Object Storage? What’s that?
For the past couple of months, I have been speaking with a few parties in Malaysia about object storage technology. And I was fairly surprised with the responses.
The 2 reports
For a start, I did not set out to talk about object storage. It kind of fell onto my lap. 2 recent Hitachi Vantara reports revealed that countries like Australia, Hong Kong and even South East Asian countries were behind in their understanding of what object storage was, and the benefits it brought to the new generation of web scale and enterprise applications.
In the first report, an IDC survey sponsored by Hitachi Vantara, mentioned that 41% of the enterprises in Australia are not aware of object storage technology. In a similar survey, this one pointing towards Hong Kong and China, the percentages were 38% and 35% respectively. I would presume that the percentages for countries in South East Asia would not fall too far from the apple tree.
How is Malaysia doing?
However, I worry that the percentage number could be far more dire in Malaysia. In the past 2 months, responses from several conversations painted a darker hue about object storage technology with the companies in Malaysia. These included a reasonable sized hosting company, a well-established systems integrator, a software development company, several storage practitioners in Openstack and a DellEMC’s regional consultant for unstructured data. The collective conclusion was object storage technology was relatively unknown (probably similar to the percentages to the IDC/Hitachi Vantara reports), but it appeared to be shunned at this juncture. In web scale applications, Redhat Ceph block and files appeared popular in contrast to Openstack Swift. In enterprise applications, it was a toss of iSCSI and NFS.
Image from https://zdnet4.cbsistatic.com/hub/i/r/2018/04/24/c79e9dfb-b4a9-46bb-b831-f2c57fdf8a1d/resize/470xauto/5e4846d1bc7a034c382baf6dcbb612ed/cloud-storage.jpg
I was rather bemused with the reports’ findings and the conversations I had. The deluge of data is already breaking the seams of traditional storage and many organizations want to use the right storage technology to delve into data analytics. These organizations have already accepted cloud storage as part of their strategy, and yet, the lack of technology understanding to take advantage of object storage could stifle their long-term plans with their data. In additional, object storage technology may be the best unstructured data storage technology for data immutability, content preservation and awareness, web scale level of scalability and reliability and security. In a growing trend, enterprise applications such as Hadoop is exploring object storage as a storage repository for smaller files. Apache Hadoop Ozone is one project addressing the billions of small files requirement, something very relevant for the trillion metadata objects produced by the object storage technology for data analytics. From geospatial data, to real-time manufacturing devices, from the descriptive details of a single genome to the metadata associated to the horizontal and vertical coordinates of the CRS, metadata is the definitive resource in the big data world.
Public and private cloud storage
Amazon Web Services S3 is probably the most prominent of the public cloud storage services, while Dropbox and other EFSS (Enterprise File Sync and Share) like Box, Google Drive for Business and OneDrive for Business simplified the concept of files and folders with the foundation of their own object storage technology.
In the private cloud storage setting, there are plenty of them out there. Personally, I have engaged with OpenIO and Minio, so much so that my company, Katana Logic, took up the Minio resellership after several meetings with Anand Babu (of GlusterFS fame) and Garima Kapoor.
What’s going on in Malaysia?
I have spent some months observing and trying to understand this oddity. Perhaps it was the complexity of setting up of the object storage that has deterred the adoption. I have tried my hand with Openstack Swift with the Okata edition some years back, and there was definitely a steep learning curve, even for me. I have also explored the Redhat Ceph object storage implementation from a non hands-on point of view, reading and picking up the commands just to understand what things were involved getting Ceph up and running in a production environment. Again, Ceph implementation did require a skilled Linux administrator to get things going. A Linux administrator, especially a good one, is a luxury and many companies do not have experienced admins or engineers to build their own object storage platforms.
My second unsubstantiated conclusion was the fear of the unknown. Nobody wanted to be a guinea pig, and many Malaysian companies exemplified that perception. Perhaps only a handful of companies have used object storage in Malaysia, especially on-premises. Malaysia beloved private TV and digital content service provider, Astro, has been the poster boy of Amazon Web Services in the South East Asia region for years, and they have been reaping the rewards of S3 object storage technology.
The need to gain competitive advantage in any industry, in any business has escalated. Data is the new oil and information derived from data is the fuel that drives insights and actions. The object storage technology is best to consolidate data from multitude of sources into a very, very large repository. Data associations in a vast repository would accelerate learning, and deepen insights. Just like our brain, the more associations and linkages in a neuron, the more intelligent we could be.
Existing file systems often bet upon by companies would be complex and costly when it is required to scale into the realms of world wide web. The inherent HTTP nature and Restful APIs would be a much, much better fit for web scale applications. Data immutability, where the data cannot be modified once the object has been created is another great feature for compliance and security. But the most powerful feature in my books is the amount of metadata that is produced in storing data in object storage. Data analytics would probably go through the roof with the right metadata.
In conclusion…
Thus, it is high time for Malaysian companies to explore and experiment with object storage. On my end, I have written about object storage in the past, sharing its technology and benefits. The lack of understanding of object storage in Malaysia and in Asia represents an opportunity that without a doubt, will grow fruitfully. I believe Katana Logic picking Minio as the best private object storage technology is a good decision.
Albert Einstein once said, “Everything should be made as simple as possible, but not simpler”. Minio, in my point of view, exemplifies Einstein’s quote. Viva la vida object storage!
