Snowflake Data Cloud [Private Offer Only]

It's all about data modernization, building a data warehouse on the cloud, and then keeping it ready for consumption.

For me, there are only two products in the market which are Databricks and Snowflake Data Cloud [Private Offer Only]. Working on either one of them would give me recognition in the market. Databricks works excellently in certain situations, and Snowflake Data Cloud [Private Offer Only] works excellently in certain situations, whether it is high frequency, low latency, or low frequency and high latency type of data pipelines with different batch processing requirements. Based on this type of requirement, you either recommend a Snowflake or recommend a Databricks solution. Before you actually recommend, you also go with a pilot or an MVP which will justify the need. You go with a set assumption and then determine whether Snowflake or Databricks would work excellently for the customer's situation. It has to be done with an evaluation, a discovery, and a pilot, which will help the customer either pick Databricks or Snowflake Data Cloud [Private Offer Only].

Data sharing is important as far as if it is a multi-node or a multi-nation where there is fiduciary involved. You keep certain data inside the boundary of a country versus certain types of data can be shared across, even if it is the same company working across different geographies. Data sharing and data clean room definitely help a lot, and Databricks is ahead in this direction.

Machine learning is not something you do once and then forget about it, forgetting about whether you need to maintain, re-tweak, or retune the model. Machine learning helps in arriving at a decision, but it does not mean that there is no follow-up. There will be a follow-up. Whenever there is a data drift, your outcome keeps changing. When the outcome changes, you need to work on the inventory and retool the algorithm and redeploy. This is what we call model risk assessment or a model risk framework. Machine learning helps, but there will be some sort of framework to put governance on it.

Data enrichment does help to improve strategic planning. Data enrichment is all about dealing with missing data, rounding off errors, and coming up with median and average of different types of data. That will give you a wholesome view into the missing data of, for example, customers, suppliers, or vendors. For example, if you have 10 people and for nine or seven of them there is data missing in a feature set, then you will have a way to fill them up and still create a rounded view so that you can take decisions. In this way, it helps strategic decisions.

Role-based and row-based security are both important in terms of data security. Both help keep the data privacy. One example is the salary of my peer. I should not be able to see it unless I am their supervisor. These measures help in data protection.

Concurrency is a two-pronged strategy. One is you can increase the capability where you house the cloud. The other one is how you manage it within the software. Both are important, and Snowflake has done a decent job in helping concurrency.

Snowflake Data Cloud [Private Offer Only] is definitely in the same direction as Databricks. They are mimicking and morphing themselves to keep the compute and the storage separate, which was one of the first things that was done in Databricks. In a few years' time, I think there will be no difference between Databricks and Snowflake Data Cloud [Private Offer Only]. Snowflake is playing catch-up compared to Databricks.

Something like Databricks Genie would help a lot for Snowflake Data Cloud [Private Offer Only].

I started with Snowflake Data Cloud [Private Offer Only] in May-June of 2024.

It is highly available because the high availability comes not from Snowflake Data Cloud [Private Offer Only], but from AWS or Azure or GCP. There is an unwritten contract between Snowflake and the hyperscaler. So, it is highly available.

It is 100% scalable.

Customer service is good.

The setup is easy and not an issue.

There is ROI. It is successful in many companies. Savings is a well-rounded question. It depends on what your cost is when you are running it on-premise versus what the cost is running it on the cloud, whether you switch off when you do not need a service versus you keep running it even when you do not need it. A lot of details go into ROI. Both Snowflake Data Cloud [Private Offer Only] and Databricks have delivered ROI to their customers. How you get to the ROI is a more complex question to answer than just a simple answer. It is myopic if I say you are saving 10%, but both of them have been able to save money. Before you claim that it has saved money, you still need to start with discovery of the current expenses versus what it is going to be on the cloud, and then look into the same data consumption, the growth in data year-over-year, and other factors. Then there is a projection for ROI. In best cases, I have seen 20 to 30% cost reduction. Some people will say 60 to 80%, but that is on a myopic level. If you consider all parameters, 10 to 20% is what I have seen.

It is all consumption-based. It is that Rolls-Royce versus a Ford car. A Ford car runs through the streets of London, versus a Rolls-Royce that still runs on the street of London, but it takes more fuel consumption. It is all about fitting Snowflake Data Cloud [Private Offer Only] or Databricks. It is about looking into the scenario. Both are quite competitive.

It is round peg to a round hole versus square peg to a square hole. Once you make your determination whether Snowflake Data Cloud [Private Offer Only] works in a certain situation, not all software can work for all situations because the world is made up of too many different types of companies. Many deal with offline data, many deal with high-frequency data, many deal with real-time data. Many deal with API, sending data from the source systems. You will have these challenges. It is hard to typecast and give a stereotyped answer.

I have seen Azure and AWS. GCP is a distant third. My rating for the product is 7.5 to 8, and I would give it an 8 overall.

We are consolidating data from multiple sources into Snowflake Data Cloud [Private Offer Only], into a gold layer, and my team is leveraged for governance and reuse.

Regarding the data sharing feature in Snowflake Data Cloud [Private Offer Only], we have built our own data sharing capabilities, so that is not something we are using.

Regarding machine learning, it is not my use case, so I cannot really comment on whether it improves my decision-making process or what my impression of machine learning is.

On the security side of Snowflake Data Cloud [Private Offer Only], I am a customer of that team. I am not building any security. We are now actually implementing access controls using Immuta.

We are using Snowflake Data Cloud [Private Offer Only] for data enrichment. We are using the data in Snowflake Data Cloud [Private Offer Only] as the raw data for us and then we enrich the data there using a tool called Collibra Catalog and Collibra DQ. Together, we created data products which are more comprehensive.

In terms of technical advantages, other than it being a very modern and good system, I can tell you that we have been looking into competitors and they are pretty much on par. I cannot see that this feature is better in Snowflake Data Cloud [Private Offer Only] versus others. I think it is very available. We have not had any availability issues, and that is something that we can say. I cannot think of anything else.

There are disadvantages in Snowflake Data Cloud [Private Offer Only]; the compute cost is borderline outrageous. I told the Snowflake team that was one of the main reasons that we decided long-term to move to Databricks.

Speaking hypothetically, I would appreciate any additional features added in Snowflake Data Cloud [Private Offer Only].

Anything around data enrichment and data integrations would be very interesting for me in Snowflake Data Cloud [Private Offer Only].

I do not know specifically about Snowflake Data Cloud [Private Offer Only], but Snowflake has been in use for more than six years.

As for the support for concurrent users and workloads in Snowflake Data Cloud [Private Offer Only], it is important to me, and as I said, it is more the data engineers that are looking into that, but as a user and for my users, we have not experienced any availability issues. Snowflake Data Cloud [Private Offer Only] is working fine.

I would agree that Snowflake Data Cloud [Private Offer Only] is approximately 99% stable and reliable, and my experience has been at that level.

I believe scalability in Snowflake Data Cloud [Private Offer Only] is easy to scale, but the limitation is pricing.

As a customer, I think the biggest advantage in Snowflake Data Cloud [Private Offer Only] for me is the technical support.

The technical support is very much available and they do not just react to problems, they actively help us solve our use cases and apply Snowflake Data Cloud [Private Offer Only] in a way that helps our needs and our solutions. Right now, for example, they are helping us stand up an MCT environment around Snowflake Data Cloud [Private Offer Only].

Regarding the technical service from Snowflake Data Cloud [Private Offer Only], if I had to rate support from zero to 10 points, I would give it a nine.

I did not participate in the deployment process of Snowflake Data Cloud [Private Offer Only], but I believe it is on the cloud. I am pretty sure it is on the cloud, but I am not 100% sure.

I am asking about deployment because Snowflake Data Cloud [Private Offer Only] partners with some cloud providers such as AWS, but I am not doing any deployments, so I do not know.

Even though it is pricey, I absolutely see value for money, and until recently, we saw no real competition or real alternatives. But as I said, we went through an evaluation process and we have decided strategically to move to Databricks.

I know we are measuring the ROI of Snowflake Data Cloud [Private Offer Only], but I do not have that data.

I am not really sure about what metrics or benchmarks we use to measure the effectiveness of Snowflake Data Cloud [Private Offer Only].

Regarding the license for Snowflake Data Cloud [Private Offer Only], it is on the higher side.

Summarizing all that we have discussed, I would give Snowflake Data Cloud [Private Offer Only] a rating of eight simply because it is very expensive.

The main projects on Snowflake Data Cloud [Private Offer Only] are with GDO company and retail companies from Italy. We have a customer that uses Snowflake Data Cloud [Private Offer Only] for data enrichment. Before the Snowflake Data Cloud [Private Offer Only] implementation, they used Excel sheets or Access, and we helped this customer to use Snowflake Data Cloud [Private Offer Only] for this feature, not only for the cloud data warehouse feature but also for data enrichment.

I think the data sharing feature of Snowflake Data Cloud [Private Offer Only] is good, particularly when you need to do services for customers or other external partners. At the moment, we do not have a project that uses that feature. The main feature we use is the cloud data warehouse service, and we are starting to work with DBT, which is in Snowflake Data Cloud [Private Offer Only] platform. We are also a partner.

I think that Snowflake Data Cloud [Private Offer Only]'s advanced machine learning capabilities are good, but at the moment we have no projects with this feature or service. We have several projects in AI and machine learning, but not in Snowflake Data Cloud [Private Offer Only] at the moment. If we could do a project with Snowflake Data Cloud [Private Offer Only], it could be good for us. Currently, we work with other platforms for AI and machine learning projects.

I think that Snowflake Data Cloud [Private Offer Only] is a good solution, especially when you have projects based on a SQL-like database or data warehouse. We have a couple of migrations from Oracle to Snowflake Data Cloud [Private Offer Only], for example, and I think that you can add a modern platform while using internal competencies, such as SQL languages or SQL-like architecture. In this regard, I think that Snowflake Data Cloud [Private Offer Only] fits well.

I think that the weak feature of Snowflake Data Cloud [Private Offer Only] at the moment is the ingestion phase. Snowflake Data Cloud [Private Offer Only] bought an open-source tool to perform this step in the implementation of the cloud data warehouse, but it is not as strong compared to others.

I do not remember the company that Snowflake Data Cloud [Private Offer Only] chose or bought, but the integration phase when you load data from the sources to the data platform is currently poor in Snowflake Data Cloud [Private Offer Only]. You have this feature stronger in the competitors. For example, if you go to Microsoft, you have Azure Data Factory, and if you choose AWS, you have Glue, and these tools are stronger than the one inside Snowflake Data Cloud [Private Offer Only] at the moment. I think that they probably will work on this feature in the future. This is the data integration feature.

We have been working with Snowflake Data Cloud [Private Offer Only] since 2021 or 2022.

In this case, we have customers who have concurrent requests or queries because they are large companies. We have different people that use Snowflake Data Cloud [Private Offer Only] platform at the same time, and in that case, we found good performance, and the customers are happy to use Snowflake Data Cloud [Private Offer Only].

I think that Snowflake Data Cloud [Private Offer Only] is well-supported. I do not remember if we use it a lot of the time, but I know that my colleague used it, and it sounds good for them.

I think that the implementation of Snowflake Data Cloud [Private Offer Only] is the most simple architecture that you can choose to implement a modern data platform at the moment. Some years ago, it did not have all the features for a modern data platform. For example, there were not AI features or other important features, and it was only a cloud data warehouse. At the moment, if you choose Snowflake Data Cloud [Private Offer Only], you have all the features that are required for a modern data platform, and you implement it in a very easy way. For example, inside Snowflake Data Cloud [Private Offer Only], you have DBT Labs inside the management monitor tools, which is very helpful because you do not need to install DBT Labs on a separate server or service. Also, the AI feature is inside Snowflake Data Cloud [Private Offer Only], so you do not need any separate installation and configuration.

I think that the price for Snowflake Data Cloud [Private Offer Only] could be competitive, but you can do good monitoring of the platform because you pay based on consumption. You need to monitor it, and we usually help our customers with a FinOps monitor. I think that when the project is live, the price is right for the solution that we have implemented. It needs control.

I rate Snowflake Data Cloud [Private Offer Only] as an eight out of ten.

I use Snowflake Data Cloud [Private Offer Only] for different layers, particularly in the fintech part, where I have handled various MGUs. For each MGU, I have to create a different environment; for instance, if it's in a platinum layer, I describe their business role in platinum and upscale it according to the uses. Mostly, the data refining part is done in the gold layer in my business rules.

The best features of Snowflake Data Cloud [Private Offer Only] revolve around its main concept, which is the separation of storage and compute. This allows for data processing or storage using virtual warehouses. If I want to create a larger computing cluster, I can manipulate it according to the layers. It offers elasticity and instance management, enabling automation for auto-scaling or auto-suspending tasks. Most importantly, the zero-copy cloning feature prevents duplicate copies of databases, saving time in larger databases by eliminating confusion about instance or database names. When configuring AI/ML workloads in Snowflake Data Cloud [Private Offer Only], it proves beneficial as it supports structured or semi-structured AI/ML workloads, allowing direct integration of LLM or GenAI models for data processing.

If I am applying any ML methodology or decision-making process in Snowflake Data Cloud [Private Offer Only], I can eliminate null data from massive data sets without needing to apply numerous filters. The ML model can bifurcate the data on its own. When integrating raw data, I can utilize feature engineering, such as Snowpark, to automate decisions in the data interface. By using streams or tasks, automation becomes a straightforward process, making it much easier to work with ML models in Snowflake Data Cloud [Private Offer Only].

In the context of Snowflake Data Cloud [Private Offer Only], sharing data across teams becomes seamless. My colleagues can access shared data live or in read-only format without needing to mention each MGU repeatedly. All business logic is embedded in that particular data set, and metadata links can be provided to users. The core methods of sharing include direct data sharing, Snowflake Marketplace, and Reader accounts, each with distinct practices. Direct data sharing allows sharing with existing Snowflake customers, while Snowflake Marketplace helps monetize data or access third-party public datasets. Reader accounts facilitate direct sharing with clients or partners.

When working with larger datasets, the data enrichment process in Snowflake Data Cloud [Private Offer Only] follows a pipeline that starts with Snowflake Marketplace. The enrichment can involve using external functions and APIs or leveraging Cortex AI and Snowpark. Snowflake Marketplace is utilized for public or commercial purposes, while external functions can provide real-time enrichment from third-party APIs. If LLM-driven enrichment is required, Cortex AI and Snowpark are options, accommodating the dynamic nature of data changes, such as interest rate fluctuations calculated based on geographical coordinates.

Snowflake Data Cloud [Private Offer Only] features built-in security measures, including automatic data encryption and TLS, which I can leverage for secure data transactions. I use dynamic data masking and Row Access Policies (RAP) for enhanced security controls. For security regarding external functions, tokenization options are available. Depending on the model strategy, I can implement perimeter hardening with strict network policies, and for governance, I utilize flatten or RBAC hierarchy methodologies.

Snowflake Data Cloud [Private Offer Only] enhances user flexibility with multi-cloud capabilities, allowing cloud agnosticism. Unlike Google BigQuery, which is tightly bound to Google Cloud, Snowflake Data Cloud [Private Offer Only] offers a Snowgrid for proper data organization across various cloud services. While Snowflake Data Cloud [Private Offer Only] has some manual operational overhead as a SaaS platform, it provides a robust structure compared to its competitors such as Databricks. The container service in Snowflake Data Cloud [Private Offer Only] allows direct app execution, offering unique advantages over traditional applications.

I have been working with Snowflake Data Cloud [Private Offer Only] for almost five years, specifically both Snowflake Data Cloud [Private Offer Only] and Streamlit, so it has been around four to five years.

Snowflake Data Cloud [Private Offer Only] inherently supports concurrent users and workloads through multiple clusters. It is designed to manage thousands of users simultaneously, especially with heavy data loads or BI databases. The native features allow handling multiple concurrent users with horizontal scaling. I can automate additional clones on a single larger server to manage workloads effectively, isolating the total data set to avoid contention.

There are performance considerations during transitions; higher complexity queries may slow down vertical scaling, while horizontal scaling effectively manages more users at once. By mimicking cluster size for complex projects, I can allocate resources appropriately. Horizontal scaling allows automatic spinning of identical clusters, facilitating quick adjustments based on demand.

Setting up Snowflake Data Cloud [Private Offer Only] is generally straightforward with account creation and initialization, though the enterprise version may pose some challenges, especially regarding multi-cluster scaling which can take a few days to finalize properly. Choosing the right cloud provider, such as AWS or Azure, depends on geographical regions, which can complicate authorization requirements. Proper configuration, especially roles such as data engineer over account admin, is crucial for successful setup.

I find Snowflake Data Cloud [Private Offer Only] cost-effective when compared to other tools, indicating a favorable return on investment.

Snowflake Data Cloud [Private Offer Only] offers a 30-day free trial, which costs $400 post-trial for MVP stage users, with different pricing based on warehouse size and features. The standard credit is $2, increasing for additional features including data sharing and security. The Enterprise version starts at $3 per credit, with further costs for critical features at $4 per credit, enhancing data security with provisions such as HIPAA. Post-setup costs are dependent on virtual warehouses and additional cloud services, which affect the overall pricing strategy.

While using Snowflake Data Cloud [Private Offer Only], I must define the working layer and establish the architecture. During MVP stages, this is manageable, but in production, it is essential to maintain structured data to avoid analytics errors. Proper architecture and workflow are keys to seamless pipeline creation in Snowflake Data Cloud [Private Offer Only], ensuring smooth operational processes. I would rate this product an 8 out of 10.