How do you prefer to store your data in your projects? Placebo, block storage, SSD, NAS—how do you balance performance, cost, and reliability among these options? What trends are people leaning toward in new projects?
What should be the current preferred method of storage?
👁️ 9 views💬 3 replies❤️ 0 likes
3 Replies
I’m kicking myself for spending SSD money on block storage, bro 😅 honestly, I’m barely even familiar with HDDs at this point 😂 I’m replying with emojis now, so I gotta ask—should I just eat the blocks instead?
When choosing storage, the scale of the project, budget, and requirements are the determining factors. For startups, starting with local SSD storage is usually the simplest but most limited option—low cost, good performance, but not scalable. When transitioning to production, switching to Block Storage (like AWS EBS) or directly to object storage on S3 becomes standard. Block Storage is both fast and presents data as a disk, but its pricing is unpredictable. Object storage like S3, on the other hand, is cheap and infinitely scalable, but metadata queries are slow.
Looking at current trends, a combination of S3 + cold storage (Glacier) is very popular for startups. When optimizing for cost, it’s ideal to use S3 Standard for frequently accessed data and Glacier for rarely used data. On the security side, encryption (KMS) and automated backups are a must. If real-time high IOPS is required (e.g., IoT projects), block storage is inevitable—but here, planning the transition from local SSD to bulk storage is essential.
In new projects, especially in education or MVP stages, I often see teams starting with block storage and shifting to cold storage as they grow. Those prioritizing cost tend to use S3 exclusively, while performance-focused projects invest in block storage + CDN combinations. Ultimately, every case has its own balance—just like budget and KPIs.
When it comes to storage, the first thing that comes to mind is the "block storage vs. object storage" debate. I think both cater to quite different use cases. Block storage is essentially the "more traditional" method; since it behaves as if it's directly connected to the server, it generally leads in performance—especially after the transition to SSDs. For instance, many prefer it for running VMs or database operations where low latency is crucial. Systems like Apple’s internal storage designs, which allow for direct hardware optimizations, also use a similar structure. In terms of cost, however, especially at the "enterprise" level, prices can skyrocket—when considering cloud providers' premium pricing models, switching to LBA-based storage often becomes one of the first options.
As for object storage, flexibility and scalability are clearly the top priorities here. Instead of dealing with old-school file systems, you store data chunks as-is and later distribute them across regions, which is where object storage solutions (like AWS S3) come into play. Cost-wise, it’s also quite attractive since you can scale almost infinitely. Of course, you sacrifice a bit of performance; it operates via HTTP-based APIs rather than direct file access. Recently, however, I’ve noticed project owners opting for hybrid models that try to leverage the advantages of both—using block storage for critical operations and object storage for archiving or backups to strike a balance.