Transcript
I'm really excited to be here. In this video, I'm going to talk about and show you how we have designed and architected Savient's Identity Security Posture Management platform to solve the real identity security needs for today's organization. So with that being said, let me get started. The way I'll be explaining you this is in three parts of our architecture, with part one being all about data ingestion and preparedness, which is part one, data ingestion plus preparedness. To ingest data, you need data sources. What are these data sources? Let's take a look. You have your identity providers, you have your identity sources, and you have your directory systems. That's your first set of systems. The second set of systems are going to be your CMDBs, third party cloud security systems. The third one is going to be your SaaS platforms and your infrastructure as service platforms. The fourth one is going to be your very important ones, your ERPs, SAPs of the world, your on-prem applications and your hybrid applications. From IDPs, you are getting all the different identities, which includes human and non-human identities. From CMDBs, you're getting information about your apps, your assets, and in some cases, even activity. SaaS platform and infrastructure as service platform gives you information about access your identities has, which could be coarse grained or which could be fine grained, as well as the most important dimension, which is identity activity. This gives you an answer of who has access to what, as well as what are they really doing with that access. The same set of information, you also get it from all these different type of applications. Once we have ingested this data, then comes the second step of preparing that data, which stands for C, E, and T. Cleansing, enriching, and transforming this data. You might be thinking about why this step is important. Identity data has been inherently poor in organizations, and it becomes extremely imperative and paramount that before you start strategizing based on this identity data, that it becomes right, it becomes enriched, and it becomes clean, so that you have the right insights to begin with. That basically concludes my part one of the architecture. Part two is about how are you processing this data, and that's where the real magic happens. Let's talk about it. Part two is all about data processing. So in data processing, what I am doing is, I have this magical box. I am calling it as my data lake, where the real magic is happening. Let's understand and demystify what this is about. As we are sending this data, we are streaming this data in form of events into the data lake, which means the first place where this data is going on is a cloud object storage. And this storage, as the name suggests, is responsible for storing all your unstructured and structured data, which sits there to be ready to be processed. Once this happens, this data is going into various different stores. Let's take a look at what these stores are. You have an RDBMS, you have a graph database, you have an analytics store, and you have a vector database. Why these different type of stores? Anytime when the relational constructs of your identity data has to be shown or derived for insights, that is being served from your relational database services. When organizations are looking to visualize who has got access to what, where are the risks residing in their access path, that's all coming in from graph database. Trend analysis, time series, crunching of what has happened historically versus where the trends are going, is all coming from your analytics database. The last but the most important one is, for large language models to interact with this massive data set easily and in a seamless manner, you have to store the embeddings in a mathematical format easily for LLMs to understand, and that's coming from your vector database. So that basically concludes what's happening in my data processing stage, which and where the data which has been ingested from all these platforms is now being processed and is ready to be consumed. That leads to me to my part three of the architecture, which is part three, insights consumption. You have all this data, and now you have derived insights which has to be consumed. Now who are the consumers of this? The consumers could be your end users, application owners, business SMEs, even CXOs, or it could be programs which want to interact with this data through APIs or through streaming information. And the way to do that is through two important elements here. One is a distributed query engine, and the other one is a large language model based NLP interface. Why distributed query engine? We wanted all the personas, whether it's a human or it's a program, to interact with this unstructured and structured data in a seamless manner, and that's what a distributed query engine gives you. LLMs play a very important role because we wanted to ensure that the reliance on BI tools, the reliance on any kind of dependency on technical resources goes away. And that is why LLMs gives you the capability of giving a NLP interface for anybody to interact with this massive data set, unlock that data set in a very simplified manner. And this basically concludes the third part of my architecture, which is you have the insights. How do you consume these insights in a very easy, simple manner? Now as I explained to this whole architecture, the benefit of any customer, any enterprise which they get from this platform are four. Number one is they build and get an inventory of all their identities, all their assets, all their applications in one single place, a massive win for any enterprise. Second is they get insights, or I would say deep insights and intelligence insights for their governance controls, audit, compliance, and risk postures. This is very important when you are thinking about strategizing your identity transformation projects. The third one is your data quality and data hygiene. As I said, the only long pole for an enterprise or an organization in this era of AI is how good the quality of their data is. And that is why we are making a purposeful decision of how ISPM can help organizations to secure, create a data quality process, which allows them to not only get clean, but stay clean as well. Last but not the least, having visibility is not enough. Giving you a way to remediate, orchestrate all your risk, which is existing in your environment and reducing that sprawl becomes paramount. So all in all, we are very excited about this architecture. We are very pumped about what we have been doing. I want to thank all my customers and partners in helping us, working feverishly with us for throughout many, many years, which is now coming to fruition. We would like to help elevate the identity security posture for every organization on this planet. So thank you once again for joining me. I hope this was useful.
