Scality RING Reviews

Scality RING is used exclusively for S3 object storage. Scality RING is an alternative to Amazon S3, allowing customers to have their own on-premises S3 infrastructure when needed. The features are very close to Amazon, with Scality RING's S3 implementation tending to be very similar to Amazon's standards, including features and user experience.

In my experience with Scality RING, I worked for a big banking company that had on-premises cloud and on-premises S3 infrastructure. They used features such as resiliency and deployment across three availability zones, as well as replication and cross-region replication because they deployed across two regions. There are new features for high performance, such as for AI workloads, but these were not implemented during my time there, as I finished my work with them three months ago and am now in another area.

Scality RING was widely used in our private cloud for storing large capacities and volumes, and also served to store backups for databases, making it very useful. Utilizing S3 is cheaper than SAN or NAS for backup storage, and we can manage large volumes for backups and other uses. While not intended for hot data, it performs well for other cases.

The separation of metadata from object data significantly influences storage performance, as metadata uses NVMe flash storage, serving as an accelerator for accessing data. While data resides on traditional spinning hard disk drives, using NVMe storage for metadata allows for fast access to data, especially for large volumes and a large number of objects. The performance regarding metadata is particularly vital when managing a high number of objects; thus, dedicating servers and fast storage to metadata improves handling those large object counts, which we optimized during our operations, resulting in satisfactory outcomes.

The improvement with Scality RING is the ability to manage huge volumes and the easy scalability when dealing with capacity growth and management. We can buy servers, install them, and hot upgrade the capacity without requiring downtime. Everything is available through an API, allowing end users to manage their storage conveniently without needing requests or tickets; they can allocate, delete, and move storage as required, making operations easier for the end user.

Scality RING has room for improvement, particularly in having a supervisor external to Scality RING, which is key for management. I believe they need to enhance the stability and ease of management for upgrades and migration of the supervisor. Additionally, improving the metadata component is essential, as while we achieved significant enhancements through dedicated servers, NVMe storage, and powerful CPUs, limitations still exist in the metadata engine that need addressing, based on my experience with occasional issues.

I have been using Scality RING for four years and a half.

I would rate the stability of Scality RING as eight.

I find the scalability to be very satisfactory, and I also rate it as eight.

I would rate the technical support from the vendor as six or seven, as it often varies based on whom you are working with; sometimes the support is excellent, while at other times, it falls short. On average, I would rate it as a seven.

Comparing Scality RING with other solutions, I find it to be behind Amazon in the S3 technology space, as Amazon is one of the pioneers. I do not prefer Dell's S3 solutions, which I find inferior. Pure Storage is new to S3 and lacks mature features, although I appreciate their block storage offerings. Regarding Red Hat, using Ceph for S3, I have not tested it but believe it is not yet widely adopted or mature. Scality maintains a strong position behind Amazon due to its experience and large user base.

I find the deployment of a brand-new Scality RING installation to be easy and not overly complicated. However, upgrades and migration are more complex and can be sensitive; sometimes we need support from the Scality team. While setting up a new Scality environment is quite straightforward, extensions, migration, and server replacements often require assistance from Scality, which occurs quite frequently.

Managing Scality RING involves a dedicated team of five or six persons, with half focusing on infrastructure and the others on developing an API for user consumption of S3, as Scality RING already has an API. We have to manage multiple instances of Scality RING, specifically five, with one large installation containing sixty petabytes of data alongside smaller installations.

I was not directly involved in the cost-saving discussions related to the pay-as-you-grow model with Scality RING, so I do not have detailed insights. From a distance, I observed the purchasing process, which was smooth and flexible. This model is indeed helpful because we do not need to purchase all required capacity upfront; we can buy only what we need based on expected growth for twelve months, with annual updates. While not always perfect due to sometimes rapid growth, we managed to anticipate demands adequately, and did not face any capacity issues, adapting as needed while working with Scality.

I assess the simplicity of managing large volumes of data with Scality RING's interface as feasible, but it requires a minimum of skills. With the right staff, managing it is easy, as all the enablers are there; however, I would rate the difficulty as intermediate.

I would rate the overhead as nonexistent with the right sizing of staff managing Scality RING platform. However, if a storage team deals with other technologies, they need to be cautious, as Scality RING requires day-to-day monitoring to check performance and potential issues. Sometimes we need to restart components, including metadata, so it is not just plug and play; day-to-day attention is necessary, but we can leverage vendor-provided monitoring tools, including dashboards to identify potential issues, allowing us to help manage overhead as much as possible.

I assess Scality RING as well-fitted to handle modern workloads such as AI and backup. They have developed a special edition optimized for AI, featuring NVMe storage and a specialized metadata engine, although I personally did not experiment with it. However, I am confident it addresses AI workload needs effectively based on presentations I have seen.

I would rate this review overall as an eight.

Scality RING serves multiple use cases for storage, for GenAI and analytics, unstructured and structured data, backups and recovery, and long-term retention. It can fulfill five to six use cases.

The best features of Scality RING include multiple aspects, but a few that we specifically appreciate are the scalability platform, which allows us to scale our growth. 

Second is metadata indexing and S3 API capabilities that allow us to access the data from one source. 

The use of Scality RING has led to significant improvements in our storage management efficiency. We can estimate the volume with around 20 petabytes, estimating having an NFS file system with 20 petabytes and 30 billion objects.

There are areas that may have room for improvement, and the Scality team is aware of it. You can obtain information directly from them about areas of improvement they need to address. 

We do experience some challenges from time to time when we have to scale up. We do see some areas where we have much higher operations or much higher queries, Scality sometimes chokes up. 

The good news is that the Scality team is aware of it. They know exactly what issues we are facing, what we are looking for, and how to address them. Because of the long-term relationship we have developed with them, when they see our use case and try to address our issues, once they resolve the issue, they ensure that the product is updated with this use case so the next customer does not run into this problem. It's a scale problem, querying problem when the number of operations becomes too high.

I have been using Scality RING for six to seven years, possibly more.

I would rate the stability of Scality RING as having occasional downtimes, bugs, and glitches that we see from time to time on the metadata indexing portion of it. 

Overall, when we see downtime and we try to reach out to Scality, we do get support pretty fast, and we are back up and operational within hours instead of days. This is a massive ecosystem, and generally, 95% of the time, it's healthy and operational, but there are occasional outages that we will see here and there that will throw us back. It comes back up faster with the help and support from their support group.

Regarding scalability, it's quite scalable, but there are some hiccups along the way. I would rate it around eight to nine.

For technical support, I would rate it as high as a nine.

Before we onboarded Scality RING, we definitely did a POC on different solutions. One of the biggest requirements for us is scale. 

As I mentioned, we started with a million objects and in six years, we have grown up to 30 billion objects, so scalability was very important for us. 

In the last five or six years, other technologies have enabled scaling. However, Scality being a partner for HPE, it's also our responsibility to ensure that the partner products are certified and tested so that we can have joint revenue coming to the companies. That's where we have to stick with our partners. If we have another partner, we will test another partner.

The deployment of Scality RING on-premises was not easy when we started. We learned over the six years, and now we know what to do. Now it's easier for us, absolutely easier for us. We know what we need to do if we have to scale up. However, when you do this type of activity of scaling, you will have to engage Scality professional support all the time. If I were to give a suggestion in this area, Scality should come up with a solution where customers can do self-upgrades. 

Customers should be able to do a lot of things by themselves and not have to rely on Scality's professional support. They are aware of it. We have given them that feedback, so this is not going to be new for them.

Every solution, including Scality RING, needs maintenance: time-to-time upgrades, time-to-time patching, and time-to-time vulnerability management. Every time we do the maintenance and this type of activity, it is required to communicate and coordinate with Scality's professional support. 

We have technical team members who can do this work without any problems and without any issues. However, they do end up saying that this type of activity needs to be done under the guidance of professional support. There is a need to do this all the time.

The solution has helped in reducing our operational overhead. If you have 1,000 NFS servers, imagine how many operational overheads I would have to manage, maintain, scale, patch, and vulnerability management on the NFS servers and volumes versus having a single cluster with 30 nodes getting this thing done and staying intact. If I have to gauge, it has definitely reduced on this ecosystem for comparative purposes by at least 20 to 30%.

When comparing Scality RING to other products and vendors on the market, you need to specify which vendors for a specific comparison. If you're talking about MinIO or Weka, each of them has a different use case. 

For our specific use case that we have onboarded Scality with, there are very limited vendors that can meet this type of requirement. Of course, with the technology change and technology shift happening, other vendors are coming up. 

For us, this is the solution that works for our use case. We continue to evaluate other solutions from time to time to see if there are use cases that we can bring in or if there are other solutions that we can try for a different use case. For the use case that we have today in our ecosystem, in our environment, and in our business, Scality is, if I have to rank the top three, it will fit in the top three outcomes at this point in time.

I would recommend Scality RING to other users. However, I always advise any customer or future prospect to do their due diligence on other products. Make sure that your use case gets the solution it deserves, not just because of customer feedback. 

The customer has a specific use case, and that's why we chose Scality. A prospect may have a different use case, and there might be a different solution associated with it. The customer should always do due diligence on multiple products to see which one meets their needs for their use case. If Scality is where it fits in, then we know that it's Scality. For our use case, MinIO did not work because of its scale, for the scale that we needed at that point in time. In somebody's use case, MinIO can work if they want to go horizontally or to smaller rings. Depending upon the customer use case, we recommend Scality if the use case that we have is similar to one that the customer may have.

Our overall rating is 8.5 out of 10.

Which deployment model are you using for this solution?

I work with Scality RING for storage, and we have a concrete strategy to offer it to business customers.

Many customers use it for different purposes, but a common use case is for user computer backups.

I like UTIPI (Unified Tiered Infrastructure Per IOPS) feature in billing. 

Another feature I like is the life cycle management that helps me with data storage efficiency. 

I like that there is no storage limit in Scality RING.  There's no inherent storage limit. Even big companies with many users can reach the practical limits of the system.  

It's a very good product, but for large-scale usage, it has limitations.

Scality RING's lack of extensive testing and potential hidden limitations.

While Scality provides precise architecture guidelines for sizing the storage infrastructure, Orange utilizes a different processor architecture for its Ovi servers, requiring additional optimization beyond what Scality offers. In essence, the product lacks third-party optimization capabilities.

I have been using it for two years. 

We have over a thousand end users and administrators.

The customer service and support are generally good. 

However, sometimes we need direct support, and they might not be fully aware of our specific situation. We have to re-explain our setup, including the conversions we use, the programs involved, and our intended functionalities.

Positive

The initial setup is easy because it is automated. We just need to follow a few pre-requisites. So, the first installation is very easy.

The initial cost (CAPEX) to set up the infrastructure is expensive due to the specific hardware required. 

However, the ongoing costs associated with Scality are expensive due to the specialized servers and additional advice needed.

I would recommend using it. It is a good product for me. 

However, Scality RING is not easy to learn for someone new. It is a little bit difficult. There are a lot of components to it, and you also need to understand them to work with it effectively. It's a complex product. 

The most valuable feature of Scality RING8 is its performance and good interface.

Scality RING8 could improve by having more features. We have to use two automation tools to meet our needs. We would prefer to use only one.

I have been using Scality RING8 for approximately three years.

Scality RING8 is a stable solution.

The scalability of Scality RING8 is good. We are able to add subsets.

Our customers are enterprise-sized companies. The amount of people using the solution depends on the company's administration. However, they are using SAN and NAS and could have more than 100 people working on the solution.

The support from Scality RING8 is good.

I did not use another solution prior to Scality RING8.

The initial setup of Scality RING8 is complex. There are many commands and scripts that need to be launched. If you want to do the implementation yourself there are training sessions available.

There is a lot of documentation for the implementation.

We did not implement the solution ourselves, we used assistance.

The cost of Scality RING8 could be less expensive. It is difficult for smaller businesses to afford it.

I rate Scality RING8 an eight out of ten.

We used Scality during the capacity extension.

I think it's the economic factor. This solution has the lowest cost for storage systems.

If I had to choose a new aspect for the system, it would be the rebalancing. The ease of use during capacity upgrades would be the most important improvement. I don't know if they manage it better right now or not, but it was really horrible to extend capacity on Scality at the time that I used it.

When you add new servers in the node or ring, Scality does rebalancing to split the workload across all the nodes and servers. It moves the data from one server to another.

I did a backup for someone else, so I only have one month of experience with Scality. I did the project though and from there I just followed up with the associate or management of the project.

This solution is stability. 

When we used this solution in 2015, it was not scalable at all. I don't know if they have improved on that, but at the time, scalability was just horrible.

It's used within Orange in France, which handles petabytes of data. I don't know exactly how much it was, but I think we are talking about five, or even more than that. When I worked on this project, it was during migration. It is definitely getting higher and higher in terms of capacity that you can use.

Scality's technical support team is good.

I'm starting to work on a project now with ScaleIO. The engineering team suggests migrating to Red Hat Ceph. I don't know yet what their reasons are, but that was the information I got two weeks ago. I am now starting to look for information comparing the products. For the project I'm working on now, we need block storage, so Scality is not an option in this case. Scality is used for object storage and here we are talking about another solution for block storage. Today we are using ScaleIO, which engineering recommends. It's a common thing to migrate to Red Hat Ceph.

My experience in 2015 was that it was not easy. Each time we needed to get the Scality professional services to help us at that time.

I think I would recommend Scality but there is another solution that I'm going to dig into more, which is OpenIO. Right now for new projects, I don't know yet if I'm going to recommend Scality or OpenIO. Most probably OpenIO, but it's not necessarily based on technical features. It's because we have a commercial agreement with OpenIO.

I would rate this solution as eight out of ten.

S3 and CDMI, restful API for object access, are useful for deploying standard S3 or CDMI-based applications such as cloud backup, email.

Erasure coding technologies permit to reduce storage costs by providing the same security as replication but with less storage resources.

Scality permit to deploy a data/metadata ring on 2 sites, a specific DRP ring on a remote site.

Each ring can have its own QoS (Rep3, EC,…) in order to provide application-based SLAs

We have benefitted from cloud storage adoption.

S3 complete compliance which is planned for v6.

I have used this solution for five years.

Only at the beginning because of a specific implementation issue.

Adding a storage node is simple but rebalancing takes a long time (several weeks).

The technical support is correct.

Scality is the first cloud storage product we have used.

The setup was complex.

This product is interesting if you target a least 10 PB.

We evaluated SwiftStack, Cleversafe, HDS Content Platform, and Mezeo.

Take care of S3 compliance and metadata ring architecture.