Related
I don't have much experience with network programming, but an interesting problem came up that requires it. The server will be transmitting multiple streams of different types of data to other machines. Each machine should be able choose which of the streams (one or more) it will like to receive. The whole setup is confined to the local network only. Initially, there will be only two clients, but I would like to design a scalable approach, if possible.
The existing server code, which is streaming only a single stream, is using TCP streaming socket for doing so. However, from some reading on the subject, I am not sure if this approach will scale to multiple streams and multiple clients well. The reason is: wouldn't two clients, who want to receive the same stream but connect via different TCP sockets, result in wastage of bandwidth? Especially compared to UDP, which allows to multicast.
Due to my inexperience, I am relying on better informed people out there to advise me: considering that i do want the stream to be reliable, would it be worth it to start from the scratch with UDP, and implement reliability into it, than to keep using TCP? Or, will this be better solved by designing an appropriate network structure? I'd be happy to provide more details if needed. Thanks.
UPDATE: I am looking at PGM and emcaster for reliable multicasting at the moment. Must have C# implementations at server side, and python implementations at client side.
Since you want a scalable program, then UDP would be a better choice, because it does not go the extra length to verify that the data has been received, thus making the process of sending data faster.
I am to design a server that needs to serve millions of clients that are simultaneously connected with the server via TCP.
The data traffic between the server and the clients will be sparse, so bandwidth issues can be ignored.
One important requirement is that whenever the server needs to send data to any client it should use the existing TCP connection instead of opening a new connection toward the client (because the client may be behind a firewall).
Does anybody know how to do this, and what hardware/software is needed (at the least cost)?
What operating systems are you considering for this?
If using a Windows OS and using something later than Vista then you shouldn't have a problem with many thousands of connections on a single machine. I've run tests (here: http://www.lenholgate.com/blog/2005/11/windows-tcpip-server-performance.html) with a low spec Windows Server 2003 machine and easily achieved more than 70,000 active TCP connections. Some of the resource limits that affect the number of connections possible have been lifted considerably on Vista (see here: http://www.lenholgate.com/blog/2005/11/windows-tcpip-server-performance.html) and so you could probably achieve your goal with a small cluster of machines. I don't know what you'd need in front of those to route the connections.
Windows provides a facility called I/O Completion Ports (see: http://msdn.microsoft.com/en-us/magazine/cc302334.aspx) which allow you to service many thousands of concurrent connections with very few threads (I was running tests yesterday with 5000 connections saturating a link to a server with 2 threads to process the I/O...). Thus the basic architecture is very scalable.
If you want to run some tests then I have some freely available tools on my blog that allow you to thrash a simple echo server using many thousands of connections (1) and (2) and some free code which you could use to get you started (3)
The second part of your question, from your comments, is more tricky. If the client's IP address keeps changing and there's nothing between you and them that is providing NAT to give you a consistent IP address then their connections will, no doubt, be terminated and need to be re-established. If the clients detect this connection tear down when their IP address changes then they can reconnect to the server, if they can't then I would suggest that the clients need to poll the server every so often so that they can detect the connection loss and reconnect. There's nothing the server can do here as it can't predict the new IP address and it will discover that the old connection has failed when it tries to send data.
And remember, your problems are only just beginning once you get your system to scale to this level...
This problem is related to the so-called C10K problem. The C10K page lists a large number of good resources for addressing the problems you will encounter when you try to allow thousands of clients to connect to the same server.
I've come across the APE Project
a while back. It seems like a dream come true. They can support up to 100k concurrent clients on a single node. Spread them across 10 or 20 nodes, and you can serve millions. Perfect for RESTful applications. Might want to look deeper for any shared namespace. One drawback is that this is a standalone server, as in supplementary to a web server. This server is of course Open Source, so any cost is hardware/ISP related.
You cannot use UDP. If the client sends a request and you don't reply immediately, a router is going to forget the reverse route in 30 seconds or less, so your server will never be able to reply to the client.
TCP is the only option, and it, too, will give you headaches. Most routers are going to forget the route and/or drop the connection after a few minutes, so your client/server code is going to have to send "keep alives" fairly often.
I recommend setting up a "sniffer", to see how the phone companies are staying in touch with your smartphone for their "push" technology. Copy whatever they're doing, because that stuff works!
As Greg mentioned, the problem you are describing is C10K (or rather "C1M" in your case )
I recently made a simple TCP echo server on linux that scales very well with the number of sessions (only tested up to 200.000 though), by using the epoll queue. On BSD, you have something similar called kqueue.
You can check out the code if you want to. Hope this helps and good luck!
EDIT: As noted in the comments below, my original assertion that there is a 64K limit based on the number of ports is incorrect, however there is a 32K limit on the number of socket handles, so my suggested design is valid.
With a typical TCP/IP server design, you're limited in the number of simultaneous open connections you can have. The server has one listening port, and when a client connects to it the server makes an accept call, and that creates a new socket on a random port for the rest of the connection.
To handle more than 64K simultaneous connections I think you need to use UDP instead. You only need one port for the server to listen on, and you need to manage the connections using a 32-bit client ID in the packet data instead of having a separate port for each client. The 32-bit client ID could be the client's IP address, and the client can listen on a known UDP port for messages coming back from the server. That port would be the only one that needs to be open on the firewall.
With this approach, your only limitation is how quickly you can handle and respond to UDP messages. With millions of clients, even sparse traffic could give you large spikes, and if you don't read the packets fast enough your input queue will fill up and you'll start dropping packets. The C10K page Greg points to will give you strategies for that.
Imagine you have many clustered servers, across many hosts, in a heterogeneous network environment, such that the connections between servers may have wildly varying latencies and bandwidth. You want to build a map of the connections between servers by transferring data between them.
Of course, this map may become stale over time as the network topology changes - but lets ignore those complexities for now and assume the network is relatively static.
Given the latencies between nodes in this host graph, calculating the bandwidth is a relative simply timing exercise. I'm having more difficulty with the latencies - however. To get round-trip time, it is a simple matter of timing a return-trip ping from the local host to a remote host - both timing events (start, stop) occur on the local host.
What if I want one-way times under the assumption that the latency is not equal in both directions? Assuming that the clocks on the various hosts are not precisely synchronized (at least that their error is of the the same magnitude as the latencies involved) - how can I calculate the one-way latency?
In a related question - is this asymmetric latency (where a link is quicker in direction than the other) common in practice? For what reasons/hardware configurations? Certainly I'm aware of asymmetric bandwidth scenarios, especially on last-mile consumer links such as DSL and Cable, but I'm not so sure about latency.
Added: After considering the comment below, the second portion of the question is probably better off on serverfault.
To the best of my knowledge, asymmetric latencies -- especially "last mile" asymmetries -- cannot be automatically determined, because any network time synchronization protocol is equally affected by the same asymmetry, so you don't have a point of reference from which to evaluate the asymmetry.
If each endpoint had, for example, its own GPS clock, then you'd have a reference point to work from.
In Fast Measurement of LogP Parameters
for Message Passing Platforms, the authors note that latency measurement requires clock synchronization external to the system being measured. (Boldface emphasis mine, italics in original text.)
Asymmetric latency can only be measured by sending a message with a timestamp ts, and letting the receiver derive the latency from tr - ts, where tr is the receive time. This requires clock synchronization between sender and receiver. Without external clock synchronization (like using GPS receivers or specialized software like the network time protocol, NTP), clocks can only be synchronized up to a granularity of the roundtrip time between two hosts [10], which is useless for measuring network latency.
No network-based algorithm (such as NTP) will eliminate last-mile link issues, though, since every input to the algorithm will itself be uniformly subject to the performance characteristics of the last-mile link and is therefore not "external" in the sense given above. (I'm confident it's possible to construct a proof, but I don't have time to construct one right now.)
There is a project called One-Way Ping (OWAMP) specifically to solve this issue. Activity can be seen in the LKML for adding high resolution timestamps to incoming packets (SO_TIMESTAMP, SO_TIMESTAMPNS, etc) to assist in the calculation of this statistic.
http://www.internet2.edu/performance/owamp/
There's even a Java version:
http://www.av.it.pt/jowamp/
Note that packet timestamping really needs hardware support and many present generation NICs only offer millisecond resolution which may be out-of-sync with the host clock. There are MSDN articles in the DDK about synchronizing host & NIC clocks demonstrating potential problems. Timestamps in nanoseconds from the TSC is problematic due to core differences and may require Nehalem architecture to properly work at required resolutions.
http://msdn.microsoft.com/en-us/library/ff552492(v=VS.85).aspx
You can measure asymmetric latency on link by sending different sized packets to a port that returns a fixed size packet, like send some udp packets to a port that replies with an icmp error message. The icmp error message is always the same size, but you can adjust the size of the udp packet you're sending.
see http://www.cs.columbia.edu/techreports/cucs-009-99.pdf
In absence of a synchronized clock, the asymmetry cannot be measured as proven in the 2011 paper "Fundamental limits on synchronizing clocks over networks".
https://www.researchgate.net/publication/224183858_Fundamental_Limits_on_Synchronizing_Clocks_Over_Networks
The sping tool is a new development in this space, which uses clock synchronization against nearby NTP servers, or an even more accurate source in the form of a GNSS box, to estimate asymmetric latencies.
The approach is covered in more detail in this blog post.
Say if I was to get shared, virtual or dedicated hosting, I read somewhere a server/machine can only handle 64,000 TCP connections at one time, is this true? How many could any type of hosting handle regardless of bandwidth? I'm assuming HTTP works over TCP.
Would this mean only 64,000 users could connect to the website, and if I wanted to serve more I'd have to move to a web farm?
In short:
You should be able to achieve in the order of millions of simultaneous active TCP connections and by extension HTTP request(s). This tells you the maximum performance you can expect with the right platform with the right configuration.
Today, I was worried whether IIS with ASP.NET would support in the order of 100 concurrent connections (look at my update, expect ~10k responses per second on older ASP.Net Mono versions). When I saw this question/answers, I couldn't resist answering myself, many answers to the question here are completely incorrect.
Best Case
The answer to this question must only concern itself with the simplest server configuration to decouple from the countless variables and configurations possible downstream.
So consider the following scenario for my answer:
No traffic on the TCP sessions, except for keep-alive packets (otherwise you would obviously need a corresponding amount of network bandwidth and other computer resources)
Software designed to use asynchronous sockets and programming, rather than a hardware thread per request from a pool. (ie. IIS, Node.js, Nginx... webserver [but not Apache] with async designed application software)
Good performance/dollar CPU / Ram. Today, arbitrarily, let's say i7 (4 core) with 8GB of RAM.
A good firewall/router to match.
No virtual limit/governor - ie. Linux somaxconn, IIS web.config...
No dependency on other slower hardware - no reading from harddisk, because it would be the lowest common denominator and bottleneck, not network IO.
Detailed Answer
Synchronous thread-bound designs tend to be the worst performing relative to Asynchronous IO implementations.
WhatsApp can handle a million WITH traffic on a single Unix flavoured OS machine - https://blog.whatsapp.com/index.php/2012/01/1-million-is-so-2011/.
And finally, this one, http://highscalability.com/blog/2013/5/13/the-secret-to-10-million-concurrent-connections-the-kernel-i.html, goes into a lot of detail, exploring how even 10 million could be achieved. Servers often have hardware TCP offload engines, ASICs designed for this specific role more efficiently than a general purpose CPU.
Good software design choices
Asynchronous IO design will differ across Operating Systems and Programming platforms. Node.js was designed with asynchronous in mind. You should use Promises at least, and when ECMAScript 7 comes along, async/await. C#/.Net already has full asynchronous support like node.js. Whatever the OS and platform, asynchronous should be expected to perform very well. And whatever language you choose, look for the keyword "asynchronous", most modern languages will have some support, even if it's an add-on of some sort.
To WebFarm?
Whatever the limit is for your particular situation, yes a web-farm is one good solution to scaling. There are many architectures for achieving this. One is using a load balancer (hosting providers can offer these, but even these have a limit, along with bandwidth ceiling), but I don't favour this option. For Single Page Applications with long-running connections, I prefer to instead have an open list of servers which the client application will choose from randomly at startup and reuse over the lifetime of the application. This removes the single point of failure (load balancer) and enables scaling through multiple data centres and therefore much more bandwidth.
Busting a myth - 64K ports
To address the question component regarding "64,000", this is a misconception. A server can connect to many more than 65535 clients. See https://networkengineering.stackexchange.com/questions/48283/is-a-tcp-server-limited-to-65535-clients/48284
By the way, Http.sys on Windows permits multiple applications to share the same server port under the HTTP URL schema. They each register a separate domain binding, but there is ultimately a single server application proxying the requests to the correct applications.
Update 2019-05-30
Here is an up to date comparison of the fastest HTTP libraries - https://www.techempower.com/benchmarks/#section=data-r16&hw=ph&test=plaintext
Test date: 2018-06-06
Hardware used: Dell R440 Xeon Gold + 10 GbE
The leader has ~7M plaintext reponses per second (responses not connections)
The second one Fasthttp for golang advertises 1.5M concurrent connections - see https://github.com/valyala/fasthttp
The leading languages are Rust, Go, C++, Java, C, and even C# ranks at 11 (6.9M per second). Scala and Clojure rank further down. Python ranks at 29th at 2.7M per second.
At the bottom of the list, I note laravel and cakephp, rails, aspnet-mono-ngx, symfony, zend. All below 10k per second. Note, most of these frameworks are build for dynamic pages and quite old, there may be newer variants that feature higher up in the list.
Remember this is HTTP plaintext, not for the Websocket specialty: many people coming here will likely be interested in concurrent connections for websocket.
This question is a fairly difficult one. There is no real software limitation on the number of active connections a machine can have, though some OS's are more limited than others. The problem becomes one of resources. For example, let's say a single machine wants to support 64,000 simultaneous connections. If the server uses 1MB of RAM per connection, it would need 64GB of RAM. If each client needs to read a file, the disk or storage array access load becomes much larger than those devices can handle. If a server needs to fork one process per connection then the OS will spend the majority of its time context switching or starving processes for CPU time.
The C10K problem page has a very good discussion of this issue.
To add my two cents to the conversation a process can have simultaneously open a number of sockets connected equal to this number (in Linux type sytems) /proc/sys/net/core/somaxconn
cat /proc/sys/net/core/somaxconn
This number can be modified on the fly (only by root user of course)
echo 1024 > /proc/sys/net/core/somaxconn
But entirely depends on the server process, the hardware of the machine and the network, the real number of sockets that can be connected before crashing the system
It looks like the answer is at least 12 million if you have a beefy server, your server software is optimized for it, you have enough clients. If you test from one client to one server, the number of port numbers on the client will be one of the obvious resource limits (Each TCP connection is defined by the unique combination of IP and port number at the source and destination).
(You need to run multiple clients as otherwise you hit the 64K limit on port numbers first)
When it comes down to it, this is a classic example of the witticism that "the difference between theory and practise is much larger in practise than in theory" - in practise achieving the higher numbers seems to be a cycle of a. propose specific configuration/architecture/code changes, b. test it till you hit a limit, c. Have I finished? If not then d. work out what was the limiting factor, e. go back to step a (rinse and repeat).
Here is an example with 2 million TCP connections onto a beefy box (128GB RAM and 40 cores) running Phoenix http://www.phoenixframework.org/blog/the-road-to-2-million-websocket-connections - they ended up needing 50 or so reasonably significant servers just to provide the client load (their initial smaller clients maxed out to early, eg "maxed our 4core/15gb box # 450k clients").
Here is another reference for go this time at 10 million: http://goroutines.com/10m.
This appears to be java based and 12 million connections: https://mrotaru.wordpress.com/2013/06/20/12-million-concurrent-connections-with-migratorydata-websocket-server/
Note that HTTP doesn't typically keep TCP connections open for any longer than it takes to transmit the page to the client; and it usually takes much more time for the user to read a web page than it takes to download the page... while the user is viewing the page, he adds no load to the server at all.
So the number of people that can be simultaneously viewing your web site is much larger than the number of TCP connections that it can simultaneously serve.
in case of the IPv4 protocol, the server with one IP address that listens on one port only can handle 2^32 IP addresses x 2^16 ports so 2^48 unique sockets. If you speak about a server as a physical machine, and you are able to utilize all 2^16 ports, then there could be maximum of 2^48 x 2^16 = 2^64 unique TCP/IP sockets for one IP address. Please note that some ports are reserved for the OS, so this number will be lower. To sum up:
1 IP and 1 port --> 2^48 sockets
1 IP and all ports --> 2^64 sockets
all unique IPv4 sockets in the universe --> 2^96 sockets
There are two different discussions here: One is how many people can connect to your server. This one has been answered adequately by others, so I won't go into that.
Other is how many ports yours server can listen on? I believe this is where the 64K number came from. Actually, TCP protocol uses a 16-bit identifier for a port, which translates to 65536 (a bit more than 64K). This means that you can have that many different "listeners" on the server per IP Address.
I think that the number of concurrent socket connections one web server can handle largely depends on the amount of resources each connection consumes and the amount of total resource available on the server barring any other web server resource limiting configuration.
To illustrate, if every socket connection consumed 1MB of server resource and the server has 16GB of RAM available (theoretically) this would mean it would only be able to handle (16GB / 1MB) concurrent connections. I think it's as simple as that... REALLY!
So regardless of how the web server handles connections, every connection will ultimately consume some resource.
I am considering using SCTP instead of TCP for a p2p app written in C. Should I do it? Also how does the speed of SCTP compare to the speed of TCP?
EDIT:
I found that SCTP can be tunneled over UDP with the only problem being tunneled SCTP is not interoperable with untunneled SCTP.
Have you considered whether your target systems will all have SCTP pre-installed on them or whether your application will need to include SCTP itself? In my experience I would not expect all systems to have SCTP installed on them, and I would expect them not to if it were Windows.
If you include SCTP in the application itself then that will more than double the number of messages being passed into an out of the Kernel which will impact performance when compared with using the pre installed TCP.
Have you considered what benefits you want from SCTP? You mentioned fault tolerance but for this to work with SCTP it requires the application to have multiple ethernet ports and and IP addresses. Is this likely on your app?
As much as I love SCTP (!) I would seriously consider sticking with TCP unless you are sure SCTP is needed or unless you control the hosts your app is deployed on.
Regards
If it's for a local area network, sure go for it.
Note however that if you plan to use it on the open internet many consumer grade firewalls aren't flexible enough to permit unrecognised IP protocols through them.
How does it help you?
You're P2P, so every peer must have at least one socket open to every other peer.
If you've got a socket open, then you can do everything you need to do over that. If you've taken the approach of one socket per file and you have multiple files being tranferred concurrently between two given peers, then SCTP will save you one socket per file. However, on a normal P2P network of any size, you will almost never have multiple files being transferred concurrently between two peers.
Just have one socket and have your own little protocol; send a packet with a header, the header indicates content type, e.g. a command, or part a file - and if so, which file, and which byte range.
Of course, you get a little overhead for that, whereas if you have one socket for commands and one per file, you're more efficient. Is saving one socket per peer (assuming one download at a time) worth the time/hassle/complexity of using SCTP?