Real-time analytics has transformed daily decisions and delivers fast dashboards from streaming data
What is our primary use case?
My main use case for SingleStore is real-time analytics and data storage and serving for my ML workloads. It enables me to run low-latency analytics and model-driven use cases at scale, which is quite difficult for OLAP and OLTP databases alone.
One specific example of how I use SingleStore for one of my model-driven use cases is as a data warehouse for fintech, which processes payments through TC worldwide. The integration with Azure Data Factory occurred without complications. It helps me in an excellent way since I am very fast in obtaining the data for my dashboards. Additionally, the compression of the information is accurate.
Regarding my main use cases with SingleStore, I think it is very useful for managing information for business intelligence processes and processing. Given that the information is brought quickly, it processes correctly when loaded into the dashboard.
What is most valuable?
The best features SingleStore offers include fast data recovery and data compression by 80 percent. Having the information in sheets helps me to process the information quickly. Simplicity in T-SQL is another aspect I appreciate.
Data recovery and sheet-based processing help my team on a day-to-day basis by enabling us to handle information efficiently.
I would like to add that the data compression by 80 percent helps us in an excellent way since we are very fast in obtaining the data for our dashboards and the compression of the information is great.
SingleStore has impacted my organization positively by enabling us to run low-latency analytics and model-driven use cases at scale, which is quite difficult for OLAP and OLTP databases alone. It has been very helpful because our internal clients are happy to have the data and make data-driven decisions easily. Making decisions based on data within a two-hour delay to the transactional database is excellent since we went from twenty-four hours to two hours. I think the best contribution is decision-making with data that is close to reality.
Reducing that delay from twenty-four hours to two hours has significantly affected my team and business outcomes by increasing productivity. We have been able to serve all our customers, and they are very happy. We can deliver to them on time.
What needs improvement?
SingleStore can be improved in the aspect that the Azure pipelines do not have many parameterization features compared to others, for example, AWS. Also, error handling needs attention. When it fails due to memory, it only indicates that but not exactly in which process it failed.
I would like to add that we need to know what the roadmap of SingleStore is.
An additional improvement needed is that the semantic layer can be better. Currently, it requires significant development experience to fine-tune queries.
For how long have I used the solution?
I have been using SingleStore for six years and some few months.
What do I think about the stability of the solution?
SingleStore is stable as I find it extremely stable. I have not seen any downtime.
What do I think about the scalability of the solution?
SingleStore's scalability is high and it can be used by any size of organization and can handle any needs of any organization.
How are customer service and support?
The customer support is very proactive and responsive twenty-four hours per day, seven days per week.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
I previously used a different solution, which was Oracle database.
I switched from Oracle to SingleStore because SingleStore outperforms based on my analysis with simple data sets within my organization.
What was our ROI?
I have seen a return on investment as real-time compute reduces the hassle of fine-tuning required, but it is minimal compared to other tools like Oracle or other RDBMS. Primarily, it was not client-focused. It was used for large enterprises with a significant amount of real-time data generation. The objective was to scale as data loads with high-performing query model responses.
What's my experience with pricing, setup cost, and licensing?
My experience with pricing, setup cost, and licensing is that it can be a bit expensive for startups, but for my organization, the cost is affordable and cost-effective.
Which other solutions did I evaluate?
Before choosing SingleStore, I evaluated other options such as PostgreSQL.
What other advice do I have?
My advice to others looking into using SingleStore is that it is extremely good for scenarios where large sets of data are generated in a day and the data is streamed. If you would like to run queries and analytics on such data, it would really scale and outperform TimesDB or Oracle in-memory options.
Before we wrap up, I have additional thoughts about SingleStore, stating it is a great tool that is very cost-effective. It shines as a unified solution for high OLTP and OLAP workloads. I rate this product a nine out of ten.
Which deployment model are you using for this solution?
Hybrid Cloud
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Singlestore is best in market compared to Google's Bigquery
What do you like best about the product?
I like the scalability, versatililty in the product with architecture itself which can be used for multi models.
What do you dislike about the product?
I feel singlestore is not beginner friendly as early stage developer in AI era.
What problems is the product solving and how is that benefiting you?
SingleStore is best when I am creating saas applications for modern scaling architecture, multitenancy and easily customisable products.
SingleStore speeds up my daily workflow
What do you like best about the product?
I use Singlestore daily, and it has great features. Their docs are very thorough, which makes it easy to find answers to question. Their support is very responsive.
What do you dislike about the product?
Some of things inherent with the implementation of SingleStore make some tasks tedious, such as updating an enum column.
What problems is the product solving and how is that benefiting you?
SignleStore allows our users to get data fast. The MySql flavor it uses makes it easy for us to query the data and find insights quickly. It also allows us to ingest massive amount of data quickly.
Database that changed our company
What do you like best about the product?
Singlestore has made our data processing much easier and faster with its tooling
What do you dislike about the product?
Pipelines can be slow and there's not very many logs around them
What problems is the product solving and how is that benefiting you?
Data Processing -> makes cleaning & processing data easier
Singlestore review
What do you like best about the product?
The fact that it's the best of both worlds of row store and column store. It's easy to use and setup. We've been using it for a lot of our high throughput use cases and it has worked wonders.
What do you dislike about the product?
A lot of the technology and functionality as it comes with a proprietary database is under the hood which makes is difficult to debug issues. And it takes a lot of back and forth with their support team to explain the issue and hopefully get a resolution.
What problems is the product solving and how is that benefiting you?
Giving us the best of both worls of row store and column store so that we can do real time aggregates as well as individual row retrieval.
Great near real-time analytics database
What do you like best about the product?
I have had a great experience using SingleStore. It can ingest massive amounts of change events from Kafka and keep up. It ingests this data within 3-5 seconds of update in our main application. Most of the queries we send to SingleStore come back within 5 seconds and work well with highly concurrent (thousands of customers) environments. For providing customer-facing dashboards and reporting, it is really fast and everything I expect. It also has a great ecosystem with its managed service running in AWS, Azure, and GCP, with support for customer-managed keys and other enterprise features we need. It also has cool newer features like database branching and fast JSON queries (it treats JSON like any other columnar data). It is like a really fast version of Snowflake or Databricks. It is easy to get up and going with SingleStore pipelines (no external tools like Kafka Connect or Apache Flink are required since they have SingleStore pipelines), and customer support is great too when I wasn't able to figure it out on my own.
What do you dislike about the product?
It isn't open source, which I prefer. Despite being proprietary, they do make it fairly easy to use their free version on Docker or their managed free forever solution (but quite limited). The sales/procurement process can be a bit onerous too.
What problems is the product solving and how is that benefiting you?
Faster ingest and query speed than most competing products. This results in a better user experience for our customers since they are often creating ad-hoc dashboards in our product and expect the experience to be snappy. As they are building new charts, reports, and dashboards, they iterate through different versions and tweak the settings, and each time they do this, they want the queries to come back within a second or two. Then, when they have 5-10 widgets on a dashboard, they expect those to come back within a couple of seconds. This great experience enables us to democratize reporting for our users and are delighted with our work management product.
User Friendly and Reliable in-memory DB
What do you like best about the product?
Accessing and Loading relational data into tables is pretty efficient, reliable and fast.
Having Singlestore employee as a consultant in the organisation helped us get answers from Singlestore to resolve a lot of minor issues and questions we had with this product.
What do you dislike about the product?
It lacks replication product to replicate data on table by table basis from one cluster to another remote cluster.
We need a reject file created for all rejected rows during pipeline load of millions of rows of a table, instead of abending at the first rejected row.
What problems is the product solving and how is that benefiting you?
Transferring Data from Mainframe DB to in-memory DB on distributed platform is our long term goal. Also one of directive from our management is to have a never-down DB and Singlestore is the DB we are going to use to achieve it for our application.
Successfull Co-Innovation Partnership
What do you like best about the product?
Query and Ingestion Speed at Scale over Multi-Model Storage w/ Simplicity .
What do you dislike about the product?
Medium UX for Self-Hosted versions and medium materialization features
What problems is the product solving and how is that benefiting you?
Chatbot, Customer Mobile Banking RT Analytics , Customer Intelligence for RT decision (multiple scenarios)
An excellent choice for diverse data processing needs with exceptional in-memory capabilities, robust failover mechanisms, easy scalability and high performance
What is our primary use case?
I use it for managing both transactional and analytical workloads within the same database. In my previous organization, I successfully implemented it for a banking system, where it accommodated transaction-based processes using row store tables and analytical requirements using column store tables. This dual functionality eliminates the need for separate databases for transactions and reporting, streamlining the overall architecture. With SingleStore's distributed architecture, it provides the scalability needed to support diverse workloads effectively.
What is most valuable?
Its in-memory storage, distributed architecture, scalability, and failover mechanisms collectively contribute to its exceptional performance and reliability, especially in demanding transactional environments like online and mobile banking systems. The ability to store data in memory is a standout feature, enhanced by robust failover mechanisms. Even in scenarios where all servers experience downtime, it ensures data safety by maintaining a copy on disk.
What needs improvement?
The critical challenge involves optimizing the distribution of data across partitions through careful design of the sharing key. Poor key distribution can significantly impact performance, requiring a backward approach in design rather than adding tables incrementally. Intricate use cases, especially those involving joins across multiple tables, pose challenges if sharing and distribution are not well-aligned. Unlike traditional databases where indexing may suffice, SingleStore may require redistributing the entire dataset, presenting a persistent challenge.
For how long have I used the solution?
I have been using it for four years.
What do I think about the stability of the solution?
The stability of SingleStore varies depending on the use case. For a transaction-based system, I would rate it around eight out of ten. However, if it's utilized for an analytical system, I would give it a rating of around seven out of ten.
What do I think about the scalability of the solution?
Scalability is its key strength. Adding servers for scalability is a straightforward process involving simply incorporating a few additional servers and recycling the cluster triggers automatic repartitioning and redistribution of data. For instance, if the initial database creation involved a hundred servers and later, four more servers are added, specific commands can be executed to increase the partitions to one hundred twenty. The data is then efficiently redistributed across the expanded partitions without the need for manual data movement, ensuring a seamless and efficient scalability process. In my current organization, approximately three projects involve the usage of SingleStore, with a team size ranging from ten to twenty individuals.
How are customer service and support?
During the onboarding process at my previous organization, SingleStore provided dedicated support for five to six months, offering invaluable assistance. Presently, with our current service providers partnered with them, support involves raising a ticket, leading to the allocation of a dedicated person for assistance. This personalized approach enables an assessment of the issue, considering factors like data volume. Additionally, the forums serve as a helpful resource for addressing queries, although responses may take a few days. I would rate it eight out of ten.
How would you rate customer service and support?
Which solution did I use previously and why did I switch?
We transitioned from using IBM Db2 to SingleStore due to a shift in our infrastructure plan. Initially designed for on-premise deployment, we sought optimized server capabilities for a banking process, with a primary goal of cost reduction compared to mainframe expenses. In our current project, SingleStore is predominantly employed for analysis and reporting purposes. Previously, Palantir and Vertica were used for reporting, but observations of drawbacks in these platforms led to the decision to migrate to SingleStore for more efficient analysis and reporting capabilities, which is proving successful in our current setup.
How was the initial setup?
The initial setup is straightforward, with comprehensive tutorials available on its website. Beginners can easily follow the step-by-step guides, either for a local installation or on cloud platforms like Azure.
What about the implementation team?
The installation process is user-friendly, requiring the selection of a cloud provider and a few configuration choices. Unlike on-premise solutions that involve server setup, SingleStore simplifies the process, making it accessible to a wide range of users. For on-premise installations, specifying server details and failover architecture is necessary, but once the server is prepared, the installation itself is uncomplicated. Database creation involves specifying configurations and requirements, and streamlining the overall setup process.
What was our ROI?
The platform's versatility allows it to cater to various use cases effectively. Unlike other databases that might require separate solutions for transactional and analytical needs, it offers a unified solution for both. This dual functionality appeals to organizations seeking cost-effective solutions, as they can invest in a single database to address multiple requirements.
What's my experience with pricing, setup cost, and licensing?
Using it for analytical purposes can be cost-effective in the long run, especially in terms of infrastructure. While building an on-premise cluster incurs an initial cost for servers with ample RAM, it becomes a one-time investment with subsequent maintenance handled internally. For cloud deployments, the cost may be relatively higher due to instances offering lower RAM. Opting for higher RAM in cloud instances increases the per-server cost. However, it's important to note that this is a one-time expenditure, and maintenance becomes more straightforward.
What other advice do I have?
I would advise individuals to consider it for transactional systems, particularly if their requirement is for millisecond-level performance. The row store feature is well-suited for such applications. However, it's essential to be mindful of the associated costs, whether deploying on the cloud or on-premise. Due to the need for substantial RAM to store data in memory, the cost can be significant, especially for larger datasets. Overall, I would rate it nine out of ten.
Which deployment model are you using for this solution?
Public Cloud
If public cloud, private cloud, or hybrid cloud, which cloud provider do you use?
Amazon Web Services (AWS)
Supports in-memory data types, storing data on RAM for high performance
What is our primary use case?
I worked for a company that outsourced tasks for SingleStore. I mainly worked with one customer who was a video platform. Their primary use case was storing metadata for their videos. They offered recording and playback services for TV shows in the US and Canada, and SingleStore efficiently managed the metadata for all these recordings.
There's a wide range of professionals using SingleStore, You can find more details on their website.
How has it helped my organization?
Since it is not as costly as Oracle or other database counterparts, that's one benefit. And the speed is very fast. It supports in-memory data types, storing data on RAM for blazing-fast performance. That's the highlight; it's perfect for both OLTP and analytical workloads.
What is most valuable?
It's a distributed relational database, so it does not have a single server, it has multiple servers. Its architecture itself is fast because it has multiple nodes to distribute the workload and process large amounts of data. I heard a client processed 3.5 billion records in seven minutes! Their data ingestion is very high, and SingleStore even markets itself as the world's fastest database.
What needs improvement?
For new customers, it's very tough to start. Their documentation isn't organized, and there's no online training available. SingleStore is working on it, but that's a major drawback.
Also, technically, SingleStore needs more features on the SQL part. Most SQL boards work in MySQL, and SingleStore integrated all its sequel with MySQL, so nearly 99% of MySQL code runs on SingleStore. But features like TVF and UDL lack depth. Users have to walk into it, and SingleStore has minimal features there.
IUDF, TVF, and stored procedures are not as advanced as SQL Server's. That's one thing I would like to see improved.
For how long have I used the solution?
I worked with SingleStore just two months before resigning from my previous company. I have over two years of experience with it.
What do I think about the stability of the solution?
It is a stable solution. Any downtime I've seen was due to application or software bugs, not SingleStore. Human errors happen, but the system itself is stable.
What do I think about the scalability of the solution?
It is a highly scalable solution. SingleStore has servers, separated into two parts. One aggregates queries (gateway nodes), and the other stores data (leaf nodes).
To increase database size, you simply add more servers. There's minimal downtime during rebalancing, maybe a minute or two. You can add as many servers as you need without taking anything offline. That's what makes it highly scalable.
How are customer service and support?
We mostly handled day-to-day maintenance. But for the real heavy lifting, there's a separate team in SingleStore. It's all handled through one ticketing tool, Zendesk. You're a registered customer, you log a ticket, and they prioritize and address them accordingly.
How would you rate customer service and support?
How was the initial setup?
SingleStore offers cloud and bare-metal installations.
Cloud hosting is simple; you pay hourly and follow their cloud UI instructions. Anyone, even someone less technical, can install it.
For bare metal VMs, it might take a day for a new technical person, but an experienced one can do it in an hour. It's quite easy.
The customer I supported had an on-premises SingleStore cluster running on bare metal.
What about the implementation team?
One technical person is enough for deployment if it's not a production-grade cluster. For testing, one or two days should be max.
For production-grade clusters, you might need professional services. SingleStore also offers database architects and consultants to help design your cluster architecture, hardware selection, license units, etc. They basically do everything for you, from designing the blueprint to setting up hardware and licenses. So, professional services are highly recommended for production environments.
If you have many clusters (over 10-15), you might need a team of 3-4 people for maintenance. But for 1-2 servers, you can handle it yourself. It's easy after the initial learning curve. And best of all, no downtime! You can perform maintenance tasks online, like adding a new service ID or scaling your cluster, without impacting your business.
The product offers high availability. Two copies of data at all times, so even if one server goes down, your application stays up. That makes maintenance even easier. You can take a server down, fix it, and put it back without impacting users. This makes maintenance very easy.
As long as one server is running, you're good. The only limitation to high availability is the increased cost. You need more servers, which means more money. Think of it like this: without high availability, you'd need X servers. With it, you'd need 2X servers. Hardware costs go up.
What was our ROI?
I had a customer I worked with for five years who kept adding new customers throughout. SingleStore provides significant cost-based value to companies.
What's my experience with pricing, setup cost, and licensing?
The pricing is available on the website, https://www.singlestore.com/pr.... They have two main options: cloud installation and bare-metal installation, each with different pricing models.
For cloud services, they offer three tiers: Standard, Premium, and Dedicated. Standard starts at just $0.80 per hour. This is for the standard cloud service.
Now, for self-managed on-premises clusters, they provide free licensing up to four units. If you need premium features or enterprise support, like direct access to their support team, you'd need to purchase their enterprise license. For specific pricing on those, I recommend contacting their sales team directly.
Which other solutions did I evaluate?
I'm currently focusing on C++, building my resume for web development. I want to be a programmer and build things, not work in a service-based company. That's why I moved on. But I'm happy to help with any service-related tasks or where my two years of SingleStore experience might be useful.
What other advice do I have?
Online resources are limited since it's a new database. So, first, read the documentation to understand the basics. Then, approach them directly and explain your specific use case for the database.
Their sales team is very responsive and can help you get started. I highly recommend this database, but do your research first because online materials are scarce. Just understand the basic terms and policies in the docs, then contact them to set up your clusters.
Overall, I would rate the solution an eight out of ten.
