So let's start at the very beginning. What is OSS? How does it compare to other methods of storing data? So the traditional way in which files are stored is on a file system. In traditional file storage there's a nested hierarchy consisting of files and directories. And there's a file table to keep track of the relationships between these files and directories to maintain this hierarchy. A typical file storage solution or a file system will run on top of block storage, which could be a traditional hard drive or a solid state drive.

Object storage differs from file storage in that there is no hierarchy. There's no concept of a directory. Each object is represented by a key and each key maps to a particular location within the object storage system. Hence there's a table of key value pairs mapping optic names to optic locations within the object storage system. A major advantage of having this simple one-dimensional key value mapping is that object storage systems can scale flexibly. You can add capacity to an object storage system very easily without having any impact on files or objects that are already stored within the system.

Traditional file storage requires the maintenance of a file table that records the relationship between files and directories. And typically there's a maximum size limit on file storage systems. Object storage doesn't have that limitation and can essentially grow infinitely. So let's take a look at some traditional storage system pain points. Let's say you're developing your own off-cloud storage solution. Well, the first thing you're going to need to do is buy a storage server. This could be something like a sand device from a company like Dell or EMC.

What happens when the volume of data increases? Well, you have to buy more storage equipment and it's very difficult to predict in advance how much space you're going to use. And you have to make sure that you plan to have overcapacity. You need to have more storage available than you actually need to make sure that you don't run out. So you have the expense of buying this equipment upfront and doing your best to try to guess how much capacity you're going to need. So you have to evaluate new storage capacity, figure out how much space you need for data backup, figure out how you're going to handle data security. And most importantly, since you have to plan in advance, you have to try not to buy too little equipment and run out of space, but you also need to not buy too much. So you have to worry about cost control.

With OSS you don't have these concerns. That's because OSS can scale transparently. So OSS scales in the background without you needing to think about adding capacity, you're charged only for the storage space you use on OSS, and you can scale up or down as often as you need. So OSS wins versus traditional storage because you can store an enormous number of files easily without worrying about how scaling is done. That happens in the background. You get enterprise grade security. That means encryption in transit, encryption for files at rest. You get high availability, this is the same system we use internally for a lot of data storage at Alibaba.

So we maintain its availability. It's a very high availability service, and it's pay as you go. If today you need 10 gigabytes of space and tomorrow you need 100 and then next month you only need 10 gigabytes again, no problem, you pay only for the space you're currently using. So what are some of the advantages of optic storage? Well, one of the big ones is scalability. A single OSS bucket can hold up to 50 petabytes of data and an essentially unlimited number of objects.

Simplicity is another key advantage. The OSS service provides a simple restful HTTP API, which is accessible from anywhere, anytime. It is in fact a public internet service. Optic sizes can vary between zero bytes and 48.8 terabytes. We have multiple upload methods, normal upload multipart upload and append, which allows you to add data to an existing OSS object. The OSS service is a fully managed service. There is no patching for you to do, no operating system for you to worry about. All you need to do is interact with the OSS API.

There's built-in monitoring, metrics support and access logging as well to help you keep track of your OSS usage. In the security department, OSS provides multi-layer protections, including anti-DDoS protection, up to five gigabits per second. Multi-user isolation and comprehensive logs to help you trace malicious access. Availability is another key advantage. All data in OSS is stored across three different physical servers attached to three different network switches. So all data stored in triplicate. That gets you 11 nines of data reliability and 99.99% service availability.

And of course, the OSS service expands or grows its storage capacity automatically. So you don't need to worry about how that's going to be done. And if you choose to turn on cross-region replication, it's even possible to make copies of all of the data within a bucket, into a bucket in another region. So your data can survive the failure of an entire region. And finally, there's inclusiveness. OSS integrates with both a variety of Alibaba cloud's own services, and can easily be integrated with third-party applications via its open HTTP API.