In blockchain networks, transaction latency is crucial for determining the quality of service (QoS). The latency of a transaction is measured as the time between its issuance and its inclusion in a block in the chain. A block proposer often prioritizes transactions with higher fees or transactions from accounts it is associated with, to minimize their latencies. To maintain fairness among transactions, a block proposer is expected to select the included transactions randomly. The random selection might cause some transactions to experience high latency following the variance in the time a transaction waits until it is selected. We suggest an alternative, age-aware approach towards fairness so that transaction priority is increased upon observing a large waiting time. We explain that a challenge with this approach is that the age of a transaction is not absolute due to transaction propagation. Moreover, a node might present its transactions as older to obtain priority. We describe a new technique to enforce a fair block selection while prioritizing transactions that observed high latency. The technique is based on various declaration schemes in which a node declares its pending transactions, providing the ability to validate transaction age. By evaluating the solutions on Ethereum data and synthetic data of various scenarios, we demonstrate the advantages of the approach under realistic conditions and understand its potential impact to maintain fairness and reduce tail latency.

Data preparation is the first and a very important step towards any Large Language Model (LLM) development. This paper introduces an easy-to-use, extensible, and scale-flexible open-source data preparation toolkit called Data Prep Kit (DPK). DPK is architected and designed to enable users to scale their data preparation to their needs. With DPK they can prepare data on a local machine or effortlessly scale to run on a cluster with thousands of CPU Cores. DPK comes with a highly scalable, yet extensible set of modules that transform natural language and code data. If the user needs additional transforms, they can be easily developed using extensive DPK support for transform creation. These modules can be used independently or pipelined to perform a series of operations. In this paper, we describe DPK architecture and show its performance from a small scale to a very large number of CPUs. The modules from DPK have been used for the preparation of Granite Models [1] [2]. We believe DPK is a valuable contribution to the AI community to easily prepare data to enhance the performance of their LLM models or to fine-tune models with Retrieval-Augmented Generation (RAG).

Blockchain leader-based protocols elect leaders for proposing the next block of transactions. Proposed blocks need to pass a validation routine in order to be added to the blockchain. Proposers may prioritize certain transactions based on their fees or accounts, which enables attackers to gain profits through block building, while simultaneously causing negative impacts on other users. A fair block selection follows a random selection of pending transactions among those that a proposer is aware of. We propose CFTO, a protocol that aims at encouraging fair block selection in a leader-based blockchain network while taking into account real network conditions, such as the network’s topology structure and the forwarding protocol. CFTO offers two main contributions. First, it provides incentives for acting honestly and diminishing malicious and dishonest nodes. To accomplish this, we use a reputation system, whereby each node is given a reputation score based on its actions. Second, it consists of an algorithm that evaluates the proposed blocks based on the zone structure of the network. Furthermore, we adapt the evaluation algorithm to fit the additional order constraints implied in Ethereum transaction ordering. We demonstrate the improved accuracy of CFTO in detecting fair blocks, in terms of increasing the probability of approving fair blocks and decreasing the probability of approving unfair blocks, by implementing experiments and comparing them with Helix (Yakira et al., 2021), a previously proposed consensus algorithm for fair block selection. As part of our experiments, we also compare certain features with those of previous studies.

Blockchain leader-based protocols elect leaders for proposing the next block of transactions. Proposed blocks need to pass a validation routine in order to be added to the blockchain. Proposers may prioritize certain transactions based on their fees or accounts. A fair block selection follows a random selection of transactions among pending transactions that a proposer is aware of. The validators may only have partial knowledge of the network transactions making it challenging to validate the random selection. We propose a protocol to encourage fair block selection in a leader-based blockchain network. Our protocol offers two main contributions. First, suggesting an algorithm that evaluates the proposed blocks based on both their transactions’ issuance times and zone structure. Second, providing incentives for acting honestly and diminishing malicious and dishonest nodes. To accomplish this, we use a reputation system, whereby each node is given a reputation score based on its actions (i.e. latest proposals and evaluations). We demonstrate the improved accuracy of our protocol by implementing experiments based on Ethereum topology, comparing it with Helix [1], an existing consensus algorithm for a fair block selection.

A functional PIR array code is a coding scheme which encodes some s information bits into a t × m array such that every linear combination of the s information bits has k mutually disjoint recovering sets. Every recovering set consists of some of the array's columns while it is allowed to read at most ℓ encoded bits from every column in order to receive the requested linear combination of the information bits. Functional batch array codes impose a stronger property where every multiset request of k linear combinations has k mutually disjoint recovering sets. Given the values of s, k, t, ℓ, the goal of this paper is to study the optimal value of the number of columns m such that these codes exist. Several lower bounds are presented as well as explicit constructions for several of these parameters.