Scientific Data Management Group LBNL

1
Scientific Data Management Group LBNL

• SRM related demos SC 2002
• Demos
• Robust File Replication of Massive Datasets on
  the Grid
• GridFTP-HPSS access Provided through HRM
• HRM enables NCARs MSS to be accessed on the Grid
• Uniform Grid Access to Different Mass Storage
  Systems
• Projects
• Earth Science Grid Use of HRMs in ESG Demo, new
  HRM-NCAR included
• Particle Physics Data Grid File replication for
  STAR project Demo

2
Demo Robust File Replication of Massive Datasets
on the Grid

• This demo illustrates
• Replicating Thousands of Files Robustly between
  Two remote HPSS installations
• Using a single command for a request to
  replicate all files
• Using an intermediate middleware storage
  management (called Hierarchical Storage
  Manager - HRM)
• Using a client HRM command-line module
• Without any modifications to the MSS
• Synchronizing pre-staging, file transfer, and
  archiving
• Robustness recovers from MSS and network
  transient failures

Junmin Gu, Vijaya Natarayan, Alex Sim, Arie
ShoshaniLBNL http//sdm.lbl.gov/srm
3
SRMs in Action PPDG
Anywhere
HRM-Client Command-line Interface
HRM-COPY (thousands of files)
BNL
LBNL
HRM-GET (one file at a time)
HRM (performs writes)
HRM (performs reads)
GridFTP GET (pull mode)
Network transfer
archive files
stage files
4
Web-Based File Monitoring Tool

• Shows
• Files already transferred- Files during
  transfer
• Files to be transferred
• Also shows for
• each file
• Source URL
• Target URL
• Transfer rate

5
Recent Measurements of large multi-file
replication
Shows that the network is the bottleneck
6
Demo GridFTP-HPSSAccess Provided through HRM

• This demo illustrates the
  capability of
• Providing GridFTP access to Mass Storage Systems
  (MSSs)
• Using an intermediate middleware storage
  management (called Hierarchical Storage
  Manager - HRM)
• Without any modifications to the MSS
• Managing queues of multiple requests to the MSS
• Minimizing tape mounts
• Recovers from MSS transient failures

Alex Romosan, Alex Sim, Junmin Gu, Arie
Shoshani http//sdm.lbl.gov/srm
7
Where does this service belongin the Grid
architecture?

.
N
G
O
R
2
S
T
O
R

O
I
E

E
O
T

Request
Workflow or
C

C
N
V
I
A
Community
Application-
Consistency Services
L
I
I
I
F
Interpretation
Request
C
A
T
V
I
A
I
Authorization
Specific Data
(e.g., Update Subscription,
C
C
R
M
U
L
and Planning
Management
E
T
E
E
P
O
E
Services
Discovery Services
Versioning, Master Copies)
L
P
S
R
P
V
Services
Services
D
I
L
S
I
A
V
T
O
C
C
E
L
L

G
R
O
1

N
S
O
C
I
E
L
E
E
F
T
V
L
Data Filtering or
A

C
General Data
Storage
Compute
Data
Monitoring/
Data
A
I
S
P
R
T
R
N
E
I
Transformation
E
Discovery
Management
Scheduling
Transport
Auditing
Federation
T
U
C
I
C
D
N
L
E
O
I
Services
Services
(Brokering)
(Brokering)
Services
Services
Services
E
V
R
U
L
S
L
G
R
O
M
E
O
E
O
R
C
S
C
E
L

S
G
E
E
N
Hierarchical
C
Resource
I
C
Compute
Data Filtering or
Database
File Transfer
R
S
R

Monitoring/
U
Resource
Resource
Transformation
Management
Service
U
G
O
N
O
Auditing
Manager
Management
Services
Services
(GridFTP)
S
I
S
R
E
E
A
R
R
H
S
Y
T
I
V
I
T
Communication
Authentication and
C
E
Protocols (e.g.,
Authorization
N
TCP/IP stack)
Protocols (e.g., GSI)
N
O
C
C
I
Other Storage
R
Mass Storage System (HPSS)
Compute
B
Networks
A
Systems
This figure based on the Grid Architecture paper
by Globus Team
F
systems
8
HRMs and GridFTP
Using HRM protocol
New GridFTP-HPSS through HRM
Client
Client
GridFTP-API
SRM-API
GridFTP entry
HRM
SRM-API
HRM
GridFTP-API
GridFTP
GridFTP move
9
GridFTP-HRM-Layerimplementation detail
Client
GridFTP-API
1a
1b
GridFTP entry GridFTP move GridFTP exit
FTP- HRM Layer
2a
Shared memory
2b
HRM
Corba
3b
3a
1a stor/retv 1b hrm_get/hrm_put
2b call_back 2a unblock semaphore
3a success_code 3b hrm_release
10
Demo HRM enables NCARs MSS to be accessed on
the Grid

• This demo illustrates
• The application of Storage Resource Management
  (SRM) technology to make NCARs MSS Grid-enabled
• NCARs MSS is a unique system developed in house
• Making HRM-HPSS grid-enabled is a big task
• HRM is a specialized SRM designed for Mass
  Storage Systems (MSS). HRM-HPSS already
  exists
• Adapting HRM-HPSS to work with HRM-HPSS is a
  small task
• Demo illustrates success and ease of this
  approach

Alex Romosan, Junmin Gu, Alex Sim, Arie
Shoshani LBNL http//sdm.lbl.gov/srm
11
How Was It Done?only MSS-specific module modified
HRM-HPSS
HRM-HPSS
Disk Resource Manager (DRM)
Disk Resource Manager (DRM)
Tape Resource Manager (TRM)
Tape Resource Manager (TRM)
HPSS-specific Access Module
NCAR-specific Access Module
Specialize for NCAR-MSS
NCAR-MSS
12
Demo details File Replication from NCAR MSS to
NERSC HPSS controlled at NCAR
13
SRMs in Action ESG

14
Demo Uniform Grid Accessto Different Mass
Storage Systems

• This demo illustrates Uniform Grid Access to
• three MSSs HPSS, Enstore, JASMine
• Using the Storage Resource Manager (SRM)
  interface
• SRM modules were developed independently for
  the 3 MSSs
• Shows the power of standard SRM functionality

Contributors JLAB Bryan Hess, Andy
Kowalski FNAL Don Petravick, Timur Perelmutov,
Rich Wellner LBNL Junmin Gu, Vijaya Natarayan,
Ekow Otoo, Alex Romosan, Alex Sim, Arie
Shoshani
15
High Level View of Demo Setup
Client (USER/APPLICATIONS)
uniform interface
SRM
SRM
JASMine
SRM
Enstore
16
Where do SRMs belongin the Grid architecture?

.
G
N
O
R
2
S
O
T
R

O
I
E

E
O
T

Request
Workflow or
C

V
C
N
I
A
Community
Application-
Consistency Services
L
I
I
I
F
Interpretation
Request
C
A
T
V
A
I
I
Authorization
Specific Data
(e.g., Update Subscription,
C
C
R
M
U
L
and Planning
Management
E
T
E
E
P
O
E
Services
Discovery Services
Versioning, Master Copies)
L
P
S
R
P
V
Services
Services
D
I
L
S
I
A
V
T
O
C
C
E
L
L

G
R
O
1
S

N
O
C
I
E
L
E
E
F
T
V

L
C
Data Filtering or
A
General Data
Storage
Compute
Data
Monitoring/
Data
I
A
S
P
R
T
R
N
I
E
Transformation
Discovery
Management
Scheduling
Transport
Auditing
Federation
E
T
U
C
I
C
D
L
N
E
O
I
Services
Services
(Brokering)
(Brokering)
Services
Services
Services
L
E
V
R
U
S
L
M
G
R
O
E
O
E
O
R
C
S
C
E
L

S
G
E
E
N
C
Resource
Storage
I
C
Compute
Data Filtering or
Database
File Transfer
R
S
R

Monitoring/
U
Resource
Resource
Transformation
Management
Service
U
G
O
N
O
Auditing
Manager
Management
Services
Services
(GridFTP)
S
I
S
R
E
E
A
R
R
H
S
Y
T
I
V
I
T
Communication
Authentication and
C
E
Protocols (e.g.,
Authorization
N
TCP/IP stack)
Protocols (e.g., GSI)
N
O
C
C
I
Other Storage
R
Mass Storage System (HPSS)
Compute
B
Networks
A
Systems
F
This figure based on the Grid Architecture paper
by Globus Team
systems
17
Main advantages of using SRMs

• Same uniform interface to all types of storage
  systems
• Disk Storage System
• Mass Storage System
• For Mass Storage Systems it provides
• Queuing and pre-staging
• Queued multi-file get requests (avoid flooding
  MSS)
• Pre-staging of files (concurrent with transfer)
• Queued archiving of put requests (avoid
  flooding MSS)
• Robustness and efficiency
• Recover in case of transient failures from MSS
• Reorders pre-staging requests to minimize tape
  mounts
• Recovery from failed GridFTP transfers
• Re-issues requests in case of failure