Skip to main content

The Machine and BigData

Listen:

HP’s „The Machine“ (1) project is in my eyes the most advanced in the IT world with the simple goal to rethink the entire computer design. And the plan is ambitious – the first edge devices shall be ready in 2018, industrialized series in 2020.

Will “The Machine” really revolutionize an entire industry mostly influenced by IBM? Let’s say it could and probably will with a high percentage of success.
Based on the idea of Memristor (2) the project uses memory based technology to store data. Nothing new here. New is the non-volatile usage. Data, stored in an Memristor, persists unless the storing bit gets cleaned and new aligned. Now, NVRRAM (non-volatile resistive RAM) it’s faster as volatile DDR4 modules (which they use at the moment until Western Digital can deliver NVRRAM modules) and factor 100x faster than current state-of-the-art SSD based technologies. The newest prototype has 40 nodes with approx. 160 TB DDR4-RAM and 1,280 Cores connected with X1 PM’s (Photonic Modules). Means: pretty fast. Anyhow, just follow the appendix (1) to get more interesting engineering facts.

The most important consideration is the pure permanent all-integrated storage itself. The part of attached storage (like HDFS, GFS, Ceph) would simply disappear and directly merge with the computation layer. The principle “local data first” will surely be a part of any fine-tuning approach but with the high density of storage that will not really matter. All pieces of computation will be at the same place (cache, volatile and permanent storage combined with fast caching) and work as one homogenous entity which can hold every state of every piece of data during the whole computation lifecycle.
I just want to consider the changing fundamentals of that idea and what that would mean to data processing. The first big difference – a trinity memristor can store 10 bits instead of 8 today. That means simply a 3 times higher data storage density than today. Additionally, the highly volatile cache a CPU uses during the calculation process will be stored permanently which allows following processes to reuse the pre-calculated subsets and that would speed up any calculation dramatically. As for example in pattern detection algorithms like MCMC (3) could highly benefit simply by picking up the already calculated subset and use it in a new chain which would revolutionize data intelligence in terms of speed and tree generation. I think thats an huge step into the AI world - ultrafast learning algorithms helping the mankind to operate high sensitive environments like deep- space flights, connected cars, CEP networks or decentralized power grids.

(1) https://www.labs.hpe.com/the-machine
(2) http://en.wikipedia.org/wiki/Memristor
(3) https://en.wikipedia.org/wiki/Markov_chain_Monte_Carlo

Comments

Popular posts from this blog

Deal with corrupted messages in Apache Kafka

Under some strange circumstances, it can happen that a message in a Kafka topic is corrupted. This often happens when using 3rd party frameworks with Kafka. In addition, Kafka < 0.9 does not have a lock on Log.read() at the consumer read level, but does have a lock on Log.write(). This can lead to a rare race condition as described in KAKFA-2477 [1]. A likely log entry looks like this: ERROR Error processing message, stopping consumer: (kafka.tools.ConsoleConsumer$) kafka.message.InvalidMessageException: Message is corrupt (stored crc = xxxxxxxxxx, computed crc = yyyyyyyyyy Kafka-Tools Kafka stores the offset of each consumer in Zookeeper. To read the offsets, Kafka provides handy tools [2]. But you can also use zkCli.sh, at least to display the consumer and the stored offsets. First we need to find the consumer for a topic (> Kafka 0.9): bin/kafka-consumer-groups.sh --zookeeper management01:2181 --describe --group test Prior to Kafka 0.9, the only way to get this inform

Hive query shows ERROR "too many counters"

A hive job face the odd " Too many counters:"  like Ended Job = job_xxxxxx with exception 'org.apache.hadoop.mapreduce.counters.LimitExceededException(Too many counters: 201 max=200)' FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.MapRedTask Intercepting System.exit(1) These happens when operators are used in queries ( Hive Operators ). Hive creates 4 counters per operator, max upto 1000, plus a few additional counters like file read/write, partitions and tables. Hence the number of counter required is going to be dependent upon the query.  To avoid such exception, configure " mapreduce.job.counters.max " in mapreduce-site.xml to a value above 1000. Hive will fail when he is hitting the 1k counts, but other MR jobs not. A number around 1120 should be a good choice. Using " EXPLAIN EXTENDED " and " grep -ri operators | wc -l " print out the used numbers of operators. Use this value to tweak the MR s

AI's False Reality: Understanding Hallucination

Artificial Intelligence (AI) has leapfrogged to the poster child of technological innovation, on track to transform industries in a scale similar to the Industrial Revolution of the 1800s. But in this case, as cutting-edge technology, AI presents its own unique challenge, exploiting our human behavior of "love to trust", we as humans face a challenge: AI hallucinations. This phenomenon, where AI models generate outputs that are factually incorrect, misleading, or entirely fabricated, raises complex questions about the reliability and trust of AI models and larger systems. The tendency for AI to hallucinate comes from several interrelated factors. Overfitting – a condition where models become overly specialized to their training data – can lead to confident but wildly inaccurate responses when presented with novel scenarios (Guo et al., 2017). Moreover, biases embedded within datasets shape the models' understanding of the world; if these datasets are flawed or unreprese