Solana promises great speeds, but achieving them isn’t straightforward. Blockdaemon is developing a High Availability (HA) Remote Procedure Call (RPC) solution coming in December 2021. This unlocks the blockchain’s true potential and provides a reliable and scalable service.
The Solana blockchain is focused on speed. It’s their main differentiator and key to their success and applicability.
This makes the protocol a seemingly perfect choice for financial enterprises, institutions and exchanges. The faster these organizations can process transactions, the more money they can make.
This of course depends on your ability to actually reach these speeds.
To achieve the stated speeds, you need a couple of things in place.
For one, you need enterprise-grade bare-metal machines. Solana needs high specification nodes. Don’t expect to run on cloud nodes or consumer-grade hardware.
You also need reliability and consistency. With a normal Solana setup, that means dedicated maintenance. After all, the mainnet is still in Beta.
Additionally, Solana is currently vulnerable to certain combinations of requests. A single user can drown a node with the right (or wrong) combination of requests. This can be via a deliberate DOS attack or an accidental trigger.
Blockdaemon has addressed all these issues with our high availability clusters. We handle the infrastructure and resilience, you benefit from the fastest possible Solana service.
3 Main Challenges With Solana Infrastructure
Solana has tremendous potential for speed, but there are challenges to overcome. If you want to run your own infrastructure, you will need to handle the following 3, non-trivial issues:
1. Solana Requires High-Specification Nodes
Solana nodes have the following minimum hardware requirements:
- The fastest AMD EPYC CPUs
- 128GB RAM (Blockdaemon uses 512GB)
- 2TB of data-center grade NVMe storage
That is a serious level of kit. It’s understandable, Solana is built for speed and the nodes must reflect that. It is a significant investment though and is a potential barrier to entry.
Nodes running on substandard machines have no chance of reaching the necessary level of operational performance. If you can’t keep up with the rest of the network you will lose out.
If you are running mission critical systems, you can’t afford this downtime.
Running on cloud-hosted infrastructure for these requirements is less than optimal so Blockdaemon runs on bare metal servers.
2. Node Maintenance is Time Consuming
A big challenge with operating a single node comes from the node startup internals. Node restarts are required after node version upgrades or simply after unexpected shutdowns.
When a node starts it needs to catch up to the network. To do this, it retrieves a Snapshot of the most recent ledgers and ingests information from the network.
This can take over an hour, depending on how fast the snapshot could be downloaded and network congestion.
This amount of downtime is unacceptable for systems used by lots of users, e.g. large dApps, exchanges and institutions. It is highly likely this timeframe would exceed any Service level agreements.
3. Individual Solana Nodes Can be Easily Overloaded
Solana nodes can be overloaded and quickly fall behind if certain RPC requests are sent very frequently. When this happens, RPC latency shoots up across the board for all RPC methods.
Common RPC methods that trigger this behavior are ‘getProgramAccounts‘ and ‘getTokenLargestAccounts‘.
HA RPC Clusters Solve these Solana Challenges
Rather than having to worry about infrastructure challenges, Blockdaemon lets you access Solana as a high availability service.
By using our Solana HA RPC Clusters, you eliminate all infrastructure effort. Instead, you can focus your attention on key business activities.
We have solved the challenges for you:
- High Speed – Our Nodes always exceed the minimum requirements and are housed in environmentally friendly data centers.
- Zero Maintenance – We handle all Node Maintenance and provide you with a flawless service
- Load Resilience – Our HA clusters analyze and schedule requests evenly across the HA cluster. This allows us to handle spikes in request volume
What Happens Behind The Scenes
Behind the scenes, our system design eliminates single points of failure and ensures you get a consistent and high-performance service:
- Rolling upgrades to make sure our service always stays up throughout node upgrades
- All nodes are monitored in real-time to instantly isolate nodes that fall behind.
- Sticky sessions provide a consistent RPC interface (simpler HA clusters suffer from Pub/Sub vs RPC desync)
- Improved snapshot startup mechanism to ensure fast node startup times
Furthermore, Blockdaemon has full-time engineers dedicated to Solana. We work together with the Solana developers to upstream any improvements being made.
We Even Optimize Network Traffic Routing For Maximum Solana Transaction Speeds
With Solana HA RPC Clusters, routing is also geographically aware for optimal performance.
Our nodes run in Solana data centers all over the world. We know how important speed is to Solana users, especially when dealing with high-frequency trading tasks. It is therefore vital to minimize routing overhead.
We use load balancers listening on anycast IP addresses and route requests to the closest available node. The load balancing edge network is powered by multiple regions and provides a consistently fast service all over the globe.
Unlock Solana’s Potential with HA RPC Clusters
If Solana is critical for your business you can’t afford failures. HA RPC clusters provide a rapid, resilient and scalable solution.
HA Clusters are Essential for Institutional Blockchain Operators
Nothing else gives you the uptime and assurance. To keep your Solana services running, use HA Clusters.
Blockdaemon operates thousands of nodes across 40+ blockchain protocols. When it comes to blockchain infrastructure, nobody has more experience.
HA Clusters are backed by our security-first approach. This includes our 99.9% uptime guarantee.