Since 1 Aug 2016, I am a Research Associate within the Digital Technology Group, working on optimising provenance systems with the goal of dealing with large quantities of data, as produced by low-level tracing systems. Within the CADETS project, such tracing systems (like DTrace) are used for detecting Advanced Persistent Threats (APTs). I also contribute to work being done on developing Distributed DTrace and to the development of new versioning models for provenance.
I have submitted my PhD thesis, entitled "Provenance-based computing" on 20 Aug 2016 and have passed my examinations (viva) on 5 July 2016. I am in the process of finalising the corrections required for submitting the final version of the thesis. [draft pdf version available on request]
I have started my PhD in the University of Cambridge Computer Laboratory in October 2011, as a member of the Digital Technology Group. My supervisor was Professor Andy Hopper and I have worked closely with Dr. Ripduman Sohan in exploring the research area of provenance, looking at building systems for tracking the history and inter-dependencies of pieces of data as they are produced or consumed in arbitrary computational workflows.
My PhD research was funded by UK's EPSRC, by a CHESS Scholarship, and by a Computer Laboratory Departamental Scholarship.
I am a member of Wolfson College.
- Data provenance, which is the topic of my PhD
research. The central idea that I put forward is the fact that
applications can query fine-grained provenance (the history of
where each piece of data comes from together with its
dependencies) in order to understand the causes of
unexpected characteristics of their execution (poor performance, errors).
- A primer on provenance. Lucian Carata, Sherif Akoush, Nikilesh Balakrishnan, Thomas Bytheway, Ripduman Sohan, Margo Seltzer, and Andy Hopper (2014). Communications of the ACM, 57(5), 52-60.
- Resource accounting and monitoring: I am
working, together with Dr Ripduman Sohan,
Oliver Chick and James Snee
on a fine-grained kernel-level resource accounting
framework that allows applications to gain a better understanding
of the resources consumed (either synchronously or asynchronously)
by kernel activities. This gives a detailed, per kernel subsystem
view of the resource usage of particular system calls and of how
this varies in the presence of workloads that compete for
the same resources. The resulting accounting data can be
aggregated according to application semantics (e.g. resources
consumed over servicing a client request).
- Shadow kernels: A general mechanism for kernel specialization in existing operating systems Oliver R.A. Chick, Lucian Carata, James Snee, Ripduman Sohan. Proceedings of the 6th Asia-Pacific Workshop on Systems (ApSys), 2015
- Soroban: Attributing latency in virtualized environments James Snee, Lucian Carata, Oliver R.A. Chick, Ripduman Sohan, Ramsey M. Faragher, Andrew Rice and Andy Hopper. USENIX Workshop on Hot Topics in Cloud Computing (HotCloud), 2015
- Resourceful: fine grained resource accounting for explaining service variability Lucian Carata, Oliver Chick, James Snee, Ripduman Sohan and Andy Hopper (2014). Technical Report, UCAM-CL-TR-859
- Linux Containers - basic concepts a talklet covering the basic mechanisms used by linux containers.
- Automated configuration of distributed systems: My intrest in this area is in understanding how certain configuration parameters are propagated/used in a loosely coupled system, and how having this information can help establishing automated feedback control loops in order to keep fulfilling SLOs even in the presence of errors (for example, during failed service updates and on human configuration errors).
- Computer Networks (IB) - since 2016
- Databases (IA, IB) - since 2016
- Computer Graphics and Image Processing (IB) - since 2012
- Algorithms (IA) - [2011 - 2015]
I have previously been an intern at both Microsoft (2010) and Google (2013), where I have worked in infrastructure teams dealing with performance evaluation and service configuration, respectively. My time within both companies has helped me gain valuable experience and a better understanding of real-world requirements at scale. This continues to drive my interests towards research that has a real impact and solves challenging engineering problems.
I have graduated (B.Eng.) as valedictorian from the "Gh. Asachi" Technical University of Iasi, Romania. My bachelor thesis dealed with real-time object recognition in video streams using interest point detectors. I have obtained a M.Sc. in Distributed Systems from the same university, with a research stint in the Institute of Computer Graphics and Algorithms from the Vienna University of Technology, during an ERASMUS Scholarship.
If you want to securely communicate with me or make sure a particular software artifact was released (signed) by me, please download my OpenPGP public key 0x1AFE0D9B85AA6C3E from one of the public key servers or from this key file (https, local copy).
importing the key into your keyring, please verify that its fingerprint
I use different subkeys for encryption/authentication/signatures (given below with their ids). Unless compromised, the validity date of most of those subkeys will be prolongued before they expire. An exception is the encryption subkey, which I might choose to rotate in order to limit the number of communication sessions that are leaked if the corresponding private key is compromised.
- 0xC4CA98A189E620F8: Encryption key 1; Use this to encrypt any data you want to send me privately. Do not use for urgent communication, as there might be significant latency before I can decrypt the message
- 0x1201F2F9E2846C6C: Encryption key 2; Encryption key that can be used for urgent communication
- 0x7006F6140EB6A1B1: Signing key used for software artifacts released by me
- 0x5C9399D6F3273144: Authentication key used privately