Chapter 3:
Building a craft-driven organization
The forces behind the scene
Zorro:
Zoho has multiple data centers (DC) across the world, storing information of millions of end users. Zorro is the team responsible for Zoho's data center operations. The team identifies and mitigates threats, provides necessary assets, and continually monitors and improves the DC's performance without compromising the security and privacy of end users.
Sysadmins and PitStop:
Sysadmins, or system administrators, monitor corporate servers and manage employee requests. Their responsibilities can be categorized as user management, infrastructure management, and asset management. Since we're a fast-growing organization, managing employees' device-related requests demanded the creation of a dedicated team: PitStop.
The PitStop team is responsible for activities relevant to employees like:
- Providing laptops for new employees.
- Installing operating systems and other required software on devices.
- Maintaining antivirus solutions for all user devices, test machines, and servers.
- Maintaining and configuring Wi-Fi throughout the campus.
- Updating security patches.
- Maintaining hardware and software inventories.
PitStop technicians are stationed on each floor. Employees can drop by or contact the team via email or internal communication tools.
User Management
- User device onboarding and offboarding
- User device requests
- Device repair
- Software installation
- Operating procedure documentation
Infrastructure Management
- Servers and application management
- Repository management and data backup
- Wi-Fi configuration
- Media disposal
- Patch management
- Antivirus management
Asset Management
- Inventory management
- Hardware and software evaluation and procurement
Responsibilities of sysadmins
Network operations center (NOC):
The NOC team monitors local area networks (LANs), wide area networks (WANs), and network devices. The scope of work includes analyzing problems in network devices, troubleshooting issues, reporting incidents, communicating with site technicians, and tracking problems to resolution. The NOC team monitors DCs and network devices such as routers, switches, and firewalls while the Zorro team manages servers and load balancers as part of server monitoring. Physical access to the center and logical access to the tools is highly restricted.
All servers, core network components, and their processes are monitored both internally and externally. Zoho uses a primary internal monitoring tool that provides failure alerts via a dashboard. This dashboard is displayed in the NOC, which is monitored around the clock. We also have an in-house developed data centre life cycle management (DCLM) tool that allows deployment, monitoring, and management of services, components, and servers, etc.
Zoho's DCLM tool is the DevOps console for the Operations (Zorro), Network (NOC), and Service Delivery (SD) teams and is the management gateway to all deployments. These three core teams work together to keep our IT infrastructure up and running smoothly. You can learn more about their roles and responsibilities in this IT operations management e-book.
AI at ManageEngine
A recurring question from customers and tech enthusiasts is, "How does ManageEngine keep up with new trends and the constantly evolving IT landscape?" The answer is that R&D has always been a focal point, regardless of which stage of growth we're in. Nearly 40% of our revenue is allotted to extensively studying and identifying what the market would demand tomorrow.
Rajesh Ganesan, president of ManageEngine, talks about the role of AI in business and IT: "Businesses always undergo phases of change and every business is leveraging technology to change their basic functions and offerings. AI is part of this next phase of reinvention that businesses will undergo. It is being incorporated heavily into company roadmaps across almost every industry."
With AI being incorporated into multiple aspects of IT infrastructure management, organizations will have to streamline their existing processes to adapt to the changes that come along with implementing this technology. At ManageEngine, we are working on AI under four broad divisions:
- Statistical machine learning
- Computer vision and optical character recognition (OCR)
- Natural language processing (NLP)
- Video and voice processing
| Feature | Description |
|---|---|
| Univariate anomaly detection | Identification of unusual spikes in any time-series plotted against single metrics |
| Multivariate anomaly detection | Detection of aberrations caused by multiple or a combination of several parameters using matrix sketching algorithms |
| Forecast engine | Prediction of future data points based on both the past and the present data. It uses customized models such as logarithmic, exponential, polynomial, and power models |
| Malware analysis | Analysis of hardware components to identify pirated data and potentially malicious files entering the organization, and protect the system |
| Markov chains | A probablistic model that aims to identify the likelihood of each transition in a sequence of events |
| Error classifier | Classification of errors using the longest common subsequence algorithm where when given the original and the corrected sentences, it differentiates between them and identifies the corrections made |
| Curve similarity (via dynamic time warping) | A tool to find a specific pattern in graphs when asked in a query |
| Blockchain | A decentralized open ledger system to store data, making it tamperproof and securing it further with public and private keys |
| Bot script | A bot framework that helps in the creation of virtual assistants and other help bots using NLP |
| Web spam detection (via UBEA) | Tool used to identify if a form is filled by a human or a bot (spam) by tracking the pointer movements |
| Sentiment analysis | Model to understand the mindset of the sender behind the text or mail |
| Ticket handling automation | An ensemble of models to automate the help desk ticket handling system |
| OCR | Tool used to extract text from text-superimposed images, files, handwritten documents, etc. |
Some of the features used across ManageEngine products and operations
Zia
Zia is a continuously learning AI-assistant. Zia is used across ManageEngine solutions like endpoint management, help desk, and advanced analytics tools to perform various functions.
What can Zia do?
- Execute commands via a voice assistant.
- Offer virtual support.
- Provide relevant predictions.
- Assist users via a chatbot.
- Provide automated insights (Ask Zia).
- Deliver contextual suggestions.
It's worth mentioning that Ask Zia is one of the first NLP assistants that we started building. The R&D team started working on this way back in early 2014. It took almost five years for the feature to release in 2018.
Privacy first: how do we do it?
Privacy is not just a legal obligation but an ethical choice.
Privacy is increasingly becoming an important subject in the industry. While the market focus is mostly on B2C, we are trying to bring the attention to B2B. Protecting user privacy has always been a top priority, and in 2020, we took a stance and blocked adjunct surveillance. This meant eradicating the trackers from our website and products, particularly the trackers from companies with ad-driven business models. We also have strong policies in place and we're constantly working on new methods to keep customer information secure.
Ulaa:
Ulaa (which means exploring in Tamil), is our privacy-centric, enterprise-enabled browser. It is built on Chromium, the open-source browser. As a business with a comprehensive suite of primarily web-based products, our teams felt the need to acquire a deeper understanding of the underlying platform—the browser. Creating a browser tailored to our needs also enables us to provide more extensive and meaningful integrations of our products for customers. Ulaa is currently in its beta version and has multiple features in the works. Our goal is to make Ulaa the default browser for employees eventually.
Security and privacy awareness training:
The objective of our annual training is to reinforce the importance of data security and privacy. It covers data security and privacy at Zoho, the challenges we face as an enterprise, and the expectations from employees to ensure a secure and privacy-first Zoho. The Central Information Security and Privacy Team is responsible for initiating, preparing, and curating the content for the program. Employees attend modules and take up assessments. Each team is assigned a privacy champ who ensures all employees take part in the session. Non-compliance results in low data protection and compliance scores, followed by corrective actions.
Data Privacy Day:
Data Privacy Day is an international effort held on January 28 every year to remind individuals of their rights, and empower employees and businesses to accept their responsibilities of safeguarding personal data and enabling trust. At Zoho, we celebrate Zoho International Privacy week (ZIP), during which the privacy team hosts events, shares resources spreading awareness, and educates teams through past incidents and hypothetical scenarios. In 2022, we also introduced a privacy bounty program for employees to identify and report privacy loopholes or bugs in our tools and processes.