Your best investment in 2015 could turn out to be litecoin.Litecoin price has increased from $1.60 to over $4.00 in just thirty days.Take a look at this logarithmic scale weekly chart of litecoin price in US dollars:The price of litecoin broke up through the long term trendline in early May and has continued to rise with conviction.Bitcoin has also broken the weekly trendline but is lagging litecoin instead of leading for a change:The amount of litecoin mined will be decreased by half on 25 August 2015 at the first litecoin block reward halving.After that date litecoin miners will only be rewarded with 25 litecoin for each block discovered instead of 50.When the first block reward halving occurred for bitcoin in Novenber 2012 bitcoin price increased from under $15 to over $250 in five months.The current rise in the price of litecoin is speculation by traders that litecoin will increase dramatically in price after the block reward halving just as bitcoin did.Over the past year the price of litecoin was driven down as scrypt ASIC miners began shipping and the supply of litecoin increased sharply.

The rate of litecoin being mined is just about to be cut in half leading to the price being bid up.If bitcoin is gold, litecoin has always been considered silver with a similarly lower price.Litecoin is the third place digital coin in market capitalization after Ripple.Compared to bitcoin, litecoin has faster transaction times and less turmoil in the community over block size because the litecoin network can handle four times the transactions per second than bitcoin.Any change in the litecoin code to handle increased transaction volume is years away.There is no easy way at present to buy litecoin with fiat; Coinbase does not offer litecoin for sale.If you have bitcoin you can exchange them for litecoin at any exchange like BTC-e.Probably the most convenient way to acquire litecoin is to just swap bitcoin or other altcoins you already own for litecoin at ShapeShift.When to buy?If you think a new bull market in litecoin price has begun you could buy in using dollar cost averaging.Another method to get a good entry price is to follow daily price of litecoin and wait for a down day on volume, preferably with the Stochastic RSI indicator in oversold region as in late June:The blue line on the chart is the 200 day simple moving average.

With litecoin price well above the 200 day SMA a new bull market in litecoin price would seem to be confirmed.You might prefer the Electrum for Litecoin wallet which can be downloaded at the same link.It is faster and easier to use than Litecoin Core.You can also use Electrum for Litecoin with a Trezor for secure cold storage of your litecoin.Once you have a litecoin wallet you can easily swap bitcoin for litecoin at ShapeShift and send to your litecoin wallet for storage.Recently there has been a race towards lower block intervals for PoW block-chain based  cryptocurrencies.First there was Bitcoin with a 10 minute interval, then was LiteCoin using a 2.5 interval, then was DogeCoin with 1 minute, and then QuarkCoin with just 30 seconds.This reminds me a joke in a comedy movie where there were two guys in a car, and one says that he had a fantastic idea for a top-selling product: a video called “Get fit by exercising 10 minutes a day”.The other replies that there is already a video teaching how to get fit by exercising 15 minutes a day.

The first explains that that’s why his own video will be a top-seller: you get the same result in less exercise time.After thinking about it a minute, the second claims he has a tremendous idea, a video teaching to get fit by exercising 5 minutes a day.But the first one then shouts terribly outraged: that’s impossible!pyopencl bitcoin miner[1] This seems similar with what is happening with confirmation intervals: every new cryptocurrency lowers it a little bit, but no one actually knows what are the implications of doing so.hacker currency bitcoin crashesTo understand how the block interval impacts the stability and capability of the cryptocurrency network, several factors must be taken into account.litecoin server mining

First of all, the most important factor that affects the viability of short confirmation intervals is the number of stale blocks generated.Two factors mainly affect the stale block rate: the propagation protocol and the block propagation time from the top miners to the top miners.litecoin feb 2014We’ll analyze in the next section these factors and some measures that can be taken to optimize them.ethereum jobs londonThe block propagation protocol: Bitcoin forwards a block by packing the block header with all the transactions contained in the block.bitcoin karte kaufenThis strategy, while being the most easy to analyze, is known to perform badly, both regarding block propagation latency and bandwidth usage.earn litecoin fast

Since the transactions on a block are generally already known to the network, there is no benefit in transmitting them again.It’s possible that a peer reconstructs a block by being sent only the header and the list of transactions hashes.The peer then decides which transactions missing in the transaction pool need to be fetched.bitcoin crime sentenceBut since each node stores the hashes of the transactions advertised by its peers, an even better strategy is that a miner also sends immediately the transactions included in the block that he knows are missing in the peer’s pool.tutorial moon bitcoinThis eliminates completely the need of a second interaction to request additional transactions.Still another optimization is that miners only include transactions that have been received before a few of seconds ago.This assures with high probability that transactions will have already been received by peers before the block is mined.

Still another improvement is to forward the block header before even checking the transactions, and only checking the block PoW and height at forward time.This allows the header to spread over the network in less than a second.Nodes can then start mining an empty block on top of the header even if the transactions are still missing, during a fixed interval.After that interval, they resume mining in whatever block they were mining before.These empty blocks reduce the effective bandwidth and block-chain storage usage.But this is generally unimportant since empty block are, by definition, small.But if the block interval is reduced beyond 1 second, then the time a transaction is included in a block begins to increase, since most of the blocks produced will be empty.Some parts of this protocol, and how to modify Bitcoin to support them, was proposed here.The type of nodes in the network: Because what matters most is inter-miner connection latency, having other kind of nodes in the network (such as users,  or activity monitors) only increase the number of stale blocks.

A cryptocurrency network could be optimized by forcing that all participants to be miners, or prioritizing miners over other nodes.This can be done, as in mining pools, by incetivizing nodes to broadcast quasi-solved block headers (blocks with a PoW that is lower than the target difficulty, but still are difficult to find).These quasi-blocks can be used by nodes to increase the priority of peers that send the highest number of quasi-solved blocks.The block reward scheme: If the block reward decreases over time until zero (as in Bitcoin), then the miner’s revenue will, on the long run, depend solely on transaction fees.This means that creating a block with no transactions is useless for the miner.But as previously stated, maximally reducing the block interval comes with a price: empty blocks may need to be created for headers with still unknown transactions.That’s why a cryptocurrency which maintains a reward forever on blocks can achieve a lower stale block rate.The incentivized topology of the miners network: a highly concentrated network, with big mining pools, tend to generate much less state blocks than a complete distributed mining topology.

In this regard cryptocoins based on SHA-256 PoW have an advantage over non ASIC-friendly PoW based crytopcoins, since mining requires an investment in ASIC boards.The real topology of the network: Bitcoin design assumes the network is similar to a random graph, having a certain average out-degree and in-degree.While this is far from true in reality,  network nodes take local decisions to avoid forming geographical clusters (at least for the out-bound connections).This is not the best topology to help block propagation.The best topology for block propagation is one that serves the top miners better, by encouraging direct connections between them or by routing blocks faster between them.Also a direct miner-to-miner backbone can help to decrease notably the number of stale blocks.This has been proposed for Bitcoin to increase resilience from attacks.The PoW function verification time: SHA-256 is  very fast to evaluate and so the Bitcoin PoW verification time is negligible.An scrypt PoW, on the contrary, may take from 3 to 30 milliseconds to evaluate depending on the parameters chosen (GPU or ASIC prevention).

To protect the network from spamming and DoS attacks, each node needs to verify the block PoW before forwarding the block again, so the verification delay gets multiplied by the number of hops in the block path between miners.It’s nevertheless possible to forward the block without checking the block PoW if the block (or the message containing the block) carries a small fast-to-verify PoW, which is verified by nodes and “pays” for the forwarding.Then, a scrypt-based coin can also benefit from the fastest possible forwarding protocols.Client networking stack: Once a node receives a block, the best it can do to reduce the appearance of stale blocks is forward it as soon as possible.This means that all other node activity should be paused or stopped.Some operations can be designed to allow immediate cancellation and accept retrials.To allow immediate forwarding, the client networking stack must not block the client in transaction verification procedures or other housekeeping activities, such as chain reorganizations.

This can be best achieved by a multi-threading client, and dynamically setting thread priorities to boost the thread that has received the block.Even a better performance could be achieved by using a kernel level device driver that takes care of block header packet forwarding even before the application receives the notification.The block overhead: Block headers in most cryptocurrencies are small (80 bytes) so the header size (over the block size) does not pose a significant overhead.But the block must also contain the generation transaction, which may account for other 250 bytes.Anyway, the block overhead generally does not pose a problem.Nevertheless, when block interval approaches  sub-second times, it reduces the channel bandwidth considerably.Since some factors affect other factors, it’s difficult to construct a model that allows to determine the transaction confirmation time with a closed-formula.Instead, I attempted to build a network simulator to play with different factors and algorithms.

There are two strategies for simulation block propagation using a discrete event simulation.A first strategy is to simulate the whole network, node by node, assuming certain graph topology.This is challenging, because the topology of the network is unknown.Many nodes do not accept incoming connections so they cannot be probed and are invisible to network crawlers.Some sites like bitnodes  attempt to measure the network, but their results are only an approximation.Also, simulating 60K nodes for several days can be very CPU consuming.A second strategy is to simulate the interaction between two top-miners, each one in another end of the graph (the hop distance between them being near the average distance between nodes).This is not the worse case, where miners are located a hop-distance equal to the diameter of the graph, but since its the best interest for top-miners to be well-connected, we can assume they perform not worse than the average.This turns to be very easy to simulate, the only simulated events being the creation of a block in one of the two locations and the propagation of the block to the opposite location.

Each location stores a best chain and a competing chain.The following results show the simulation of Bitcoin, LiteCoin, and FastCoin5.FastCoin5 simulation represents the use of the best parameters to reduce stale blocks for a 5 block interval.The chosen optimizations for FastCoin5 are: – Block push heuristic, SHA.256 based PoW.– Transmission of blocks with transaction hash list, but not full transactions.Pushing missing transaction in a block to peers.– Mining of empty blocks while parent block transactions are still being fetched (with a maximum wait of 10 seconds) – Node route prioritizing.Nodes stop whatever they are doing when they receive a block to forward it as quickly as possible.– Multi-threaded, non-locking client networking stack.The fields “avg conf time” estimate the confirmation time in seconds for the reversal probability shown just before the field, as a percentage (total blocks lost multiplied by 100 over total blocks created).The average time to confirm a block with a 0.1% probability of being reverted because of a chain reorganization is, for each system: 25 minutes for Bitcoin, 11.25 minutes for LiteCoin and 25 seconds for FastCoin5.

The only drawback of FastCoin5 seems to be that the channel usage due to empty blocks is only 62.4%, but since transactions are transmitted only once instead of twice as in Bitcoin, FastCoin5 effectively achieves a higher channel utilization than Bitcoin (currently below 50%).It must be noted that the client application need not to internally switch the best chain before 3 confirmations have been received for any candidate chain, since the user should not be notified if  a transaction has less than 3 confirmations.This eliminates all computing bottle-neck in normal clients.Miners, on the contrary, must verify the transactions of each block to decide which chain to mine on top of.It’s interesting that a merchant may target a different number of confirmation blocks for services or products that are differently priced.For example, if the merchant is willing to accept to loose 1 USD/day and prefers smaller confirmation times, and sells 100 products a day which are priced under 1 USD each, he can choose to accept only 3 confirmation blocks for these products, to achieve a 1% reversal probability, which leads to a 20 second confirmation time.

It would be interesting to implement a client that can show the confirmation status depending on the transaction value, and a pre-defined minimal expected loss.It must be taken into account that reversal probabilities shown are based on the fact that miners will forget to include the transactions taken from the stale blocks in the newly created blocks, and that competing blocks transaction sets do not intersect.This is a simulation defect that increases the confirmation times.This is not the case in the Satoshi client, where all transactions that belong to the stale blocks are re-inserted into the transaction pool, and competing blocks share most transactions.A transaction with enough fees may have actually a reversal probability much higher than the value presented in this post, and the FastConf5 confirmation time could actually be much lower.In this post we’ve analyzed the factors that influence the stale block ratio and the expected confirmation time for a certain confirmation-reversal probability.