7 min to complete
An overview and introduction. You’ll also see how to start a ScyllaDB cloud cluster, which you can use for the hands-on lab. It’s possible to use the free-trial option.
so Silver Cloud you have actually several ways how to use Sila the most convenient one of course is to get someone else started for you which is basically what Silver Cloud is about and Sera cloud is a service that we are providing as a company and it’s a fully managed service and right now we support fully two Cloud providers which is AWS and gcp and there are several ways how to run the Clusters there in AWS we support the running the cluster under our account or if you want to leverage your account which has eventually some I don’t know savings or some discounts you can just use that account as well uh in the Google Cloud there is only the option to have it under our account and uh of course you have or you might have heard that Siva supports various apis we are mainly focused around being compatible with Cassandra and being fully secure compatible but we added the some time ago already dynamodb API compatibility as well so you can choose from those two apis which one you want to use and of course then you can pick the region where you want to host your cluster so it’s as close as possible to your clients and then you can choose the size of the cluster and the instance type I’ll probably explain it on the next slide so this is a slide about instances in AWS so there are some preset instances that we support and we have the best practice with them they generally use local storage to get the fastest latencies from Reading from the disk and the important part there is that the architecture of the Silver Cloud or the service that we are managing for you is to be highly available because of that you usually get your cluster in the end tables which in case is basically three so it’s either three nodes six notes or nine nodes and the reason is very simple because you have like different availability zones and you have usually or you can have three copies of your data which is like the best practice this is the replication Factor the RF which is on the slides and this application Factor uh gives like the availability of the data and uh you basically have three copies and if for example one copy goes down you can still basically work with your data and your application shouldn’t really feel any any problem or any any discomfort uh that’s it that’s it sometimes it’s very hard to figure out what’s the best cluster for your workload and we can help you with that you can get something from Silver there are even blogs about the sizing so you can read the blocks if you don’t want to talk to anyone and we want to figure out for yourself and generally it’s around how much data you have and how many operations per second you want to use and uh what are the types of operations of course if you treat or rights and how big is the payload size uh all that said you can go and run your cluster for free for some small instances so in AWS the smaller instance is P3 micro and this is what I’m going to show today during the demo in gcp the situation is slightly different so again there are different combinations of the machines uh the price is uh vary a little bit as well so you can compare with what which one you want I think gcp is slightly cheaper for some combinations and the rule of a thumb is the same here so again get help with sizing if you don’t dare to do it yourself or try to read the block how to basically size it based on your uh input parameters and in the gcp I think we have a better offer there is an input High mem uh eight cluster which is definitely faster than the T3 micro I will be showing today and again there are various combinations uh I I think what is better on adverse is that it has a higher storage density gcp I think has smaller storage density but then it is I think faster networks so they’re always like pros and cons you can compare them however if you don’t want to use uh any of those prepared managed services that we are giving you you can just deploy the cluster by yourself it’s not that hard to do you can the fastest for developer uh in my opinion is to get like a Docker so you can just go to the link which is on the first part of the slide and get get Docker instance set up or you can even get a cluster off of Docker instances set up it’s slightly harder of your network but it can be figured out from the documentation and if you don’t like that either you can just go and create on your own Hardware cluster we are providing the binaries you can download them and you can install slab by yourself on whatever system you want to use we even support next so if you want to have a cluster on Max you can start a small cluster on Max as well and uh you can of course like use it use it on other clouds and when you will install them I would suggest you guys to use like some supported image so we don’t use something exotic or something that’s supported by the cloud provider because then your performance will suffer that said if you want to basically manage your costs yourself we make it easy for you as well in this case if you will pick one of those big uh supported calls gcp or AWS you can just take our image and you will get a pre-tuned image that is basically tested and very well tested like civil QA and we even have like an Azure Enterprise image too