8 min

Tags in this article

, ,

Cloudian, a specialist in object storage applications, is becoming increasingly popular, especially because object storage is very suitable for storing massive amounts of unstructured data. According to Mike Tso, CEO of Cloudian, the public cloud is not very suitable for storing large amounts of data. It is better to store it closer to the source, such as in the (local) data center or even directly at the edge. This is something that the increasingly popular storage specialist is strongly committed to.

Before we talk about the facts regarding Cloudian, it might be useful to explain what object storage is again and what advantages this method of data storage has.

With object storage, data is, roughly speaking, stored as objects. These objects are given a unique identification number, ‘hash’, and extensive metadata. The addition of the unique identification number and the metadata make it possible to index the (data) objects more easily, to identify them and to retrieve them more quickly.

The unique hash of an object also ensures that data is always directly accessible. In this way, the objects can also be replicated quickly and easily.

All hardware suitable as a storage medium

Object storage does not necessarily have to be stored on special storage equipment. In fact, all standard x86 servers are suitable for this. This makes it easy to expand storage, because it is not bound to the maximum storage capacity of dedicated storage applications.

Companies can decide for themselves which hardware they want to use for storing object storage. When the maximum capacity is reached, they can easily add new capacity by adding nodes. This makes the scalability and the ability to store object storage data endless.

The direct addressing and corresponding simple replication also make it possible that the storage of object storage is not bound to a single location. Data can easily be stored in multiple locations in different data centers, anywhere in the world. This actually creates a kind of data fabric and a copy of the data is always available.

Another advantage of direct addressing is that higher performances can be achieved with large quantities of objects due to a reduced latency that object storage has when retrieving the data than with standard file servers or NAS solutions. The S3 API via http(s) is used to retrieve objects.

Large amounts of unstructured data

All these features, especially the metadata and the unique hash, make object storage especially suitable for large amounts of unstructured data and of course the emerging field of big data. Important applications are the storage of large quantities of X-rays within hospitals or (heavy) video files. Companies like Facebook, Spotify and Dropbox use object storage to store their gigantic data collections, such as user data, photos, music and all kinds of other data.

HyperStore from Cloudian

Cloudian offers its own platform that should facilitate the storage and management, anywhere in the world, of large datasets for companies and organizations. The HyperStore platform, now version 7, offers not only unlimited storage capacity but also extensive additional functionality such as data protection and various extensive and smart management tools, such as the RESTful API for automating management tasks, says CEO Mike Tso in our interview. With the acquisition of Infinity Storage in 2018, the platform can now also store and manage file-based storage. As a result, the application now delivers the best of both worlds.

Appliances, software-based and suitable for public cloud

The platform is available through dedicated data center appliances and a software solution. These appliances are available in 1U and 4U versions, with the latest version combining two nodes in one device. The available raw storage capacity of these appliances varies between 48 TB and 840 TB. Naturally, the platform also exists in a (virtual) software version. This software solution can run on-premise as well as in cloud environments.

The use of the S3 API makes it possible for end-users to combine the Cloudian solution with all major public cloud environments, such as Amazon AWS S3, Microsoft Azure and Google Cloud Platform. In this way, scalability and redundancy can easily be extended even further.

Public cloud unsuitable for large amounts of storage

Despite the above integration with public cloud environments of our own solution, it is striking that Mike Tso indicated in our conversation that object storage is still much more suitable for storage in local data center environments than in the large public cloud environments. This is what the company is responding to, with its appliances and virtual application.

Cloudian’s CEO compares the rate of data growth with that of Moore’s Law for chipsets. All sensors, from cameras or X-ray equipment, generate all that data that has to be stored somewhere, on flash or something else. When the chips within these sensors are improved, they generate more data, which in turn leads to more storage capacity being needed. This can be done with more storage equipment, but that does not solve the problem, according to the top man. Nobody wants to add (expensive) storage equipment constantly.

Public cloud environments may be the solution, but Cloudian saw from the beginning that sooner or later, there will be problems with the underlying network infrastructure for the speed of data transport. End-users may find that latency occurs in the data transport between the location and the public cloud environments or that certain functionality does not work properly.

According to Tso, the reason for this is that the fibre-optic capacity required for data transport does not automatically grow in line with the amount of data being created. The fibre-optic capacity increases every now and then, when new cabling is laid, but remains the same afterwards. The amount of data will continue to increase constantly, but the capacity and bandwidth problems, and hence the latency, will continue to exist. According to the CEO, this problem can only be solved if there is a parallel growth of data and bandwidth.

Local data center still has a right to exist

From this, you can conclude that instead of a step towards the public cloud, especially for object storage, there is still a place in the (local) data center. According to Cloudian’s CEO, the current image of the market that the (local) data center will eventually move entirely to public cloud environments does not apply.

In fact, he says, there is a development that the public cloud environments actually go to the place where the data is created, especially when it comes to the large, heavy (unstructured) data files of today, which gives(local) data centers the right to continue to exist. With its appliances and virtual solution, the storage specialist is naturally able to take advantage of this. Actually, this makes the mission of the storage specialist really clear: supplying the storage capacity of the public cloud in the (local) data center.

Storage goes to the edge

For the future, Tso sees that the location where data is to be stored will move even closer to the edge. This is because, for example, the arrival of 5G and use cases such as the connected car will cause the enormous growth of data to take place there. It’s time for a Cloudian solution in Tesla cars, the CEO jokes.

More specifically, he indicates that Cloudian is already seriously working on solutions to store data from edge environments such as vehicles, as quickly as possible and with high quality. We are now working with the city of Montebello in the United States to ensure that data from security cameras in city buses, large files, can be uploaded quickly and without problems. For the collaboration, this city first used a NAS solution on which the local storage in the bus was directly uploaded every three hours via a 4G connection for further analysis. This often caused problems when the mobile connection was disconnected, causing the files to appear incomplete or corrupted.

Now the city of California is using Cloudian’s HyperStore solution in their data center. Now offloading takes place via 4G/wifi and the data is stored as S3 files on the appliances of the storage specialist. Advanced and proactive mechanisms within the HyperStore solution ensure that corrupted files and bit rot are a thing of the past. The stored data can, therefore, be analyzed more easily with the help of the storage specialist’s artificial intelligence application Stream. Ultimately, this resulted in greater scalability, cost reduction, S3 interoperability and better metadata for the customer.

Other developments

Besides bringing more storage capacity to the edge, Cloudian is also working hard on the application of artificial intelligence, especially for the metadata brought along by object storage. The storage specialist works together with the AI-specialist Machine Box in the field of machine learning in order to find and search hidden content in object storage even more easily.

It goes without saying that we also work closely with all the major technology giants in the field of networking; from Cisco or HPE to Veeam and from Splunk to Veritas or Commvault. In combination with the various collaborations with distributors and resellers, the storage specialist creates its own ecosystem, which extends as wide as possible.

Focus on growth and customers

Cloudian’s vision in the field of object storage and the importance of the (local) data center has been of great benefit to the company lately. The storage specialist had seen its customer base grow considerably in 2018. According to Tso, Cloudian is particularly popular in the EMEA sector. This region would now represent about half of the company’s total turnover. The company was also able to expand into more and more vertical markets, such as the healthcare sector. The company, which is now in a real scale-out phase or perhaps has already passed, also received no less than 95 million dollars (84 million euros) in new external funding over the past year.

In addition, the CEO has identified many people switching from traditional storage providers, such as NetApp, or products from manufacturers such as Dell Isilion. The most important reason to use Cloudian, as the CEO points out, is that companies and organizations are wondering where they can store their growing data at low cost and with high reliability. In addition, he often sees companies switch to data management or data protection if they come in first with a small problem. These customers then discover that with Cloudian’s compact solution, they save a lot more costs and usually find more use cases that fit in well with the solution.

So, Cloudian has been working hard lately. The company sees a lot of new customers coming and will focus on them before 2019. We hope that this focus on market growth will not be at the expense of new solutions. With its innovation in the field of object storage and the (local) data center, the specialist is fairly ‘disruptive’ and it would be a shame if this focus is overshadowed by only focusing on growth and customers. We remain curious about the future technical developments that the company will bring.