Ernesta, the direct-to-consumer rug brand launched last year by three Peloton founders, is testing physical retail with a New York store.

Rug Startup Ernesta, Created By Peloton Founders, Opens First Showroom

Shares of exercise equipment company Peloton (NASDAQ:PTON) fell 8% in the morning session as major indices pulled back and yields soared after the Bureau of Labour Statistics reported CPI (Consumer Price Index - a gauge of the average price consumers pay for goods and services) for the month ...

Why Peloton (PTON) Shares Are Falling Today

March is Women&#x27;s History Month, and Peloton will be hosting an event in London to celebrate. Head below to learn more.

Peloton to celebrate Women's History Month with event in London

BACKORDERED - Order now and we will ship once more arrive! NOTE: Payment is charged when order ships. Extended backorders may require DVE Store to reach out to reconfirm payment information once ready to ship. The Decimator MD-HX HDMI and SDI Cross ...

Decimator MD-HX Miniature HDMI/SDI Cross Converter with Scaling & Frame Rate Conversion

Paperton

paperton

Authors:
(1) Halil Yigit Oksuz, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany;
(2) Fabio Molinari, Control Systems Group at Technische Universitat Berlin, Germany;
(3) Henning Sprekeler, Exzellenzcluster Science of Intelligence, Technische Universit¨at Berlin, Marchstr. 23, 10587, Berlin, Germany and Modelling Cognitive Processes Group at Technische Universit¨at Berlin, Germany;
(4) Jorg Raisch, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany. Table of Links Abstract and Introduction Problem Setup Federated fair over-the-air learning (FedAir) Algorithm Convergence Properties Numerical Example Conclusion and References VI. CONCLUSION In this paper, we have introduced the FedFAir algorithm which uses Over-the-Air Computation to carry out efficient decentralized learning while providing fairness and improved performance. We have shown that the FedFAir algorithm converges to an optimal solution of the minimax problem. Furthermore, we have also illustrated our theoretical findings with a numerical example. Future research will include the development of resilient federated learning algorithms when there are malicious agents in the system. REFERENCES [1] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics. PMLR, 2017, pp. 1273–1282. [2] V. Smith, C.-K. Chiang, M. Sanjabi, and A. S. Talwalkar, “Federated multi-task learning,” Advances in neural information processing systems, vol. 30, 2017. [3] K. Bonawitz, H. Eichner, W. Grieskamp, D. Huba, A. Ingerman, V. Ivanov, C. Kiddon, J. Konecny, S. Mazzocchi, B. McMahan et al., “Towards federated learning at scale: System design,” Proceedings of Machine Learning and Systems, vol. 1, pp. 374–388, 2019. [4] Q. Yang, Y. Liu, T. Chen, and Y. Tong, “Federated machine learning: Concept and applications,” ACM Transactions on Intelligent Systems and Technology (TIST), vol. 10, no. 2, pp. 1–19, 2019. [5] A. Nedic, “Distributed gradient methods for convex machine learning problems in networks: Distributed optimization,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 92–101, 2020. [6] V. P. Chellapandi, A. Upadhyay, A. Hashemi, and S. H. Zak, “On ˙ the convergence of decentralized federated learning under imperfect information sharing,” IEEE Control Systems Letters, 2023. [7] T. Omori and K. Kashima, “Combinatorial optimization approach to client scheduling for federated learning,” IEEE Control Systems Letters, 2023. [8] T. Sery, N. Shlezinger, K. Cohen, and Y. C. Eldar, “Over-the-air federated learning from heterogeneous data,” IEEE Transactions on Signal Processing, vol. 69, pp. 3796–3811, 2021. [9] T. Gafni, N. Shlezinger, K. Cohen, Y. C. Eldar, and H. V. Poor, “Federated learning: A signal processing perspective,” IEEE Signal Processing Magazine, vol. 39, no. 3, pp. 14–41, 2022. [10] M. Ye, X. Fang, B. Du, P. C. Yuen, and D. Tao, “Heterogeneous federated learning: State-of-the-art and research challenges,” ACM Computing Surveys, vol. 56, no. 3, pp. 1–44, 2023. [11] J. Koneˇcn `y, H. B. McMahan, F. X. Yu, P. Richt´arik, A. T. Suresh, and D. Bacon, “Federated learning: Strategies for improving communication efficiency,” arXiv preprint arXiv:1610.05492, 2016. [12] T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, “Federated learning: Challenges, methods, and future directions,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, 2020. [13] P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. N. Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummings et al., “Advances and open problems in federated learning,” Foundations and Trends® in Machine Learning, vol. 14, no. 1–2, pp. 1–210, 2021. [14] H. Y. Oksuz, F. Molinari, H. Sprekeler, and J. Raisch, “Federated learning in wireless networks via over-the-air computations,” in 2023 62nd IEEE Conference on Decision and Control (CDC), 2023, pp. 4379–4386. [15] F. Molinari, N. Agrawal, S. Sta´nczak, and J. Raisch, “Max-consensus over fading wireless channels,” IEEE Transactions on Control of Network Systems, vol. 8, no. 2, pp. 791–802, 2021. [16] M. Frey, I. Bjelakovi´c, and S. Sta´nczak, “Over-the-air computation in correlated channels,” IEEE Transactions on Signal Processing, vol. 69, pp. 5739–5755, 2021. [17] K. Yang, T. Jiang, Y. Shi, and Z. Ding, “Federated learning via overthe-air computation,” IEEE Transactions on Wireless Communications, vol. 19, no. 3, pp. 2022–2035, 2020. [18] F. Molinari and J. Raisch, “Exploiting wireless interference for distributively solving linear equations,” IFAC-PapersOnLine, vol. 53, no. 2, pp. 2999–3006, 2020. [19] D. Bertsekas, A. Nedic, and A. Ozdaglar, Convex analysis and optimization. Athena Scientific, 2003, vol. 1. [20] M. Mohri, G. Sivek, and A. T. Suresh, “Agnostic federated learning,” in International Conference on Machine Learning. PMLR, 2019, pp. 4615–4625. [21] Z. Hu, K. Shaloudegi, G. Zhang, and Y. Yu, “Federated learning meets multi-objective optimization,” IEEE Transactions on Network Science and Engineering, vol. 9, no. 4, pp. 2039–2051, 2022. [22] D. P. Bertsekas, “Necessary and sufficient conditions for a penalty method to be exact,” Mathematical programming, vol. 9, no. 1, pp. 87–99, 1975. [23] K. Srivastava, A. Nedi´c, and D. Stipanovi´c, “Distributed min-max optimization in networks,” in 2011 17th International Conference on Digital Signal Processing (DSP). IEEE, 2011, pp. 1–8. [24] R. Ahlswede, “Multi-way communication channels,” in Second International Symposium on Information Theory: Tsahkadsor, Armenia, USSR, Sept. 2-8, 1971, 1973. [25] A. Giridhar and P. Kumar, “Toward a theory of in-network computation in wireless sensor networks,” IEEE Communications Magazine, vol. 44, no. 4, pp. 98–107, apr 2006. [26] F. Molinari, N. Agrawal, S. Sta´nczak, and J. Raisch, “Over-the-air max-consensus in clustered networks adopting half-duplex communication technology,” IEEE Transactions on Control of Network Systems, 2022. [27] W. Rudin et al., Principles of mathematical analysis. McGraw-hill New York, 1976, vol. 3. [28] D. Tse and P. Viswanath, Fundamentals of Wireless Communication. Cambridge University Press, 2012. [29] A. F. Molisch, Wireless communications. John Wiley & Sons, 2012. [30] B. Sklar, “Rayleigh fading channels in mobile digital communication systems. i. characterization,” IEEE Communications magazine, vol. 35, no. 7, pp. 90–100, 1997. [31] R. G. Bartle and D. R. Sherbert, Introduction to real analysis. Wiley New York, 2000. [32] B. T. Polyak, Introduction to optimization. optimization software. Inc., Publications Division, New York, 1987. [33] Y. I. Alber, A. N. Iusem, and M. V. Solodov, “On the projected subgradient method for nonsmooth convex optimization in a hilbert space,” Mathematical Programming, vol. 81, pp. 23–35, 1998. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Halil Yigit Oksuz, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany; (2) Fabio Molinari, Control Systems Group at Technische Universitat Berlin, Germany; (3) Henning Sprekeler, Exzellenzcluster Science of Intelligence, Technische Universit¨at Berlin, Marchstr. 23, 10587, Berlin, Germany and Modelling Cognitive Processes Group at Technische Universit¨at Berlin, Germany; (4) Jorg Raisch, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany. Authors: Authors: (1) Halil Yigit Oksuz, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany; (2) Fabio Molinari, Control Systems Group at Technische Universitat Berlin, Germany; (3) Henning Sprekeler, Exzellenzcluster Science of Intelligence, Technische Universit¨at Berlin, Marchstr. 23, 10587, Berlin, Germany and Modelling Cognitive Processes Group at Technische Universit¨at Berlin, Germany; (4) Jorg Raisch, Control Systems Group at Technische Universitat Berlin, Germany and Exzellenzcluster Science of Intelligence, Technische Universitat Berlin, Marchstr. 23, 10587, Berlin, Germany. Table of Links Abstract and Introduction Abstract and Introduction Problem Setup Problem Setup Federated fair over-the-air learning (FedAir) Algorithm Federated fair over-the-air learning (FedAir) Algorithm Convergence Properties Convergence Properties Numerical Example Numerical Example Conclusion and References Conclusion and References VI. CONCLUSION In this paper, we have introduced the FedFAir algorithm which uses Over-the-Air Computation to carry out efficient decentralized learning while providing fairness and improved performance. We have shown that the FedFAir algorithm converges to an optimal solution of the minimax problem. Furthermore, we have also illustrated our theoretical findings with a numerical example. Future research will include the development of resilient federated learning algorithms when there are malicious agents in the system. REFERENCES [1] B. McMahan, E. Moore, D. Ramage, S. Hampson, and B. A. y Arcas, “Communication-efficient learning of deep networks from decentralized data,” in Artificial intelligence and statistics. PMLR, 2017, pp. 1273–1282. [2] V. Smith, C.-K. Chiang, M. Sanjabi, and A. S. Talwalkar, “Federated multi-task learning,” Advances in neural information processing systems, vol. 30, 2017. [3] K. Bonawitz, H. Eichner, W. Grieskamp, D. Huba, A. Ingerman, V. Ivanov, C. Kiddon, J. Konecny, S. Mazzocchi, B. McMahan et al., “Towards federated learning at scale: System design,” Proceedings of Machine Learning and Systems, vol. 1, pp. 374–388, 2019. [4] Q. Yang, Y. Liu, T. Chen, and Y. Tong, “Federated machine learning: Concept and applications,” ACM Transactions on Intelligent Systems and Technology (TIST), vol. 10, no. 2, pp. 1–19, 2019. [5] A. Nedic, “Distributed gradient methods for convex machine learning problems in networks: Distributed optimization,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 92–101, 2020. [6] V. P. Chellapandi, A. Upadhyay, A. Hashemi, and S. H. Zak, “On ˙ the convergence of decentralized federated learning under imperfect information sharing,” IEEE Control Systems Letters, 2023. [7] T. Omori and K. Kashima, “Combinatorial optimization approach to client scheduling for federated learning,” IEEE Control Systems Letters, 2023. [8] T. Sery, N. Shlezinger, K. Cohen, and Y. C. Eldar, “Over-the-air federated learning from heterogeneous data,” IEEE Transactions on Signal Processing, vol. 69, pp. 3796–3811, 2021. [9] T. Gafni, N. Shlezinger, K. Cohen, Y. C. Eldar, and H. V. Poor, “Federated learning: A signal processing perspective,” IEEE Signal Processing Magazine, vol. 39, no. 3, pp. 14–41, 2022. [10] M. Ye, X. Fang, B. Du, P. C. Yuen, and D. Tao, “Heterogeneous federated learning: State-of-the-art and research challenges,” ACM Computing Surveys, vol. 56, no. 3, pp. 1–44, 2023. [11] J. Koneˇcn `y, H. B. McMahan, F. X. Yu, P. Richt´arik, A. T. Suresh, and D. Bacon, “Federated learning: Strategies for improving communication efficiency,” arXiv preprint arXiv:1610.05492, 2016. [12] T. Li, A. K. Sahu, A. Talwalkar, and V. Smith, “Federated learning: Challenges, methods, and future directions,” IEEE Signal Processing Magazine, vol. 37, no. 3, pp. 50–60, 2020. [13] P. Kairouz, H. B. McMahan, B. Avent, A. Bellet, M. Bennis, A. N. Bhagoji, K. Bonawitz, Z. Charles, G. Cormode, R. Cummings et al., “Advances and open problems in federated learning,” Foundations and Trends® in Machine Learning, vol. 14, no. 1–2, pp. 1–210, 2021. [14] H. Y. Oksuz, F. Molinari, H. Sprekeler, and J. Raisch, “Federated learning in wireless networks via over-the-air computations,” in 2023 62nd IEEE Conference on Decision and Control (CDC), 2023, pp. 4379–4386. [15] F. Molinari, N. Agrawal, S. Sta´nczak, and J. Raisch, “Max-consensus over fading wireless channels,” IEEE Transactions on Control of Network Systems, vol. 8, no. 2, pp. 791–802, 2021. [16] M. Frey, I. Bjelakovi´c, and S. Sta´nczak, “Over-the-air computation in correlated channels,” IEEE Transactions on Signal Processing, vol. 69, pp. 5739–5755, 2021. [17] K. Yang, T. Jiang, Y. Shi, and Z. Ding, “Federated learning via overthe-air computation,” IEEE Transactions on Wireless Communications, vol. 19, no. 3, pp. 2022–2035, 2020. [18] F. Molinari and J. Raisch, “Exploiting wireless interference for distributively solving linear equations,” IFAC-PapersOnLine, vol. 53, no. 2, pp. 2999–3006, 2020. [19] D. Bertsekas, A. Nedic, and A. Ozdaglar, Convex analysis and optimization. Athena Scientific, 2003, vol. 1. [20] M. Mohri, G. Sivek, and A. T. Suresh, “Agnostic federated learning,” in International Conference on Machine Learning. PMLR, 2019, pp. 4615–4625. [21] Z. Hu, K. Shaloudegi, G. Zhang, and Y. Yu, “Federated learning meets multi-objective optimization,” IEEE Transactions on Network Science and Engineering, vol. 9, no. 4, pp. 2039–2051, 2022. [22] D. P. Bertsekas, “Necessary and sufficient conditions for a penalty method to be exact,” Mathematical programming, vol. 9, no. 1, pp. 87–99, 1975. [23] K. Srivastava, A. Nedi´c, and D. Stipanovi´c, “Distributed min-max optimization in networks,” in 2011 17th International Conference on Digital Signal Processing (DSP). IEEE, 2011, pp. 1–8. [24] R. Ahlswede, “Multi-way communication channels,” in Second International Symposium on Information Theory: Tsahkadsor, Armenia, USSR, Sept. 2-8, 1971, 1973. [25] A. Giridhar and P. Kumar, “Toward a theory of in-network computation in wireless sensor networks,” IEEE Communications Magazine, vol. 44, no. 4, pp. 98–107, apr 2006. [26] F. Molinari, N. Agrawal, S. Sta´nczak, and J. Raisch, “Over-the-air max-consensus in clustered networks adopting half-duplex communication technology,” IEEE Transactions on Control of Network Systems, 2022. [27] W. Rudin et al., Principles of mathematical analysis. McGraw-hill New York, 1976, vol. 3. [28] D. Tse and P. Viswanath, Fundamentals of Wireless Communication. Cambridge University Press, 2012. [29] A. F. Molisch, Wireless communications. John Wiley & Sons, 2012. [30] B. Sklar, “Rayleigh fading channels in mobile digital communication systems. i. characterization,” IEEE Communications magazine, vol. 35, no. 7, pp. 90–100, 1997. [31] R. G. Bartle and D. R. Sherbert, Introduction to real analysis. Wiley New York, 2000. [32] B. T. Polyak, Introduction to optimization. optimization software. Inc., Publications Division, New York, 1987. [33] Y. I. Alber, A. N. Iusem, and M. V. Solodov, “On the projected subgradient method for nonsmooth convex optimization in a hilbert space,” Mathematical Programming, vol. 81, pp. 23–35, 1998. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Boosting Fairness and Robustness in Over-the-Air Federated Learning: Conclusion and References

Why You Have to Love 1-to-1 Meetings to Succeed in Your Career

Controlling Automated Vehicles on Large Lane-free Roundabouts (Extended Version): References

Technological advancements like augmented and virtual reality have opened many doors in various sectors. For example, in the online shopping industry, these technologies allow customers to try something before buying it, bridging the gap between brick-and-mortar and online stores. The financial sector has also seen huge impacts from these technologies—or technology in general—and that’s what we’re here to talk about. Customer Service The increasing technologies in finance, particularly in customer service, have caused traditional banks to start adapting. The finance industry market is becoming increasingly competitive, and this is a result of . AR technology When it comes to loans, online lending companies are often seen as a preferred option compared to traditional lenders. For example, approval and fund disbursement are fast when you request a loan from a . payday loan organization That said, some banks now use AR so customers can easily locate nearby ATMs and branches and look out for deals and offers while strolling around a mall. In the UK and Australia, there is also such a thing as a “home finder” app that uses AR technology so potential home buyers can look at the information on houses for sale when they pass them. Using augmented reality and virtual reality in customer service increases user engagement and satisfaction, leading to a positive brand experience. Data Visualization As the finance industry continually grows, the data revolving around it is also growing. Both AR and VR technologies make data visualization easier, so companies can organize and visualize large chunks of data faster. This tool is important specifically for traders as it will help them make crucial decisions in managing their client’s wealth. For example, Salesforce utilizes Oculus Rift to generate a 3D environment to analyze data. Additionally, Fidelity Labs used Oculus Rift technology to create StockCity, a virtual 3D city of stock portfolios where investors can be fully immersed in the data. Virtual Reality Payments Virtual reality payments bring us a step closer to creating the metaverse in its truest sense, and it’s exciting to see this technology unfold. For example, MasterCard teamed up with Wearality, a VR glasses designer based in Orlando. They developed a prototype that enables customers to find an item within Mastercard’s “Priceless,” a VR golf experience, and purchase it without even leaving the virtual realm. Additionally, the winning team in the last Money20/20 Hackathon held in Las Vegas created a virtual reality mall where consumers can go into a virtual store and make an instant purchase. Financial Education and Training Virtual training provides value to both clients and employees in the financial industry. The rollout of new information is also made more manageable, especially in an industry where new technologies are constantly emerging. Avantica, a software engineering company founded in California, found that using AR or VR in training and remote conferences resulted in a compared to traditional methods. These technologies can help learners gain hands-on experiences and understand abstract concepts in a low-risk setting. 30% increase in information retention Virtual Branches In the presence of digital-only and mobile banks, it’s not far from reality that we’d get virtual bank branches as well. Virtual branches can provide a safe, time-saving, and cost-friendly alternative to visiting a bank’s physical branch. The goal is for these virtual branches to be able to offer the same services as a physical branch but solely in a virtual environment. This experience will be helpful to clients. However, banks will also see value in this in that they can reduce the overhead costs necessary for physical branch locations. Virtual Trading Virtual trading is a type of stock simulator where investors and traders can test-drive a trading platform and practice trading with no real-life consequences. Such systems aim to facilitate learning and familiarity with the tools available in a platform and understand how it works before deciding to do business with them. You’ll be trading virtual money in a real market environment. Real-life examples of virtual trading platforms are Webull, paperMoney, Paper Trading, and TradeStation Simulator. The last item on the list, TradeStation Simulator, is known among professional traders as it offers all of the features of its trading platform. You simply register online and start trading using virtual money. The platform will then provide you with a report of gains and losses. Security When augmented reality is integrated into biometric security and connected with virtual reality, financial companies can provide highly secure and significant protection against cyber threats. Making payments and ATM transactions can be made safer. For example, many mobile banking apps already feature user fingerprint authentication. With AR technology, voice recognition and iris identification are not far from reality. Customer Acquisition The use of AR and VR technologies in the finance sector helps reduce customer support costs, enhance user experience, and improve the efficiency of banking services. These technologies have proven beneficial in sales and marketing. When customers know how easy it is to do business with your company, they will continue to do business with you. Additionally, Millennials and newer generations grew up in a technological world, so they’re likely to be drawn to companies that value and implement such technological advancements. In fact, data from the shows that 73% of Millennials search for financial products on Apple, Google, and PayPal. Millennial Disruption Index Recruitment Financial institutions need top talent to ensure they’re providing the highest service quality to customers. To achieve this, banks can use VR technology to show recruits how technologically advanced and innovative they are. What The Future Looks Like In today’s highly technological era, technologies will continue to evolve and emerge. That said, technologies adjacent to reality, such as AR and VR, continue to become mainstream. These technologies are impacting the way we do business with financial institutions and how they operate and manage data. They also aid in customer acquisition and retention and enhance the user experience. The financial industry, particularly traditional banks, just needs to keep up and evolve to implement these systems, especially since believe that they will not need the services of a bank at all in the future. 33% of Millennials

Looking Into the Impact of AR and VR in the Finance Industry

What is Mobile Money? For the uninitiated, Agency Banking/Mobile Money has blown up across Africa and Asia in the last decade. Pioneered by telecom service providers like Vodafone and Safaricom with in Kenya in 2007, the global Mobile money market now processes over in transactions yearly. mPesa $490bn Mobile Money or Agency Banking is the generic term used to refer to the technology that facilitates the movement of money between individuals and businesses who may not have a Bank account domiciled at a traditional bank. This may be in terms of CICO (Cash-in-cash-out) transactions — where a customer might exchange cash (cash-in) for electronic value or vice versa. Mobile Money transactions can be completed in many ways. One way is with the use of a mobile device to move money to individuals or businesses who have a mobile phone number and are on the same telecom network — this is the most common form of Mobile Money transactions in East and West Africa. Mobile Money transactions can also take place via Mobile Money Agents — individuals who are licensed by a Mobile Money Service Provider to collect cash deposits and provide cash-out services on the company’s behalf. In most cases, these individuals acquire the licenses, devices and set up stalls using their own capital. In the United States, a cashier at your local Walmart or Walgreens can be considered a mini Mobile Money Agent since they can collect cash deposits and add the equivalent dollar amount to your CashApp app with the CashApp PaperMoney feature. Why Mobile Money? Around and over own a bank account. That means, more than half of young Nigerians do not own a bank account and are thus excluded from formal financial services. This presents a huge opportunity for Mobile money and Agency Banking platforms to bring the unbanked and underbanked into formal financial services. 40% of Nigerians aged 15 Since 2010, the number of Banks per capita (100,000 adults) in Nigeria has dropped from . For context, the number of Banks per capita in the US is 32. On the surface, the reduction in the number of physical banks might seem great since the internet and mobile phones are now widely available. 7 to 4 Many banks have found ways to lean out their operations by introducing internet and mobile banking services — a testament to the level of innovation within the Nigerian financial technology space. In Nigeria, thanks to , inter-bank transfers are completed nearly instantly. I’ve previously recorded a time of 2 seconds from the time I hit send on my Kuda app to receiving it in my UBA Bank account. This is not the same experience in other parts of Africa — in the Ivory coast, for example, Bank to Bank transfers can take up to 3 days to settle. NIBSS The CBN (Nigeria’s Financial regulator) has also initiated several policies to encourage individuals to go cashless. Despite these advancements in the financial services space in Nigeria, when we look a little closer look into the situation of things, the CBN tells us that the amount of cash in circulation in October 2021 ($7.3 billion) compared to the same period the month before — proving, once again, that cash is still king and a major portion of the population still like to transact in cash. The love for cash based transactions is also prevalent in Vietnam, where over . While we believe that a major portion of the Nigerian population is only comfortable with cash, we believe that with the widespread adoption of mobile phones, better technology can be deployed to bring more people into formal financial services by meeting them where they are i.e without the need for formal banks; cue — and . rose by 4.6% to 3 trillion naira 80% of transactions are cash based Mobile Money Agency Banking In Nigeria, the volume of POS transactions processed in 2021 alone was over and the sector continues to grow year on year. Over the next decade, and with the continuous innovation and investments surrounding mobile money, we should see more and more people being able to access basic financial services bearing in mind that they may never own a bank account at a traditional banking institution. The big bet is on Mobile Money services to accelerate this future. $8bn With that said, Let’s take a look into how a typical Mobile Money or Agency banking transaction works. How does Mobile Money/Agency Banking Work? Mobile Money transactions can be completed in a number of ways: Sometimes, they are completed using a mobile phone over a mobile network, Other times, they’re completed using a card issued by a Digital bank or a traditional financial institution and a Mobile POS Terminal. In many cases, they’re completed with an Agent receiving cash from a customer and making a transfer from their mobile phone to the wallet/bank account of the recipient. Regardless of what method is used, transactions are usually completed within 5 mins — a testament to the innovation that has occurred within the space. When a transaction is completed using a mobile phone alone, your phone number serves as your account number, and your wallet/account balance is stored on your mobile device — either via the app or via a menu option on your mobile phone added by your telecom provider. Getting on the mPesa network is as simple as buying a sim from Safaricom. With your sim in your mobile phone, you’re ready to participate in the mobile money revolution. You can begin transacting with other users and Businesses on the mPesa network. With a mobile phone as cheap as $5, you can pay bills, send money to other mPesa users, take a loan, and more. In rural areas without Banks or where the closest bank is too far away from residents, mobile money Agents (mPesa Agents in this case) facilitate cash deposits and withdrawals — think of a human ATM. Here’s a walkthrough of a transaction at a typical Mobile Money Agent. Picture a Shop owner running a small business that sells groceries. At the end of the day, this Customer takes all the cash they’ve received from their daily sales and walks to the nearest Mobile Money Agent close to them — usually, these Agents are much closer to the customer’s location than a physical Bank. The Shop Owner takes the cash they’ve collected to a Mobile Money Agent who collects the cash and deposits the same value in the customer’s wallet less the transaction charges. The Shop owner is happy — they have value in their wallet which they can use to purchase more goods or send to their family member’s wallet on the same telecom network in a neighboring city. The Mobile Money Agent is also happy since they’ve made a commission for processing that transaction. In most of Sub-saharan Africa, this is how most Mobile Money transactions go. In Nigeria, it’s a little different. The most common trend in the last few years has been a network of Agents spearheaded by giant payment companies or MMOs (Mobile Money Operators) such as , , Paga, , and more. Typically, these MMOs deploy Mobile POS terminals that Agents can use to start their own Agent Banking business. Opay Moniepoint Xpressteller With increasing innovation in this space, MMOs have started to deploy android based devices that can process card transactions and print receipts. These devices can also be used to pay bills, purchase airtime, perform inter-bank transfers, and more — meaning that the customers of these Agents can perform all the services they’ll typically do at a traditional bank at a Mobile Money Agent. Ivory Coast has the largest population of adults with a Mobile Money account in all of West Africa — at 75%. In stark contrast to that, only 20% of the adult population owns a Bank account. What’s the difference between a Mobile Money Agent and a POS Agent? Traditionally, in most of West Africa and East Africa, Mobile Money has been a medium for carrying out financial transactions where the store of value is in a on a mobile phone device linked to a customer’s mobile phone number. Transactions are facilitated on the network of the Telecom operator’s the user is subscribed to — popular options include MTN in Benin Republic, Oranje in Guinea and Safaricom in Kenya. wallet In Benin Republic for example, where MTN’s MoMo is dominant, users move money around using their MoMo wallet which can be done using an App on their Mobile device or using from their mobile phone. MTN processed with only a quarter of the entire population of Benin Republic as active users! USSD shortcodes $709m in transactions in July of 2021 To paint a better picture, over 11m Mobile Money accounts were registered in 2021 compared to 2m Bank accounts proving the mass adoption of mobile money. Until recently when the Central Bank of Nigeria (Payment Service Bank) licenses to Telecom Operators in Nigeria, MTN was unable to replicate their MoMo model in Nigeria. Since obtaining that license, MTN has registered MoMo (short for Mobile Money) as a payment bank in Nigeria and are beginning to scale their Agent network across the country. With Telcom companies now playing in the financial service space, it means more competition for the likes of TeamApt’s MoniePoint, Opay and Palmpay amongst others — who are Mobile Money Operators but not telecom service providers. For MTN, the numbers are already there since they have the . awarded PSB largest subscriber base in Nigeria At , we think about Mobile Money a little differently. Because of the variations in how Mobile Money operates around the world, we like to think about Mobile Money in the sense of mobility — we see Mobile Money as any digital system that allows the banked, underbanked and unbanked to participate in financial services by removing/reducing the barriers to participation associated with traditional financial institutions; this may be via the use of Mobile POS terminals, Mobile Phones that double as POS terminals, QR codes linked to QR Cards/Mobile Devices (like in the case of ) or scan to pay apps. Moni Wave Mobile Money Looking at Mobile Money and thus Mobile Money Agents in this way, allows us to build a suite of products to cater to Mobile Money Agents and their lifestyle. How much does a Typical Mobile Money Agent Make? Revenue earned by Mobile Money Agents can vary swiftly from place to place. While some areas are high volume and high traffic — consider a market for example, other areas may be high volume but low traffic. In a busy market, a Mobile Money Agent is likely to process over 100 transactions within a day with high volumes. On the other hand, in a residential/commercial area under development, a mobile money Agent may not see a lot of foot traffic but may process a high volume of transactions as vendors and contractors move money around. In Nigeria, we’ve seen that Mobile Money Agents are able to do an average turnover of N2.4m (~ $5000) a month. Assuming normal charges on transactions, a Mobile Money Agent may earn around N3,500 (~ $8) per day which works out at roughly N100k ($215) a month. With more float, and/or with a better location, Mobile Money Agents can 10x their earnings. In Benin Republic, a Mobile Money Agent can earn around CFA70k depending on the number of transactions they complete and the Agent’s location. On average, in the Nigerian market, the cost of acquiring a Mobile POS Terminal is around N20k — N25k(~$54). Besides the cost of acquiring a terminal, Mobile Money Agents will also need to get a location for their business and may need to pay rent for the space. In our experience, many Mobile Money Agents use existing Mom and Pop outlets — which may or may not belong to them —  as their primary place of Business. Starting a Mobile Money Business in Africa is not all rosy despite all the innovation occurring in the space. For one, individuals interested in starting their own Mobile Money/Agency Banking outlet need working capital to get started — this working capital may not be readily available, in which case they’d need to take a loan from friends and family since a traditional Bank will not offer them that level of credit without collateral. Besides the issue with working capital, in Benin Republic, setting up shop as a MoMo Agent can take anywhere from 2–4 months due to the paperwork and bureaucracy involved. There’s still room for more innovation within this space. Mobile Money transactions are still expensive in many parts of Africa and is one of the companies tackling this problem with up to 70% lower fees on Mobile Money transactions. wave Wrapping Up To wrap up, we’ll continue to see innovation in the Agency Banking and mobile money space across Africa as technology companies find ways to help the banked, unbanked, and underbanked access financial services through Agency Banking networks or as we like to call them at Moni — human ATMs. Recently, Moni was admitted into Y Combinator’s . Our approach to lending is different. We’re aware of the major problem Mobile Money Agents face across Africa - float or working capital for their businesses. At the same time, we are also aware that traditional institutions may be unable to provide credit to Mobile Money Agents because they’re unable to satisfy the Bank’s credit requirements. YC W22 batch At , we provide credit/float/working capital to Mobile Money Agents using a CoFi (Community Financing) model based on their social reputation. Our model allows Agents to facilitate more transactions thus ensuring that they earn more and prosper! Moni Also Published here

Mobile Money: The Key to Financial Inclusion in Emerging Markets

Courtesy The first step in optimizing a query is looking at it’s execution plan. An execution plan is a visual representation of the operations performed by the engine in order to return the data required by your query. database The execution plan for a query is your view into the Server query optimizer and query engine. It will reveal which , object like: SQL objects a query uses tables indexes etc. type of uses: in which order how they were accessed (seek or scan) types of joins data size filtering sorting aggregations calculated columns derivation foreign keys access etc. within the database, and how it uses them. There are 3 types of execution plans: Estimated Actual Cached Estimated execution plan: To get a query’s estimated execution plan: Switch to the database in concern highlight the query in concern click Query click Display Estimated Execution Plan The shortcut for this is Ctrl+ L. When you get this plan, some sections of the query might show a very high percentage as cost. Remember that this is just an estimate, and many a times what we see as the most expensive section comes up as the most cheapest or cheap. Btw, each symbol in the execution plan is an operator, and each operator has additional information within its properties that define the work performed by that operator. Actual execution plan: To get a query’s actual execution plan: Switch to the database in concern highlight the query in concern click Query click Include Actual Execution Plan The shortcut for this is Ctrl+ M. One of the queries gave this actual execution plan. Look at the actual vs estimated numbers, looks like the estimate is good enough. One of other queries gave this execution plan: Looks like the estimates and actuals are far different. This proves that . Btw, we read execution plans from left to right and top to bottom. actual execution plan are always good to tune compared to estimates The estimated percentage costs at the front level are . So to tune the queries, it looks like we need to go through each of the blocks one-by-one. This looks like a lot of pain. This is where the 3rd type of execution plans comes into the picture — the . still estimates cached execution plan Cached execution plan: There is a to find out which queries we actually need to tune. Just hit and grab . free cool utility http://www.brentozar.com/blitzcache sp_BlitzCache Just execute this . This query will give the top 10 queries with the most execution time. In a column it also shows the possible problems like missing indexes, downlevel cardinality estimates and implicit conversions in the column. In the column, it gives the actual execution plan map link which can be clicked upon. EXEC sp_BlitzCache @top=10, @sort_order='duration' Warnings Query Plan Query Tuning: The single largest use for execution plans is query tuning. It’s the most common tool for understanding what you want to tune. It’s invaluable for identifying and fixing poorly performing code. I can’t run through every potential cause of poor query performance here, or even close, but we can look briefly at a few of the common culprits and the warning signs you’ll find in the execution plan. Poorly Designed Queries Like with any language, there are common code smells within T-SQL. I’ll list out a few of them as well as what you might see in the execution plan. Search Conditions with Functions Placing a function on a column in the , or clause (e.g. can lead to very poor performance because it renders the predicate non-SARGable, which means that SQL Server cannot use it in an index seek operation. Instead, it will be forced to read the entire table or index. In the execution plan, look out for a Scan operator against a table or index where you expected a Seek operation, i.e. a single or small set of lookups within an index. ON HAVING WHERE WHERE SomeFunction(Column) = @Value ) Nesting Views Having a view that calls to other views or has operations to other views can lead to very poor performance. The query optimizer goes through a process called simplification, where it attempts to eliminate tables that are not needed to fulfil the query, helping to make it faster. For simple queries involving views or functions, the optimizer will simply access the underlying base tables, and perform table elimination in the usual fashion. However, use of nested views and functions will lead quickly to a very complex query within SQL Server, even if the query looks simple in the T-SQL code. The problem is that beyond a certain point, SQL Server can’t perform the usual simplification process, and you’ll end up with large complex plans with lots of operators referencing the tables more than once, and so causing unnecessary work. JOIN Incorrect Data Type Use Of course, to do this it must apply the conversion function to the column, in the , or clause and we are back in the situation of having a non-SARGable predicate, leading to scans where you should see seeks. You may also see a warning indicating an implicit conversion is occurring. ON HAVING WHERE Row-by-row Processing Row-by-row approach is expressed via cursors and sometimes as loops that often lead to extremely poor performance. In such cases, rather than a single execution plan, you’ll see one plan for each pass through of the loop, and SQL Server essentially generates the same plan over and over, to process each row. WHILE Common Warning Signs in the Execution Plan Warnings These are red crosses or yellow exclamation marks superimposed on an operator. There are sometimes false positives, but most of the time, these indicate an issue to investigate. Costly Operations — the operation costs are estimated values, not actual measures, but they are the one number provided to us so we’re going to use them. The most costly operation is frequently an indication of where to start troubleshooting Fat Pipes Within an execution plan, the arrows connecting one operator to the next are called pipes, and represent data flow. Fat pipes indicate lots of data being processed. Of course, this is inevitable if the query simply has to process a large amount of data, but it can often be an indication of a problem. Look out for transitions from very fat pipes to very thin ones, indicating late filtering and that possibly a different index is needed. Transitions from very thin pipes to fat ones indicates multiplication of data, i.e. some operation within the T-SQL is causing more and more data to be generated. Extra Operators Over time you’ll start to recognize operators and understand quickly both what each one is doing and why. Any time that you see an operator and you don’t understand what it is, or, you don’t understand why it’s there, then that’s an indicator of a potential problem. Scans We say this all the time, scans are not necessarily a bad thing. They’re just an indication of the scan of an index or table. If your query is , with no clause then a scan is the best way to retrieve that data. In other circumstances, a scan can be an indication of a problem such as implicit data conversions, or functions on a column in the clause. SELECT * FROM TableName WHERE WHERE Let’s take a look at a few of the warning signs that can appear, and what we can do in response. DECLARE @LocationName AS NVARCHAR(50); SET @LocationName = 'Paint'; SELECT  p.Name AS ProductName ,        pi.Shelf ,        l.Name AS LocationNameFROM    Production.Product AS p        JOIN Production.ProductInventory AS pi ON pi.ProductID = p.ProductID        JOIN Production.Location AS l ON l.LocationID = pi.LocationIDWHERE   LTRIM(RTRIM(l.Name)) = @LocationName;GO And the problems are as follows: The INDEX SCAN on the column – there’s an index on that column a SEEK is expected - culrit: , in the clause Location.Name LTRIM(RTRIM(l.Name)) WHERE The FAT PIPE coming out of the Clustered Index Scan on the ProductInventory table — in this example, it’s processing over 1000 rows at this stage, and eventually only returning 9 rows. — culprit: missing index on the column. The HASH MATCH join — for the number of rows returned I’m expecting a NESTED LOOPS join — culprit: bad or missing indexes. Poor Indexing At the top of some execution plans, you may even see an explicit suggestion, indicating that the optimizer recognized that if it had a certain index, it might be able to come up with a better plan. Note these, but do not assume they’re accurate. Test them before applying them. Missing Index Nesting Views A view is just a query that the optimizer resolves just like any other query. However, if you decide to JOIN a view to a view, or nest views inside of each other, it causes problems for SQL Server. Firstly, as the number of referenced objects increases, due to deep nesting, the optimizer, which only has so much time to try to optimize the query, will give up on simplification and just build a query for the objects defined. For each object, the optimizer must estimate the number of rows returned and, based on that, the most efficient operation to use to return those rows. However, once we go beyond levels deep on obfuscation, the optimizer stops assigning costs and just assumes one row returned. All of this can result in very poor choices of plan and a lot of avoidable work. three To illustrate how nesting views can hurt the server, consider this: SELECT  pa.PersonName,        pa.City,        cea.EmailAddress,        cs.DueDateFROM    dbo.PersonAddress AS paJOIN    dbo.CustomerSales AS cs        ON cs.CustomerPersonID = pa.PersonIDLEFT JOIN dbo.ContactEmailAddress AS cea        ON cea.ContactPersonID = pa.PersonIDWHERE   pa.City = 'Redmond'; All the 3 referenced objects are views. And the execution plan will be so complex like this: The simplication of the above query can be done by actually accessing the underlying tables like this: SELECT  p.LastName + ', ' + p.FirstName AS PersonName ,        a.City ,        ea.EmailAddress ,        soh.DueDateFROM    Person.Person AS p        JOIN Person.EmailAddress AS ea              ON ea.BusinessEntityID = p.BusinessEntityID        JOIN Person.BusinessEntityAddress AS bea              ON bea.BusinessEntityID = p.BusinessEntityID        JOIN Person.Address AS a              ON a.AddressID = bea.AddressID        LEFT JOIN Person.BusinessEntityContact AS bec              ON bec.PersonID = p.BusinessEntityID        JOIN Sales.Customer AS c              ON c.PersonID = p.BusinessEntityID        JOIN Sales.SalesOrderHeader AS soh              ON soh.CustomerID = c.CustomerIDWHERE   a.City = 'Redmond'; The execution plan is much simplified now: Lack of Database Constraints Consider this query: SELECT  soh.OrderDate ,        soh.ShipDate ,        sod.OrderQty ,        sod.UnitPrice ,        p.Name AS ProductNameFROM    Sales.SalesOrderHeader AS soh        JOIN Sales.SalesOrderDetail AS sod              ON sod.SalesOrderID = soh.SalesOrderID        JOIN Production.Product AS p               ON p.ProductID = sod.ProductIDWHERE   p.Name = 'Water Bottle - 30 oz.'        AND sod.UnitPrice < $0.0; With no constraints in place, the execution plan for this query will look as shown: This query returns zero rows, since there are no products with a price less than $0, and yet SQL Server still performs an index scan, two index seeks and a couple of nested loop joins. The reason is that it has no way to know they query will return zero rows. However, what if we have a constraint in place on the SalesOrderDetail table that requires that the UnitPrice for any row be greater than zero (this constraint already exists in the AdventureWorks database). ALTER TABLE Sales.SalesOrderDetail  WITH CHECK  ADD  CONSTRAINT CK_SalesOrderDetail_UnitPrice CHECK  ((UnitPrice>=(0.00))); If we rerun the above code with this constraint in place, then the execution plan is interesting: Missing or out-of-date Database Statistics The statistics can be turned off or on by the following command: SET AUTO_UPDATE_STATISTICS OFF | ON; It is recommended to not set them off to avoid stale statistics. Maybe it’s not the Database? It’s also possible that the database is not the culprit, the application can cause performance problems. So, profile the application code as well. Sources: Execution plan benefits Types of execution plans Photos

Execution plan in SQL Server

How Digital Platforms Undo Paperwork In Real Estate

The function — the poor man’s debugger — is every JavaScript developer’s best friend. We use it to verify that a certain piece of code was executed or to check the state of the application at a given point in time. We may also use to send warning messages or to explain what happened when things have gone wrong. Logging makes it easy to debug your app during local development. console.log console.warn console.error But what about debugging your Node.js app while it’s running in a hosted cloud environment? The logs are kept on the server, to which you may or may not have access. How do you view your logs then? Most companies use application performance monitoring tools and observability tools for better visibility into their hosted apps. For example, you might send your logs to a log aggregator like or where logs can be viewed and queried. Datadog, Sumo Logic, Papertrail In this article, we’ll look at how we can configure an app that is hosted on to send its system logs to Papertrail by using . By the end, you’ll have your app up and running — and logging — in no time. Render Render Log Streams Creating Our Node.js App and Hosting It with Render Render is a cloud hosting platform made for developers by developers. With Render, you can easily host your static sites, web services, cron jobs, and more. We’ll start with a simple and Express app for our demo. You can . To follow along on your machine, fork the repo so that you have a copy running locally. You can install the project’s dependencies by running , and you can start the app by running . Node.js find the GitHub repo here yarn install yarn start Easy enough! Now it’s time to get our app running on Render. If you don’t have a Render account yet, create one now. It’s free! Once you’re logged in, click the “New” button and then choose the “Web Service” option from the menu. This will take you to the next page where you’ll select the GitHub repo you’d like to connect. If you haven’t connected your GitHub account yet, you can do so here. And if you have connected your GitHub account but haven’t given Render access to your specific repo yet, you can click the “Configure account” button. This will take you to GitHub, where you can grant access to all your repos or just a selection of them. Back on Render, after connecting to your repo, you’ll be taken to a configuration page. Give your app a name (I chose the same name as my repo, but it can be anything), and then provide the correct build command ( , which is a shortcut for ) and start command ( ). Choose your instance type (free tier), and then click the “Create Web Service” button at the bottom of the page to complete your configuration setup. yarn yarn install yarn start With that, Render will deploy your app. You did it! You now have an app hosted on Render’s platform. Creating Our Papertrail Account Let’s now create a Papertrail account. Papertrail is a log aggregator tool that helps make log management easy. You can create an account for free — no credit card required. Once you’ve created your account, click on the “Add your first system” button to get started. This will take you to the next page which provides you with your syslog endpoint at the top of the screen. There are also instructions for running an install script, but in our case, we don’t actually need to install anything! So just copy that syslog endpoint, and we’ll paste it in just a bit. Connecting Our Render App to Papertrail We now have an app hosted on Render, and we have a Papertrail account for logging. Let’s connect the two! Back in the Render dashboard, click on your avatar in the global navigation, then choose “Account Settings” from the drop-down menu. Then in the secondary side navigation, click on the “Log Streams” tab. Once on that page, you can click the “Add Log Stream” button, which will open a modal. Paste your syslog endpoint from Papertrail into the “Log Endpoint” input field, and then click “Add Log Stream” to save your changes. You should now see your Log Stream endpoint shown in Render’s dashboard. Great! We’ve connected Render to Papertrail. What’s neat is that we’ve set up this connection for our entire Render account, so we don’t have to configure it for each individual app hosted on Render. Adding Logs to Our Render App Now that we have our logging configured, let’s take it for a test run. In our GitHub repo’s code, we have the following in our file: app.js app.get('/', (req, res) => {
    console.log('Log - home page');
    console.info('Info - home page');
    console.warn('Warn - home page');
    console.error('Error - home page');
    console.debug('Debug - home page');

    return res.sendFile('index.html', { root: 'public' });
}); When a request is made to the root URL of our app, we do a bit of logging and then send the file to the client. The user doesn’t see any of the logs since these are server-side rather than client-side logs. Instead, the logs are kept on our server, which, again, is hosted on Render. index.html To generate the logs, open your demo app in your browser. This will trigger a request for the home page. If you’re following along, your app URL will be different from mine, but . my app is hosted here Viewing Logs in Papertrail Let’s go find those logs in Papertrail. After all, they were logged to our server, but our server is hosted on Render. In your Papertrail dashboard, you should see at least two systems: one for Render itself, which was used to test the account connection, and one for your Render app (“render-log-stream-demo” in my case). Click on the system for your Render app, and you’ll see a page where all the logs are shown and tailed, with the latest logs appearing at the bottom of the screen. You can see that we have logs for many events, not just the data that we chose to log from our file. These are the syslogs, so you also get helpful log data from when Render was installing dependencies and deploying your app! app.js At the bottom of the page we can enter search terms to query our logs. We don’t have many logs here yet, but where you’re running a web service that gets millions of requests per day, these log outputs can get very large very quickly. Best Practices for Logging This leads us to some good questions: Now that we have logging set up, what exactly should we be logging? And how should we be formatting our logs so that they’re easy to query when we need to find them? What you’re logging and why you’re logging something will vary by situation. You may be adding logs after a customer issue is reported that you’re unable to reproduce locally. By adding logs to your app, you can get better visibility into what’s happening live in production. This is a reactive form of logging in which you’re adding new logs to certain files and functions after you realize you need them. As a more proactive form of logging, there may be important business transactions that you want to log all the time, such as account creation or order placement. This will give you greater peace of mind that events are being processed as expected throughout the day. It will also help you see the volume of events generated in any given interval. And, when things do go wrong, you’ll be able to pinpoint when your log output changed. How you format your logs is up to you, but you should be consistent in your log structure. In our example, we just logged text strings, but it would be even better to log our data in JSON format. With JSON, we can include key-value pairs for all of our messages. For each message, we might choose to include data for the user ID, the timestamp, the actual message text, and more. The beauty of JSON is that it makes querying your logs much easier, especially when viewing them in a log aggregator tool that contains thousands or millions of other messages. Conclusion There you have it — how to host your app on Render and configure logging with Render Log Streams and Papertrail. Both platforms only took minutes to set up, and now we can manage our logs with ease. Keep in mind that Render Log Streams let you send your logs to any of several different log aggregators, giving you lots of options. For example, Render logs can be . You just need to create a Cloud Syslog Source in your Sumo Logic account. Or, you can as well. sent to Sumo Logic send your logs to Datadog With that, it’s time for me to off. log Thanks for reading, Happy coding, and happy logging!

Logging to Papertrail with Render Log Streams

Solarwinds scrutiny is well-earned these days. Despite being a security company by nature and owning two logging companies - Loggly & Papertrail - I can’t help but notice they aren't taking care of their own business. For example, if you follow the Loggly or Papertrail default instructions, your syslog will be configured to send everything over the internet as plain-text. Of course, they support TLS and have advanced/optional instructions for configuring that but after decades of experience in software and devops we know the danger of defaults against a motivated attacker. At Wrble we offer unencrypted channels if you must but default and strongly recommend configuring TLS and it is in there by default - saving you time while strengthening security posture. This is critical and many customers who otherwise require vendors to secure SOC2 accreditation or fill out endless security questionnaires are relying on companies that don't take it seriously for their own products' day-to-day operation. How to conduct a self-audit of your logs First off, do you know what you’re using for logs right now? If you aren't sure what tool you're using or someone other than you (who has maybe moved on to another project, department, or company!) originally configured Syslog on all your servers. Considering what’s at stake, it’s definitely worth a look. We recommend that everyone who sends logs, especially on older systems, review how they're sending and make sure it’s secure as can be. Here are some tips on how to check: Whether you're configuring a Syslog daemon, using a custom library, or a custom daemon from a vendor - there are two main ways to get logs off a system: Syslog and HTTP. Now since there are secure and insecure versions of each, we'll run through them and talk about how you can make sure that they're sending over TLS: Syslog TCP/UDP Verify Configuration Unencrypted Syslog traffic generally flows over port 514 on either TCP or UDP. If it's encrypted, you should see one or both of the lines in your conf file: ActionSendStreamDriver gtls 6514 This config file will probably be a ".conf" file in the directory, but could also be or If you're running inside Docker, it might be configured inside the image or on the host system so it's good to check both places. /etc/rsyslog.d/ /etc/rsyslog.conf /etc/rsyslog.conf. Check Traffic Checking configuration is great, but can be complicated and hard to tell which config file is authoritative. Here we'll go over how to check for both unencrypted and encrypted traffic originating from Syslog so you can make sure it's configured correctly: First, let's check for unencrypted traffic on 514. We'll be using a tool called that prints out traffic going over a specific port. You can install it via yum: or apt-get: apt-get install (you may need to update your package cache with apt-get update). tcpdump yum install tcpdump tcpdump OK, now that we have that installed here's how to print all unencrypted Syslog traffic on the TCP port: tcpdump -i any -nn -v port 514 Once that's running it will print out any packets coming or going over TCP 514, trigger something in your application that causes a log line to be sent and see if it shows up there. If it's empty you should be all good, but there are a few more things to check, here's how to look at 514 for UDP traffic: tcpdump -i any -nn -v udp port 514 Trigger log traffic again and see if anything shows up. If nothing shows up you're in good shape. Now we've verified the negative lack of unencrypted traffic but let's look to see if we can prove it is transiting out of the system securely by watching secure traffic: TCP: tcpdump -i any -nn -v port 6514 UDP: tcpdump -i any -nn -v udp port 6514 Hopefully, at this point, you'll see some lines come across as logs are triggered. It will be all garbled as you're seeing the TLS packets themselves but their existence and garbled nature are a great sign... congratulations! HTTP If you aren't using Syslog, your logs are probably going over HTTP(S) even if it's wrapped up in several layers of vendor libraries or daemons. It's really hard to list all the configurations to check, but we can still audit the traffic with tcpdump. HTTP(S) traffic can be pretty noisy on many hosts, when we're checking this we'll limit to traffic originating from your host as that should be much less noisy. NOTE: change "10.0.0.10" to your host's IP address in all the tcpdump commands below. tcpdump -i any src host 10.0.0.10 and port 80 Trigger log sending and see if any traffic shows up, if it's nothing or just other outgoing HTTP connections from your box you're in good shape. Let's also check HTTPS: tcpdump -i any src host 10.0.0.10 and port 443 Traffic here can be hard to discern from other outbound connections but look for your logging vendor's hostname as a target of one of the packets, if they're showing up here you're in good shape! Insecure HTTPS If a vendor is controlling the HTTPS send via a library or daemon it could still be possible for them to "break" the encryption by bypassing TLS security checks when configuring their HTTPS library. This can be especially common on old or embedded system where TLS certs are hard to configure. The most common bypasses are checking the hostname and certificate validity, here are instructions to verify both of those. This is a little harder to verify but worth your time if you're sending critical data to your logging provider. First, we'll need to set up an "HTTP echo server" on a separate machine like this: docker run -d -p 443:443 danthegoodman/https-echoserver Note that it's listening on 443 and will terminate TLS but it is not a secure server, the TLS certificate is self-signed. This is what we want in this rare case! Now configure the box that you're application is running on by editing /etc/hosts like this: 10.0.0.1 ingest.loggingprovider.com but replace with the IP of the machine running your echo server and with the hostname, your logging provider uses to receive logs. 10.0.0.1 ingest.loggingprovider.com If you now trigger log lines to be sent and they show up in the echo server you'll know that they've disabled hostname and/or certificate checking as we don't have a valid certificate for their domain! Report this to your provider as that’s one big glaring security hole. Whew, that was a lot of work! We know security is hard and it's a confusing world out there (Syslog was first released in the '80s when security was an afterthought at best!). But we hope these tips were able to make it a little less painful and a bit quicker. Reach out directly if I can help walk you through the securing of your systems or to learn how Wrble can reduce your logging spend while improving security.

Solarwinds' Loggly and Papertrail Default to Sending Unencrypted Logs

What can this post teach me? This post will teach you how to connect to and then automate card creations on Trello. Paperform Zapier Backstory So a while back I discovered this awesome Online Form company from Australia — . Paperform I was curious on understanding how Zapier works together with Paperform and as the best way of learning is to give yourself a project I decided to make one. Although I am retired from the software business I still take the time to hang out with cool freelance tech geniuses that can feed the inner geek in me. A few times a week these freelancers meet and work out of various coffeeshops. The locations are fixed but the weekdays can vary. I thought it would be a good idea to make a an easy app that could help our little band of brothers to quickly communicate where they would be on a weekly basis but not have to login to a calendar or any other service. The app should: ① allow any freelancer, in our group to tell the rest of the team where he/she will be working on any particular weekday ② allow anyone in the group to see where everybody will be on a weekly basis ③ not require users to login anywhere ④ make input super quick and intuitive ⑤ make results intuitive and easy to read Components of the Solution a Trello board to see the weekly schedule of everybody a zap to integrate Paperform with Trello a form hosted by Paperform to get the input Example of the input form I made using the Paperform admin panel Managing Output & Input The Trello board to capture the output. Basically the dashboard that shows where everybody will be Monday to Friday The front-end of the form used to grab the input from users, (allowing freelancers to pick where they’ll go) The Finished App Integrate all components into one HTML page so that users only have to go to one place to see and do everything. Before we begin I really recommend . It is a small business from Australia run by some GREAT people that listen to their customers and make changes quite fast. Their platform also has more logic and functionality than some of the other form builders. . Paperform See this article I wrote a while back on how to make a Ghost contact form using Paperform You will also work with and Zapier Trello Both Zapier and Trello have plans that you can use until your workflow requires more functionality and thus have to upgrade to a paid plan. free This story is brought to you by Hacker Noon’s weekly sponsor, . Create beautiful forms. For Hacker Noon readers, . Paperform 10% off all plans forever The process when working with Zapier is: ① allow access to the service you want Zapier to manipulate ② use Zapier to browse inside the service you want e.g. or Trello and grab the data you want to manipulate ③ tell Zapier what to do with the data you have grabbed Paperform Don’t worry, all these platforms we will be working with are pretty intuitive. Ok, let’s do it! — Here are the steps: Create accounts at , at and at (1) Paperform Zapier Trello You want to start by making a Zap on Zapier (2) A Zap is Zapier language for an integration Next you have to decide what app you want to use to trigger an action. Let’s choose Paperform (3a) A Trigger app is Zapier language for the app that will trigger a later action. In our case it is the submission of a form at Paperform. You have to connect Zapier to your Paperform account. Once you click on ‘Connect a New Account’ you should see a popup box that asks for your paperform login. (3b) If successful, you should see below: (3c) Now it is time to pick the form we want to access in Paperform (4) Ok, we should be done on the Paperform side. Let’s connect Trello to Zapier. This is done in the same way as in step 3b above, when we connected Paperform to Zapier (5) Add the action to use Trello as the next step in Zapier (6) We need to choose a Trello action that we want Zapier to execute for us. Scroll down in the list and pick ‘Find list’ and then continue. (7) An action is Zapier language for something to be executed in an app. In our case it is finding the right Trello board/list, as well as, creating cards Pick the board and the list we want to create cards in. (8) This is when things get interesting because we want to pick a list based on the input from the Paperform. In this case if a user has entered ‘Monday’ in the form, then we want Zapier to go into Trello and pick the ‘Monday’ list to create cards in. This is why we choose ‘What Day Of The Week Will You Go? Answer’ as the ‘List Name’. Next we add another Trello action to our Zapier flow. This time pick ‘Create Card’. (9) Basically we want Zapier to write a new card based on the information from the Paperform that users have filled in. The Name of the card and the Description of the Trello card all comes from fields users have already filled in when they submitted the Paperform. (10) The Gotcha’s — When you are in Zapier, picking fields from the Paperform, make sure to always choose ‘ ‘ fields because they are the ones that hold the data you want. PICK THE RIGHT FIELDS Answer E.g. The two fields below: ‘Pick Your Name — Answer’ and ‘What Day Of The Week Will You Go? Answer. will in this case create the card with the name ‘Andrew @ Taimo Cafe’ because those are the provided by the submitted form. answers — Papeform has a wonderful new function called pre-fill key and it uses a small unique code in the URL. This will sometimes screw up displaying the correct field name in Zapier PRE-FILL KEYS E.g. The field below: ‘What Day Of The Week Will You Go? Answer’ will sometimes show up in Zapier as and that is because if we look at the pre-fill key unique code for that field answer in the Paperform admin, the code is: Ei2sb Ei2sb Don’t worry though, Zapier will always process the field correctly and it knows this particular field is ‘What Day Of The Week Will You Go? Answer’ — In Step 10 above, if you look at the List field, it says ‘ID’ and not the Name of the list. Why is that? ZAPIER FIELD ID’s Let’s look at an extract of that screenshot. It is because Zapier sometimes can’t process the name of a field and absolutely must have the unique ID of the field. This is explained more here: Zapier Field ID’s You just need to be aware of this and sometimes give Zapier the ID of a field instead of the name of a field. Often the reason a Zap doesn’t work is exactly because of this. Conclusion & Additional Information I hope this post shows what is possible with some Zapier and Paperfom magic If you want to see the complete proof-of-concept app, go to take a look. here Below I have added some additional resources. Happy Integrating Everybody! Reference Links: Connecting Paperform to Zapier Tutorial Other Zapier Apps that connect to Paperform What’s pre-filling on Paperform? Zapier Field ID’s This is story is brought to you by Hacker Noon’s weekly sponsor, . Create beautiful forms. For Hacker Noon readers, . Learn more about (spoiler alert — this form is powered by ). Paperform 10% all plans forever becoming Hacker Noon’s Weekly Sponsor Paperform

Stories, News & Mentions about PAPERTON

Paperton’s Stories on HackerNoon

Paperton's News & Mentions