4 ms·

NATS – An open-source, high-performance, lightweight messaging system

59 points by stargrave 11y ago

jedisct1 11y ago
Nats is really neat, and super simple to deploy. For many use cases, the lack of persistence might be a showstopper, though.
- bjflanne 11y ago
  Hi guys I'm on the Community/Ecosystem side of NATS...timely comment. Persistence is actually on the roadmap but we only just discussed end of August at our Bay Area meetup:https://youtu.be/jxpYszUi8JQ https://youtu.be/jxpYszUi8JQ Please feel free to ping me with further questions, comments, ideas etc if you'd like to chat further: brian@apcera.com
  - bjflanne 11y ago
    Also, I should add we've made some significant updates since the article you referenced, and are currently performing additional benchmarking, but some examples: https://twitter.com/derekcollison/status/635541819393245184 https://twitter.com/derekcollison/status/635541819393245184 It would be helpful to understand what you are trying to achieve, and what characteristics in messaging infrastructure are most important to you?
fbomb 11y ago
How does performance compare to 0mq?
- justinsaccount 11y ago
  This is covered on the other page they link to: http://bravenewgeek.com/dissecting-message-queues/ http://bravenewgeek.com/dissecting-message-queues/ "ZeroMQ is capable of sending over 5,000,000 messages per second but is only able to receive about 600,000/second. In contrast, nanomsg sends shy of 3,000,000/second but can receive almost 2,000,000." So.. nanomsg is 10x faster. Why would one use nats?
  - zo1 11y ago
    >"So.. nanomsg is 10x faster. Why would one use nats?" The article comparing the performance of all those messaging systems conveniently "groups" them into two groups. Brokerless, and brokered. Putting Nats into the latter, despite it not sharing any of the main things that we are used to with a brokered system. I.e. Transactions and persistance. So, it's safe to say that in that context and those properties, there is no reason to use NATS over nanomsg.
- tylertreat 11y ago
  Not a fair comparison. ZeroMQ is a socket library. NATS is a message broker.
  - beagle3 11y ago
    A pretty fair comparison, I think. NATS requires a broker, ZeroMQ doesn't (though it generally supports a "forwarding device", which -- in the absence of transactional semantics as it is in NATS -- is mostly indistinguishable from a broker). From the user's perspective, ZeroMQ provides a superset of the functionality, and a superset of the topologies. Why would you say the comparison is unfair?
    - tylertreat 11y ago
      A latency/throughput benchmark comparing the two is effectively comparing the latency/throughput of a socket from A to B. The same benchmark for NATS would be a socket from A to NATS and NATS to B. Obviously the latency of going over the wire once is going to be lower than going over the wire twice. A proper comparison would need to include the forwarding device for ZeroMQ.
      - beagle3 11y ago
        Not at all. It is an internal implementation detail that, in almost all relevant topologies, ZeroMQ can do with just one socket connection but NATS needs two. Furthermore, in LAN pub/sub topologies, ZeroMQ can do with just one copy of the message on the outgoing wire + the occasional retransmit (letting the switch handle the broadcast/multicast) whereas NATS requires a copy for every subscriber. > Obviously the latency of going over the wire once is going to be lower than going over the wire twice. Latencywise, and all things being optimal, yes. But not all things are optimal. Throughputwise, and all things being optimal, there shouldn't be any difference. But there is a huge one on both receive and transmit. NATS and ZeroMQ are obviously not equivalent, but I maintain that this is an apples-to-apples proper and useful comparison, even without the forwarding device.
eikenberry 11y ago
At least once or at most once?
- tylertreat 11y ago
  NATS is at-most-once delivery, which is partly why it's so fast. Sounds like there are plans to introduce stronger guarantees in the future though.
perlgeek 11y ago
I'd like to know more about: * clustering - what kinds of clusters are supported? How do they scale? Do they favor availability or consistency? * security - is there any authentication? authorization for certain resources? TLS? * guarantuees for delivery? Transactional modes? * what does the wire protocol look like? Support for meta data? I clicked around the website and documentation a bit, and didn't find any answers.
- tylertreat 11y ago
  > * clustering - what kinds of clusters are supported? How do they scale? Do they favor availability or consistency? There is no persistence so consistency is irrelevant. NATS does everything in the name of availability to the point of dropping connections for slow consumers. > * security - is there any authentication? authorization for certain resources? TLS? TLS is planned, not sure on timeline. It sounds like Apcera has plans for some form of authentication and multitenancy as well. > * guarantuees for delivery? Transactional modes? At-most-once delivery. No transactions. There may be some form of more reliable delivery coming in the future, but I'm not sure of the details. > * what does the wire protocol look like? Support for meta data? Very simple text-based protocol. I don't think it's formally documented anywhere currently though.
  - vitroth 11y ago
    The NATS philosophy on guaranteed delivery is that delivery isn't what you care about, its the action taken as a result of delivery. If the far end system receives your request but crashes before processing it, was it really received? The business logic of most applications requires end-to-end acknowledgement of transactions, so NATS considers that an application problem. (Disclaimer: I work for Apcera)
jacques_chester 11y ago
NATS is legit -- Derek Collison has experience in messaging systems going back a very long way. Cloud Foundry still has a bunch of NATS-based services because Derek was one of the founders of the first versions, back when Cloud Foundry was a VMWare skunkworks effort (VCAP). Disclaimer: I work for Pivotal Labs, a division of Pivotal, which contributes the largest amount of engineering effort to Cloud Foundry. Apcera is, nominally, a competitor. I should also mention that my understanding of CF's early history is based on accounts I've heard and might be flawed.

59 points by stargrave 11y ago

Nats is really neat, and super simple to deploy. For many use cases, the lack of persistence might be a showstopper, though.

bjflanne 11y ago
Hi guys I'm on the Community/Ecosystem side of NATS...timely comment. Persistence is actually on the roadmap but we only just discussed end of August at our Bay Area meetup:https://youtu.be/jxpYszUi8JQ https://youtu.be/jxpYszUi8JQ Please feel free to ping me with further questions, comments, ideas etc if you'd like to chat further: brian@apcera.com
- bjflanne 11y ago
  Also, I should add we've made some significant updates since the article you referenced, and are currently performing additional benchmarking, but some examples: https://twitter.com/derekcollison/status/635541819393245184 https://twitter.com/derekcollison/status/635541819393245184 It would be helpful to understand what you are trying to achieve, and what characteristics in messaging infrastructure are most important to you?

fbomb 11y ago

How does performance compare to 0mq?

justinsaccount 11y ago
This is covered on the other page they link to: http://bravenewgeek.com/dissecting-message-queues/ http://bravenewgeek.com/dissecting-message-queues/ "ZeroMQ is capable of sending over 5,000,000 messages per second but is only able to receive about 600,000/second. In contrast, nanomsg sends shy of 3,000,000/second but can receive almost 2,000,000." So.. nanomsg is 10x faster. Why would one use nats?
- zo1 11y ago
  >"So.. nanomsg is 10x faster. Why would one use nats?" The article comparing the performance of all those messaging systems conveniently "groups" them into two groups. Brokerless, and brokered. Putting Nats into the latter, despite it not sharing any of the main things that we are used to with a brokered system. I.e. Transactions and persistance. So, it's safe to say that in that context and those properties, there is no reason to use NATS over nanomsg.
tylertreat 11y ago
Not a fair comparison. ZeroMQ is a socket library. NATS is a message broker.
- beagle3 11y ago
  A pretty fair comparison, I think. NATS requires a broker, ZeroMQ doesn't (though it generally supports a "forwarding device", which -- in the absence of transactional semantics as it is in NATS -- is mostly indistinguishable from a broker). From the user's perspective, ZeroMQ provides a superset of the functionality, and a superset of the topologies. Why would you say the comparison is unfair?
  - tylertreat 11y ago
    A latency/throughput benchmark comparing the two is effectively comparing the latency/throughput of a socket from A to B. The same benchmark for NATS would be a socket from A to NATS and NATS to B. Obviously the latency of going over the wire once is going to be lower than going over the wire twice. A proper comparison would need to include the forwarding device for ZeroMQ.
    - beagle3 11y ago
      Not at all. It is an internal implementation detail that, in almost all relevant topologies, ZeroMQ can do with just one socket connection but NATS needs two. Furthermore, in LAN pub/sub topologies, ZeroMQ can do with just one copy of the message on the outgoing wire + the occasional retransmit (letting the switch handle the broadcast/multicast) whereas NATS requires a copy for every subscriber. > Obviously the latency of going over the wire once is going to be lower than going over the wire twice. Latencywise, and all things being optimal, yes. But not all things are optimal. Throughputwise, and all things being optimal, there shouldn't be any difference. But there is a huge one on both receive and transmit. NATS and ZeroMQ are obviously not equivalent, but I maintain that this is an apples-to-apples proper and useful comparison, even without the forwarding device.

eikenberry 11y ago

At least once or at most once?

tylertreat 11y ago
NATS is at-most-once delivery, which is partly why it's so fast. Sounds like there are plans to introduce stronger guarantees in the future though.

perlgeek 11y ago

I'd like to know more about: * clustering - what kinds of clusters are supported? How do they scale? Do they favor availability or consistency? * security - is there any authentication? authorization for certain resources? TLS? * guarantuees for delivery? Transactional modes? * what does the wire protocol look like? Support for meta data? I clicked around the website and documentation a bit, and didn't find any answers.

tylertreat 11y ago
> * clustering - what kinds of clusters are supported? How do they scale? Do they favor availability or consistency? There is no persistence so consistency is irrelevant. NATS does everything in the name of availability to the point of dropping connections for slow consumers. > * security - is there any authentication? authorization for certain resources? TLS? TLS is planned, not sure on timeline. It sounds like Apcera has plans for some form of authentication and multitenancy as well. > * guarantuees for delivery? Transactional modes? At-most-once delivery. No transactions. There may be some form of more reliable delivery coming in the future, but I'm not sure of the details. > * what does the wire protocol look like? Support for meta data? Very simple text-based protocol. I don't think it's formally documented anywhere currently though.
- vitroth 11y ago
  The NATS philosophy on guaranteed delivery is that delivery isn't what you care about, its the action taken as a result of delivery. If the far end system receives your request but crashes before processing it, was it really received? The business logic of most applications requires end-to-end acknowledgement of transactions, so NATS considers that an application problem. (Disclaimer: I work for Apcera)

jacques_chester 11y ago

NATS is legit -- Derek Collison has experience in messaging systems going back a very long way. Cloud Foundry still has a bunch of NATS-based services because Derek was one of the founders of the first versions, back when Cloud Foundry was a VMWare skunkworks effort (VCAP). Disclaimer: I work for Pivotal Labs, a division of Pivotal, which contributes the largest amount of engineering effort to Cloud Foundry. Apcera is, nominally, a competitor. I should also mention that my understanding of CF's early history is based on accounts I've heard and might be flawed.