker bio\n\nPiyush Verma heads Site Reliabilty Engineering at TrustingSocial.com. At present\, their syst ems span across 6 datcenters and 3 major clouds across 3 countries.\nHe ha s been working with Infrastrcuture Problems for over a decade and in past life he has built datscale.io\, oogway.in\, and siminars.com.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123109Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/observability-and -control-theory-BgJgBnLfi9nB8gPkCJ7oXc BEGIN:VALARM ACTION:display DESCRIPTION:Observability and control theory in Main auditorium in 5 minut es TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Fast object distribution using P2P DTSTART:20191116T045000Z DTEND:20191116T052000Z DTSTAMP:20260421T020940Z UID:session/XtmR2PHucZNpd8x2Jg96EK@hasgeek.com SEQUENCE:3 CATEGORIES:Full talk,Beginner,Discussion,Distributed systems,Full talk (40 mins) CREATED:20190716T045152Z DESCRIPTION:* Motivation for fast object distribution\n* Overview of bitto rrent protocol\n* Goals for a torrent-based distribution layer\n* Design\, architecture and functional flows of Shatabdi\n - Indexing store and l eader election using etcd\n - Achieving HA for torrent tracker\n - C omparison of open-source torrent clients\n* API contracts of daemon binary which abstracts http and p2p downloads\n* Performance tuning of QBittorre nt client and uncharted territories of libtorrent\n* Benchmarking containe r deployments powered by Shatabdi\n\n### Speaker bio\n\nAnkur and Yashpal work in Platform-as-a-service team in Flipkart Cloud Platform. We build de v-centric platforms that work at Flipkart scale. We busy ourselves with op timizing systems that are best left alone\, defying CAP and not succeeding and sneaking GOT references in sprint planning sessions.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123150Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/shatabdi-p2p-laye r-for-fast-object-distribution-XtmR2PHucZNpd8x2Jg96EK BEGIN:VALARM ACTION:display DESCRIPTION:Fast object distribution using P2P in Main auditorium in 5 min utes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Morning beverage break DTSTART:20191116T052000Z DTEND:20191116T055000Z DTSTAMP:20260421T020940Z UID:session/68Ln36fhsYLe3yYKWgL8hp@hasgeek.com SEQUENCE:0 CREATED:20190628T095728Z DESCRIPTION:\n LAST-MODIFIED:20191106T044713Z LOCATION:Hyderabad ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Morning beverage break in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Taming infrastructure workflow at scale DTSTART:20191116T055000Z DTEND:20191116T063000Z DTSTAMP:20260421T020940Z UID:session/YE9vdH7gaB83AUzKCFWRbw@hasgeek.com SEQUENCE:3 CATEGORIES:Awaiting reviewer to evaluate,Automation,Full talk (40 mins) CREATED:20190904T052234Z DESCRIPTION:As more operations choices are added to your data center\, whe ther through company acquisitions\, a growing development team\, or genera l technical debt\, managing infrastructure complexity becomes a nightmare. Yet the end goal is still the same — safely deploy your application to your infrastructure. We need to tame our data centers by managing change a cross systems\, enforcing policies\, and by establishing a workflow for bo th developers and operations engineers to build in a collaborative environ ment.\n\nThis talk will discuss the problems faced in managing a modern cl oud infrastructure\, and how a set of innovative open source tools like Te rraform can be used to tame the rising complexity curve.\n\nJoin me as I t ake you on a journey of exploring Infrastruture as Code techniques as we t ake control of our cloud infrastructure. This goal of this demo driven tal k is to showcase how you can build multi-tier application infrastructure s upporting multiple cloud platforms and services using IAC.\n\n### Speaker bio\n\nAnubhav Mishra is a Technical Advisor to the CTO at HashiCorp. He i s passionate about Developer Advocacy and helping developers and operators do better. \nPreviously\, he worked at [Hootsuite](https://hootsuite.com) \, where he created [Atlantis](https://www.runatlantis.io) - An Open Sourc e project that helps teams collaborate on Infrastructure using [Terraform] (https://terraform.io). Anubhav loves working with distributed systems and sharing his experience with microservice delivery platforms. He also love s open source software and is continuously finding ways to contribute to p rojects that excite him and helping developers and operators do better. Th at has led him to contribute to projects like [Virtual Kubelet](http://vir tual-kubelet.io) and [Helm](https://helm.sh) (CNCF projects). In his free time\, he DJs\, makes music and plays football. He's a huge [Manchester Un ited](https://www.manutd.com) supporter.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123145Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/taming-infrastruc ture-workflow-at-scale-YE9vdH7gaB83AUzKCFWRbw BEGIN:VALARM ACTION:display DESCRIPTION:Taming infrastructure workflow at scale in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Achieving repeatable\, extensible and self serve infrastructure at Gojek DTSTART:20191116T063000Z DTEND:20191116T071000Z DTSTAMP:20260421T020940Z UID:session/Cs9s5nd6wDNaz7ePcAhJfV@hasgeek.com SEQUENCE:3 CATEGORIES:Automation,Full talk (40 mins) CREATED:20190823T061709Z DESCRIPTION:*Goals\n*Architecture and discussion around\n- olympus which h osts our cloud infrastructure config \n- terraform module structuring and how we run it our CI/CD platform\n- how proctor (https://github.com/gojek/ proctor) helps us in achieving the self serve model for infrastructure \n* Lessons learnt\n*Impact\n\n### Speaker bio\n\nTasdik is a Product Engineer at Gojek where he works with the systems team. Contributor to oVirt (unde r Redhat)\, before Gojek he was part of the early systems team of Razorpay . He has presented his talks at different national and international confe rences including Pycon Taiwan\, Devopsdays India and devconf India. A full list of the talks and slides/videos can be found at https://tasdikrahman. me/speaking/.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123140Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/achieving-repeata ble-extensible-and-self-serve-infrastructure-at-gojek-Cs9s5nd6wDNaz7ePcAhJ fV BEGIN:VALARM ACTION:display DESCRIPTION:Achieving repeatable\, extensible and self serve infrastructur e at Gojek in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of Feather (BOF) session: On AIOps DTSTART:20191116T063000Z DTEND:20191116T072500Z DTSTAMP:20260421T020940Z UID:session/BCdL26pZ5FWhUiBwgMmAwJ@hasgeek.com SEQUENCE:0 CREATED:20191022T061034Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20191106T045345Z LOCATION:Birds Of Feather (BOF) session area - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Birds of Feather (BOF) session: On AIOps in Birds Of Feather ( BOF) session area in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Lunch break DTSTART:20191116T071000Z DTEND:20191116T080000Z DTSTAMP:20260421T020940Z UID:session/HkNZwHS8RLX3DS6fM84AjQ@hasgeek.com SEQUENCE:0 CREATED:20190628T095832Z DESCRIPTION:\n LAST-MODIFIED:20191106T044722Z LOCATION:Hyderabad ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Lunch break in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Deploying and managing CSP: the browser-side firewall DTSTART:20191116T080000Z DTEND:20191116T084000Z DTSTAMP:20260421T020940Z UID:session/5Qo1NFim6oG8gxNv6Qgwpn@hasgeek.com SEQUENCE:3 CATEGORIES:DevSecOps,Full talk (40 mins) CREATED:20190722T053124Z DESCRIPTION:- Introduction to client-side Data Exfiltration attacks\n- Int roduction to Content-Security Policy\n- Content Security Policy to prevent Data Exfiltration attacks\n ○ What is possible\n ○ What are the lim itations\n- How to design and deploy CSP to detect/prevent Data Exfiltrati on attacks \n- How to monitor policy violations and alerts\n\n### Speaker bio\n\nLavakumar Kuppan is the founder of Ironwasp Security. He is a secur ity researcher and a product developer. He is currently developing product s that automatically perform vulnerability detection and attack monitoring for the DOM. \nHe has done extensive research on web security with specia l focus on JavaScript security. He has discovered several novel attacks ve ctors and vulnerabilities and has spoken about his research in several int ernational conferences.\nHe has also done extensive work on developing ope n source tools to discover security issues in JavaScript\, through both st atic and dynamic analysis.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123135Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/deploying-and-man aging-csp-the-browser-side-firewall-5Qo1NFim6oG8gxNv6Qgwpn BEGIN:VALARM ACTION:display DESCRIPTION:Deploying and managing CSP: the browser-side firewall in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of Feather (BOF) session: On Infrastructure as Code DTSTART:20191116T080000Z DTEND:20191116T090000Z DTSTAMP:20260421T020940Z UID:session/FvaP7RuRKvwtKJPRZmNMBw@hasgeek.com SEQUENCE:0 CREATED:20191022T061234Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20191106T044859Z LOCATION:Birds Of Feather (BOF) session area - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Birds of Feather (BOF) session: On Infrastructure as Code in B irds Of Feather (BOF) session area in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of Feather (BOF) session: Doing DevSecOps in your organizati on DTSTART:20191116T084000Z DTEND:20191116T092000Z DTSTAMP:20260421T020940Z UID:session/HBK4o7dBGsamaR429J3cwj@hasgeek.com SEQUENCE:1 CREATED:20191022T061413Z DESCRIPTION:In this session\, we will discuss:\n\n1. Are you planning to i mplement DevSecOps? What challenges are you facing in the initial stages?\ n2. What is your DevSecOps implementation plan including tooling\, process etc?\n3. Is "DevSecOps" a new buzzword? Can **security** piggyback DevSec Ops?\n4. Does Devsecops provide value for the dollars invested?\n5. Are yo u handling false positives with automation? what are the challenges in the process?\n6. How do you measure a successful "DevSecOps" implementation?\ n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20230108T103046Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Birds of Feather (BOF) session: Doing DevSecOps in your organi zation in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Flash talk: Dashboards as code DTSTART:20191116T092000Z DTEND:20191116T092500Z DTSTAMP:20260421T020940Z UID:session/LRMJ153jfEnxaDqesubPf7@hasgeek.com SEQUENCE:2 CATEGORIES:Crisp talk,Intermediate,Lecture CREATED:20190722T053035Z DESCRIPTION:How to make use of the world's top technology to build your Da shboards with more robustness without tech debt. \n\nWhy dashboards?\n\nEx isting methods of maintaining dashboards.\n\n - At current scale of using dashbaords and alerts\, difficult to scale without proper automation.\n - Consistency issues \n - Dashboards can be easily edited by any layman r esulting in consistency issues.\n - Alerts disbaled during maintenance/ deployment leading to undetected incidents.\n - Automate challenges\n - 90% of the dashboards are still being created/modified by clicking on the UI.\n - Manual configuration can lead to errors.\n - No history.\n - No rollbacks possible during an unintended modification.\n\n\nSolution 1 : Git\n\n - Will solve Consistency issues and maintains history.\n - It st ill cannot validate the huge json files\, manual review is required.\n - A utomation of stored json files to dashboards are still not solved.\n\nSolu tion 2: Git + Terraform\n\n - Terraform providers/plugins. - Automating da shboards using terrafrom providers. \n - Manage state\, conflicts\, suppor t validation\, rollback\, everything with zero tech debt.\n - Add value to your skills.\n\nDemo of how to create a dashboard with Terraform and Graf ana\n\n### Speaker bio\n\nSanooj Mananghat is working as a DataSRE @ Intui t. With 9 years of experience in Devops world. A blogger\, opensource enth usiast with contributions to multiple opensource projects including terraf orm providers.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20230810T072606Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/dashboards-as-cod e-LRMJ153jfEnxaDqesubPf7 BEGIN:VALARM ACTION:display DESCRIPTION:Flash talk: Dashboards as code in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Flash talk: Literate DevOps DTSTART:20191116T092500Z DTEND:20191116T093000Z DTSTAMP:20260421T020940Z UID:session/BuciWBstYogB4ma1qJqSWK@hasgeek.com SEQUENCE:0 CREATED:20191106T052624Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20191106T052946Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Flash talk: Literate DevOps in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Birds of Feather (BOF): On Observability DTSTART:20191116T093000Z DTEND:20191116T103000Z DTSTAMP:20260421T020940Z UID:session/JtV98SaVYapNGTJd8cjxPr@hasgeek.com SEQUENCE:0 CREATED:20191111T043339Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20191111T043350Z LOCATION:Birds Of Feather (BOF) session area - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Birds of Feather (BOF): On Observability in Birds Of Feather ( BOF) session area in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Flash talk: The consensus problem in distributed systems DTSTART:20191116T093000Z DTEND:20191116T093500Z DTSTAMP:20260421T020940Z UID:session/SppoxPFtyzFVHDKwuCP41C@hasgeek.com SEQUENCE:2 CATEGORIES:Flash talk (5 mins),Distributed systems CREATED:20191106T052726Z DESCRIPTION:What is a consensus in a distributed system?\n---------------- ----------------------------\nIn the context of distributed systems design \, a consensus is often loosely used to mean some form of agreement. Conse nsus involves multiple servers agreeing on values. Once they reach a decis ion on a value\, that decision is final. Typical consensus algorithms make progress when any majority of their servers is available\; for example\, a cluster of 5 servers can continue to operate even if 2 servers fail. If more servers fail\, they stop making progress (but will never return an in correct result).\n\n i.e 2f+1 nodes to survive f failed nodes\n\nThere are a few properties we expect from a solution to consensus:\nAgreement: Ever y correct process must agree on the same value.\nValidity: If all processe s propose the same value v\, then all correct processes decide v\nTerminat ion: Every correct process decides some value. If the protocol never termi nates\, then the processes are vacuously agreeing on the same thing\, whic h is not deciding. \n\nTo summarize\, fundamentally\, the goal of consensu s is not that of the negotiation of an optimal value of some kind\, but ju st the collective agreement on some value that was previously proposed by one of the participating servers in that round of the consensus algorithm. With the help of consensus\, the distributed system is made to act as tho ugh it were a single entity.\n\nAn example scenario:\n-------------------- \nFor the purpose of simplicity\, let's assume a distributed storage syste m with 2f+1 nodes participating to form a cluster and these participants a ct at their own speed\, may fail at any time and rejoin after recovering f rom the failure. And these nodes are connected via a network which transmi ts messages asynchronously at an arbitrary speed. In short\, everything ca n fail at any time\; after failure\, participants can recover and rejoin t he system. Yes\, we are looking at a fault tolerant storage system. As the se nodes can fail at various stages\, it's important to have more than one copy of our data. For now\, let's assume all the data is replicated acros s all the cluster nodes (but in reality it may affect overall performance) \nAnd we have a client which is not part of the cluster\, requesting for some operation from our distributed storage\, like a write or read to a d ata file. Read operation can be served by any node in our cluster\, but wr ite has to be handled by one node at a time to avoid any kind of conflicts . Also\, read of the same data which is being written can be served only a fter the write has been committed. And no other client can request write t o the same data file when a write is in progress. In order to achieve thi s data integrity in our distributed storage\, a member of the cluster shou ld be elected as a leader to accept writes or redirect the read requests t o other cluster members. Wait\, what if the leader which receives the wri te requests loses its leadership? So the node which is handling the write should lock the data file until the write is committed. But how can we ac hieve this? How will the cluster know it’s membership?\n\nWhat is Apache Zookeeper?\n-------------------------\nZooKeeper is a distributed\, open- source coordination service for distributed applications. By providing a r obust implementation of a few basic operations\, ZooKeeper simplifies the implementation of many advanced patterns in distributed systems. \nThe ser vice itself is simple\, replicated\, distributed\, fast\, highly reliable and exposes a file-system-like API so that clients can manipulate simple d ata files called znodes. These znodes can have some special properties lik e being ephemeral or sequential\, but nonetheless\, they are small data fi les.\nEphemeral: Clients create such znodes\, and they either delete them explicitly or let the system remove them automatically when the session th at creates them terminates.\n Sequential: These Znodes when created\, gets a unique number (sequence) suffixed to its name.\nUnlike standard file sy stems\, each znode in a ZooKeeper namespace can have data associated with it as well as children.\nFig 1 shows the high-level components of the ZooK eeper service. With the exception of the request processor\, each of the s ervers that make up the ZooKeeper service replicates its own copy of each of the components.\n\n\nAll requests that update ZooKeeper state are forwa rded to the leader. The leader executes the request and broadcasts the cha nge to the ZooKeeper state using a Zookeeper Atomic Broadcast (ZAB) protoc ol. \nMany prominent Apache projects like Hadoop\, HBase\, Kafka are using Zookeeper to solve their distributed complexities. \n\nHow can we use it? \n------------------\nComing back to our distributed storage cluster\, we can use Zookeeper to elect a leader using sequential znodes and create a w rite lock using Ephemeral znodes. \n\n1. Leader Election\nA simple way of doing leader election with ZooKeeper is to use the SEQUENCE|EPHEMERAL flag s when creating znodes that represent "proposals" of clients. The idea is to have a znode\, say "/election"\, such that each znode creates a child z node "/election/n_" with both flags SEQUENCE|EPHEMERAL. With the sequence flag\, ZooKeeper automatically appends a sequence number that is greater t han anyone previously appended to a child of "/election". The process that created the znode with the smallest appended sequence number is the leade r.\nThat's not all\, though. It is important to watch for failures of the leader so that a new client emerges as the new leader in the case the curr ent leader fails. A trivial solution is to have all application processes watching upon the current smallest znode\, and checking if they are the ne w leader when the smallest znode goes away (note that the smallest znode w ill go away if the leader fails because the node is ephemeral). But this c auses a herd effect: upon of failure of the current leader\, all other pro cesses receive a notification\, and execute getChildren on "/election" to obtain the current list of children of "/election". If the number of clien ts is large\, it causes a spike in the number of operations that ZooKeeper servers have to process. To avoid the herd effect\, it is sufficient to w atch for the next znode down on the sequence of znodes. If a client receiv es a notification that the znode it is watching is gone\, then it becomes the new leader in the case that there is no smaller znode. Note that this avoids the herd effect by not having all clients watching the same znode. Here's the pseudo code:\nLet ELECTION be a path of choice of the applicati on. To volunteer to be a leader:\nCreate znode z with path "ELECTION/n_" w ith both SEQUENCE and EPHEMERAL flags\;\nLet C be the children of "ELECTIO N"\, and i be the sequence number of z\;\nWatch for changes on "ELECTION/n _j"\, where j is the largest sequence number such that j < i and n_j is a znode in C\;\nUpon receiving a notification of znode deletions:\nLet C be the new set of children of ELECTION\;\nIf z is the smallest node in C\, th en execute leader procedure\;\nOtherwise\, watch for changes on "ELECTION/ n_j"\, where j is the largest sequence number such that j < i and n_j is a znode in C\;\n2. Locks\nFully distributed locks that are globally synchro nous\, meaning at any snapshot in time no two clients think they hold the same lock. These can be implemented using ZooKeeper. Clients wishing to ob tain a lock do the following:\nCall create() with a pathname of "_locknode _/lock-" and the sequence and ephemeral flags set.\nCall getChildren() on the lock node without setting the watch flag (this is important to avoid t he herd effect).\nIf the pathname created in step 1 has the lowest sequenc e number suffix\, the client has the lock and the client exits the protoco l.\nThe client calls exists() with the watch flag set on the path in the l ock directory with the next lowest sequence number.\nif exists() returns f alse\, go to step 2. Otherwise\, wait for a notification for the pathname from the previous step before going to step 2.\nThe unlock protocol is ver y simple: clients wishing to release a lock simply delete the node they cr eated in step 1.\n\nThe example discussed here is a simple distributed sys tem for the purpose of this presentation only but nevertheless\, Apache Zo okeeper can be used to solve many distributed system complexities.\n\n### Speaker bio\n\nGeethanjali Eswaran\, DevOps Engineer for Large-Scale Data Cloud in Salesforce. Passionate about distributed computing\, BigData clou d\, Apache projects\, Kerberos protocol\, and many more...\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20230810T072606Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/obtaining-consens us-in-distributed-systems-using-apache-zookeeper-SppoxPFtyzFVHDKwuCP41C BEGIN:VALARM ACTION:display DESCRIPTION:Flash talk: The consensus problem in distributed systems in Ma in auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Flash talk: Supporting utf8 characters in a utf8 mysql table DTSTART:20191116T093500Z DTEND:20191116T094000Z DTSTAMP:20260421T020940Z UID:session/7poaaTqFUEv2giibxJ7JEk@hasgeek.com SEQUENCE:2 CATEGORIES:Crisp talk (20 mins),DevOps CREATED:20191106T052817Z DESCRIPTION:1. How we supported storing utf8 content in our app.\n2. How u pgrading the rails version caused MySQL to run in strict mode causing the problems.\n3. How we evaluated our options and chose the final solution.\n 4. The actual migration and dealing with ad-hoc issues.\n\n### Speaker bio \n\nLead Software Engineer @ Freshworks.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20230810T072606Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/supporting-utf8-c haracters-in-a-utf8-mysql-table-7poaaTqFUEv2giibxJ7JEk BEGIN:VALARM ACTION:display DESCRIPTION:Flash talk: Supporting utf8 characters in a utf8 mysql table i n Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Regulating cloud service providers DTSTART:20191116T094000Z DTEND:20191116T095000Z DTSTAMP:20260421T020940Z UID:session/Pjy9EL5XekR1pXQFxjbseV@hasgeek.com SEQUENCE:2 CATEGORIES:Crisp talk,Beginner,Lecture CREATED:20191121T195348Z DESCRIPTION:* Regulating CSPs\n* TRAI consultation paper\n* Why?\n\n### Sp eaker bio\n\nSrikanth is a software professional with interest in technolo gy policy\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20230810T072606Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/regulating-cloud- service-providers-Pjy9EL5XekR1pXQFxjbseV BEGIN:VALARM ACTION:display DESCRIPTION:Regulating cloud service providers in Main auditorium in 5 min utes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Reading data from database DTSTART:20191116T095000Z DTEND:20191116T095500Z DTSTAMP:20260421T020940Z UID:session/M7NpqigvXVY46xq4z9Gp3Z@hasgeek.com SEQUENCE:0 CREATED:20200801T094601Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20200801T094601Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Reading data from database in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Debugging distributed systems DTSTART:20191116T095500Z DTEND:20191116T100000Z DTSTAMP:20260421T020940Z UID:session/ENvoDrDuWH5kmEFFNZtGHt@hasgeek.com SEQUENCE:0 CREATED:20200801T095648Z DESCRIPTION:\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20200801T095648Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Debugging distributed systems in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Evening beverage break DTSTART:20191116T100000Z DTEND:20191116T101500Z DTSTAMP:20260421T020940Z UID:session/A9j1HpmSf5ZRprSj2VYcbW@hasgeek.com SEQUENCE:0 CREATED:20190628T100028Z DESCRIPTION:\n LAST-MODIFIED:20200801T095630Z LOCATION:Hyderabad ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com BEGIN:VALARM ACTION:display DESCRIPTION:Evening beverage break in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Designing microservices around your data design DTSTART:20191116T101500Z DTEND:20191116T105000Z DTSTAMP:20260421T020940Z UID:session/TPZZ9NNYfm3JLvrZ25fadC@hasgeek.com SEQUENCE:3 CATEGORIES:Full talk,Intermediate,Lecture,Distributed systems,Full talk (4 0 mins) CREATED:20190722T060653Z DESCRIPTION:- Introduction\n- Usecases Definition\n 1. Image Classific ation\n 2. Product Recommendations\n 3. AI Stylist\n- Architecture and workflow of each usecase\n- Merge the workflows into single architect ure\n- Core Platform Architecture v0.1\n- Design Considerations\n 1. L atency & throughput\n 2. Deployment strategy\n 3. datastore design \n- Core platform architecture v0.2\n- Learnings from our design\n- Case s tudy: Microservice Registry\n 1. Problem statement\n 2. Architectu re\n 3. Config change for single service\n 4. Config change betwee n microservices\n- Questions ?\n\n### Speaker bio\n\nI am Jambunathan\, Di rector of Engineering at MadStreetDen. For the past 2 years\, we have been working on scaling our machine learning platform on AWS which supports th e multiple products at https://vue.ai.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123130Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/designing-microse rvices-around-your-data-design-TPZZ9NNYfm3JLvrZ25fadC BEGIN:VALARM ACTION:display DESCRIPTION:Designing microservices around your data design in Main audito rium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT BEGIN:VEVENT SUMMARY:Scaling hotstar.com for 25 million concurrent viewers DTSTART:20191116T105000Z DTEND:20191116T113000Z DTSTAMP:20260421T020940Z UID:session/EPxorRdfpJo4WdXkWyTka5@hasgeek.com SEQUENCE:3 CATEGORIES:Full talk (40 mins),SRE,SRE,Full talk (40 mins) CREATED:20190807T043627Z DESCRIPTION:About Hotstar \nWhy traditional autoscaling doesn't work for H otstar \nChallenges in scaling\nGamedays and load testing for millions \nH ow we prepare for peak traffic (50M)\nChaos engineering at high level\n\n# ## Speaker bio\n\nGaurav is a Cloud Architect at Hotstar\, India’s leadi ng OTT provider that holds a global concurrency record for live streaming to 25Mn+ viewers. At Hotstar\, he loves building cost-efficient infrastruc ture that can scale to millions in minutes. He is also passionate about ch aos engineering and cloud security.\n GEO:17.4457222;78.3488215586 LAST-MODIFIED:20240123T123117Z LOCATION:Main auditorium - T-Hub\nHyderabad\nIN ORGANIZER;CN=Rootconf:MAILTO:no-reply@hasgeek.com URL:https://hasgeek.com/rootconf/2019-hyderabad/schedule/scaling-hotstar-c om-for-25-million-concurrent-viewers-EPxorRdfpJo4WdXkWyTka5 BEGIN:VALARM ACTION:display DESCRIPTION:Scaling hotstar.com for 25 million concurrent viewers in Main auditorium in 5 minutes TRIGGER:-PT5M END:VALARM END:VEVENT END:VCALENDAR